EFFECTS OF INDIVIDUAL CHARACTERISTICS AND SYMPTOMS ON PHYSICAL
FUNCTION IN PERSONS WITH LUMBAR DEGENERATIVE CONDITIONS
By
Teri Lynn Holwerda

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Nursing-Doctor of Philosophy
2014

ABSTRACT
EFFECTS OF INDIVIDUAL CHARACTERISTICS AND SYMPTOMS ON
PHYSICAL FUNCTION IN PERSONS WITH LUMBAR DEGENERATIVE CONDITIONS
By
Teri Lynn Holwerda
Background/Significance: Back pain affects 80 percent of persons at some point in their lives.
Lumbar disc degeneration, stenosis and facet joint degeneration have been associated with low
back pain. Degenerative changes increase with age. Genetic influences affect the spinal
degenerative process and the experience of pain. The symptoms that accompany degenerative
spinal conditions include back pain, leg pain, numbness and weakness. Low back and leg pain
are associated with reduced physical function. Physiological, situational and psychological
patient characteristics influence physical function in degenerative lumbar conditions. These
characteristics include genotype, BMI, smoking, age, employment status, insurance type,
worker’s compensation claim and depression. Problem: Little is known about the interaction
among patient characteristics and symptoms and the outcome of physical function for persons
with lumbar spinal degeneration. Purpose: This study was undertaken to explore the
contribution of patient characteristics and symptoms to the outcome of physical function in a
population of individuals experiencing lumbar degenerative conditions. Specific Aims: 1)
Determine the contribution of physiological (BMI, sex, age, smoking status), situational
(employment status, worker’s compensation claim, insurance type), and psychological
(depression) factors in persons receiving non-surgical interventions for degenerative lumbar
conditions to symptoms and physical function, 2) Develop a predictive model for the outcome of
physical function in persons receiving non-surgical interventions for lumbar degenerative

conditions, using symptoms (back and/or leg pain, numbness, and weakness) and physiological,
situational, and psychological patient factors, and 3) Explore the impact of the physiological
factor genotype (disc structural genes and pain genes) on symptoms (back and/or leg pain,
numbness, and weakness) and on physical function in persons experiencing lumbar degenerative
conditions. Instruments: Physical function is the primary outcome, measured by the physical
function subscale of the SF-36 and the Oswestry Disability Index, (ODI). Methods: Using a
cross-sectional, observational design, 163 subjects were randomly selected from an existing
database of completed SF-36 and ODI questionnaires at a tertiary outpatient spine center. Data
on symptoms and physiological, situational, and psychological characteristics were obtained
from the medical record. A random subset of 28 subjects consented to provide saliva samples
for genotyping. Results: Aim 1: Smoking, having Medicaid insurance or no insurance were
negatively associated with the symptom pain VAS. Higher BMI and smoking were associated
with worse ODI scores, while having Commercial insurance or Medicare was associated with
better ODI scores. Higher BMI, smoking, older age, and having Medicaid insurance were
associated with worse SF-36 physical function subscale scores. Aim 2: Higher BMI, smoking,
higher pain VAS and numbness predicted 35% of the variance in ODI scores. Higher BMI, older
age and the symptom higher pain VAS predicted 26% of the variance in SF-36 physical function
subscale scores. Aim 3: No genotype was significantly associated with symptoms. OPRM1
A/A carriers had significantly worse physical function scores than those with */G alleles.
Implications: This study is an important step in identifying the combination of patient
characteristics, (including genotype) and symptoms that impact physical function in this
population, in order to tailor interventions to preserve physical function.

Copyright by
TERI LYNN HOLWERDA
2014

ACKNOWLEDGEMENTS

My only comfort in life and death is that I belong to my faithful savior, Jesus
Christ. My hope is in Him.
Over the past five years, I have received assistance from many individuals. Dr.
Debra Schutte has been a valuable mentor, friend and role model. Her sage and patient
counsel have steadied my focus.
I would like to thank the Chair of my Dissertation Committee, Dr. Barbara Given.
I have benefitted from her knowledge and experience as a nurse researcher. I thank her for
holding me to a high standard. I would also like to thank the members of my Dissertation
Committee, Dr. Amy Hoffman, Dr. Barbara Smith, Dr. Paul Stephenson and Dr. Daniel
Vaughn. I am indebted to Amy for her symptom expertise and kind encouragement. I
thank Barbara Smith for her invaluable critique of my writing, and Dan for his expertise in
physical function and for sharing his priceless skills as a journal editor. I am thankful for
the contagious enthusiasm Paul has for statistical analysis. Because of his patient
guidance, I was enthralled by the revelations of my data, even in the last stages. I survived
because of all of you. I am also grateful to Dr. Alan Davis, Director of Research at Grand
Rapids Medical Education Partners, for his kind assistance.
I am indebted to my family, husband BJ, and daughters Andrea (and Kyle) and
Olivia for their unwavering support and encouragement. Words cannot express how much
I love you all.
I could not have completed this work without the generous support in the form of a
dissertation small grant from the Saint Mary’s Foundation. I am also grateful for the

v

financial support I received from the College of Nursing and the Graduate School,
specifically the John F. Dunkel Memorial Endowed Scholarship, College of Nursing
Fellowships, College of Nursing Scholarships, a Graduate School Fellowship, and a
Dissertation Completion Fellowship.
Last, and not least, I would like to thank the staff of the Research & Innovation
department at Mercy Health Saint Mary’s for their guidance and direction. Their positive
approach kept me energized.

vi

TABLE OF CONTENTS

LIST OF TABLES

x

LIST OF FIGURES

xiv

CHAPTER I Introduction
Background and Significance
Factors Affecting Outcome in Persons with Lumbar Degenerative Conditions
Purpose of the Study
Specific Aims
Outcome of Interest
Physical Function Definition
Physical Function in Persons with Lumbar Degenerative Changes
Lumbar Degeneration
Anatomic Changes
Economic Problem
Diagnostic and Treatment Variation
Genetics and Lumbar Degeneration
Genetics and the Experience of Low Back Pain
Patient Characteristics and Effects on Various Outcomes in
Lumbar Degeneration
The Effect of Symptoms on Outcomes
Knowledge Gap

1
1
2
3
3
4
4
5
5
5
6
6
7
7
8
10
11

CHAPTER II Conceptual Framework
Use of the TOUS in This Study

13
26

CHAPTER III Review of the Literature
29
Lumbar Spinal Anatomy and Degenerative Changes
30
Physiological Factors
31
Obesity
31
Sex and Age
32
Smoking
33
Genotype
34
Selected Candidate Genes for Disc Structure
34
Collagen IX Alpha 2 and Alpha 3 (COL9A2 and COL9A3)
Genes
35
Aggrecan (ACAN) Gene
36
Vitamin D Receptor(VDR) Gene
37
Selected Candidate Genes for Pain
38
Opioid Receptor, mu-1 (OPRM1) Gene
38
Catechol-o-Methyltransferase (COMT) Gene
40
Situational Factors

42
vii

Employment Status
Worker's Compensation
Insurance Type
Psychological Factors
Depression
Symptoms in Persons with Lumbar Degeneration
Physical Function
CHAPTER IV Methods
Research Design
Sample
Setting
Instruments and Measures
Patient Factors
Physiological factors
Body Mass Index
Sex
Age
Smoking status
Genotype
Situational factors
Employment status
Insurance type
Psychological factors
Depression
Symptoms
Pain
Numbness
Weakness
Physical Function
Oswestry Disability Index (ODI)
Physical function subscale of the SF-36
Medications
Procedures
Recruitment Procedures for Aims 1 and 2
Recruitment Procedures for Aim 3
Data Collection Procedures
Gentoyping Procedures
Data Management
Data Analysis
Aim 1 Specific Strategies
Aim 2 Specific Strategies
Exploratory Aim 3 Specific Strategies
Limitations
Human Subjects
Human subjects characteristics and involvement
Sources of material

viii

42
43
44
45
45
45
47
51
51
52
54
54
55
55
55
55
55
56
56
57
57
58
58
58
59
59
62
62
64
64
67
69
69
70
70
72
72
74
75
75
76
76
77
78
78
79

Potential risks
Protection against risk
Potential benefits of the proposed research to subjects and
others
CHAPTER V Results and Interpretation
Organization of Results Chapter
Medical Records Reviewed
Demographic Information and Patient Characteristics for the Sample
Symptoms for the Study Population
Outcome Measures for the Sample
Aim 1 Analysis
The relationship between patient characteristics and pain VAS
The relationship between patient characteristics and weakness and
numbness
The relationship between patient characteristics and pain location
The relationship between patient characteristics and outcome
measures
The relationship between pain location and physical function
Aim 2 Analysis
Summary of Findings for Aims 1 and 2
Aim 3 Study Subjects
Demographics and Patient Characteristics for Genotyped Subjects
Outcome Measures for the Genotyped Subjects
Genotyping Results
Aim 3 Analysis
Relationship between genotype and symptoms
Relationship between genotype and outcome measures
CHAPTER VI Discussion and Implications
Discussion of Sample Patient Characteristics
Discussion of Sample Physiological Characteristics
Discussion of Sample Situational Characteristics
Discussion of Sample Psychological Characteristic
Discussion of Symptoms of Sample
Discussion of Sample Outcome Measures
Discussion of Results for Specific Aim 1
Discussion of Associations between Patient Characteristics and
Symptoms
Discussion of Associations between Patient Characteristics and
Physical Function
Discussion of Additional Results
Discussion of Results for Specific Aim 2
Discussion of Results for Exploratory Aim 3
Discussion of Associations between Genotype and Symptoms
Discussion of Associations between Genotype and Physical
Function

ix

79
80
80
82
82
82
84
88
89
91
91
93
97
99
104
106
109
110
110
113
114
115
116
119
123
123
123
124
125
127
128
129
129
131
134
135
138
138
140

Study Limitations
Implications for Nursing Practice
Implications for Research
Implications for Policy
Conclusion/Summary

141
144
146
149
151

APPENDICES
Appendix A Figures
Appendix B Oswestry Disability Index
Appendix C Data Collection Tool
Appendix D Permission to Use TOUS
Appendix E Informed Consent
Appendix F Data Use Agreement

153
154
168
169
173
176
182

REFERENCES

184

x

LIST OF TABLES

Table 1 Study Variables

63

Table 2 Genes Selected for Genotyping with SNPs (Single-Nucleotide Polymorphisms)
Tested
74
Table 3 Sample N and % of Overweight, Obese and Class III Obese (N = 163)

85

Table 4 Insurance Coverage for the Sample and for Working Subjects
in the Sample

86

Table 5 Sample Patient Characteristics, N and % for Categorical
Variables (N = 163)

87

Table 6 Sample Patient Physiological Characteristics, Range, Minimum,
Maximum, Mean and SD for BMI and Age (N = 163)

87

Table 7 Sample Symptom Continuous Variable: Pain VAS (N = 163)

88

Table 8 Sample Symptom Categorical Variables: Pain Location, Weakness
and Numbness (N = 163)

89

Table 9 Sample Physical Function Scores: Range, Minimum, Maximum,
Mean and SD

90

Table 10 Coefficients and Observed Levels of Significance for the Full and
Final Multiple Regression Models for Predicting Pain VAS Using
Patient Characteristics (BMI, Sex, Age, Smoking, Employment
Status and Depression) (N = 163)

92

Table 11 Coefficients and Observed Levels of Significance for the Full and
Final Multiple Regression Models for Predicting Pain VAS Using
Patient Characteristics (BMI, employment status and smoking) and
Insurance Type (N = 163)

93

Table 12 Logistic Regression for Predicting Weakness Using Patient
Characteristics, Full and Final Models (N = 163)

95

Table 13 Classification Table for Full and Final Logistic Regression for
Predicting Weakness Using Patient Characteristics (N = 163)

95

Table 14 Logistic Regression for Predicting Numbness Using Patient
Characteristics, Full and Final Models (N = 163)

96

xi

Table 15 Classification Table for Full and Final Logistic Regression for
Predicting Numbness Using Patient Characteristics (N = 163)

96

Table 16 Discriminant Analysis Patient Characteristics (BMI, Age,
Sex, Smoking, Employment Status and Depression) as Predictors of
Pain Location (Back Pain Only, Back and Leg Pain, Leg Pain Only) (N = 162) 98
Table 17 Chi-square Categorical Patient Characteristics (Sex, Depression,
Worker’s Compensation) as Predictors of Pain Location (Back Pain
Only, Back and Leg Pain, Leg Pain Only) (N = 162)

98

Table 18 Analysis of Variance for Continuous Patient Characteristic
(BMI) as Predictor of Pain Location (Back Pain Only, Back Pain
and Leg Pain, Leg Pain Only) (N = 162)

98

Table 19 Coefficients and Observed Levels of Significance for the Full and
Final Backward Regression Models for Predicting ODI Score Using
Patient Characteristics (N = 162)

100

Table 20 Coefficients and Observed Levels of Significance for the Full and
Final Multiple Regression Models for Predicting ODI Scores Using
Patient Characteristics (BMI, Employment Status and Smoking) and
Insurance Type (N = 162)

101

Table 21 Coefficients and Observed Levels of Significance for the Full and
Final Backward Regression Models for Predicting SF-36 Physical
Function Subscale Score Using Patient Characteristics (N = 161)

102

Table 22 Coefficients and Observed Levels of Significance for the Full and
Final Multiple Regression Models for Predicting SF-36 Physical
Function Subscale Scores Using Patient Characteristics (BMI,
Depression, Age, Employment Status and Smoking) and Insurance
Type (N = 161)

103

Table 23 One-way ANOVA for Between Groups Difference for Pain Location
and ODI Scores (N = 162)

104

Table 24 Multiple Comparison Test to Determine Which Means Differed for
Pain Location and ODI Scores (N = 162)

105

Table 25 One-way ANOVA for Between Groups Difference for Pain Location
and SF-36 Physical Function Subscale Scores (N = 161)

106

Table 26 Coefficients and Observed Level of Significance for the Final
Backward Step-wise Multiple Regression for Predicting ODI
Scores Using Patient Characteristics (BMI, Smoking, Pain VAS
and Symptoms (Extremity Numbness) (N = 162)

108

xii

Table 27 Coefficients and Observed Level of Significance for the Full and
Final Backward Step-wise Multiple Regression for Predicting SF-36
Physical Function Subscale Scores Using Patient Characteristics
(BMI and Age) and Symptoms (Pain VAS) (N = 161)

109

Table 28 Genotyped Subjects Patient Characteristics, N and % for Categorical
Variables (N = 28)

112

Table 29 Genotyped Subjects Patient Characteristics, Range, Minimum,
Maximum, Mean and SD for Continuous Variables (N = 28)

113

Table 30 Genotyped Subjects Symptom Continuous Variable:
Pain VAS (N = 28)

113

Table 31 Genotyped Subjects Physical Function Scores: Range, Minimum,
Maximum, Mean and SD

114

Table 32 One-way ANOVA for Between Groups Difference (Genotypes)
COL9A2, COL9A3, OPRM1, COMT and VDR and Pain VAS (N = 28)

117

Table 33 Chi-square Tests for Genotype as Predictors of Pain Location
(Back Pain Only, Back Pain and Leg Pain, Leg Pain Only) (N = 28)

118

Table 34 One-way ANOVA for Between Groups Difference (Genotypes)
COL9A2, COL9A3, OPRM1, COMT and VDR and ODI Scores (N = 27)

120

Table 35 One-way ANOVA for Between Groups Difference (Genotypes)
COL9A2, COL9A3, OPRM1, COMT and VDR and SF-36 Physical
Function Subscale Scores (N = 28)

122

xiii

LIST OF FIGURES

Figure 1 The Theory of Unpleasant Symptoms with Study Variables

154

Figure 2 Neurosurgery/Spine Health History

155

Figure 3 SF-36

159

Figure 4 IRB Approval

165

Figure 5 Pain Diagram Overlay

167

xiv

CHAPTER I
Introduction

Background and Significance
Back pain and the symptoms that accompany lumbar degenerative conditions are a highly
prevalent and important health problem. More than 20% of adults responding to the 2009
National Health Interview Survey experienced back pain in the three months prior to the survey
(National Center for Health Statistics). Back pain affects about 80 percent of persons at some
point in their lives (Healthy People 2020, 2012). Individuals with back pain are likely to
continue to have recurrent episodes of back pain over time and two to ten percent of back pain is
chronic (Healthy People, 2020, 2012). Approximately one-third of persons develop persistent
low back pain one year after an acute pain episode (Von Korff & Saunders, 1996).
Though most episodes of back pain are self-limiting, back pain can affect physical
function (Carey, et al., 1996; Thomas et al., 1996; Samartzis et al., 2011; Chung-Wei, et al.,
2011). In fact, back pain is a frequent reason for visits to physicians, Emergency Departments,
and hospitalizations (Healthy People 2020, 2012). Conditions involving the low back comprise
the fifth most frequent cause of hospitalization and the third most common reason for surgery
(Healthy People 2020, 2012). Back pain is the second leading cause of work absence, after the
common cold (Healthy People 2020, 2012).
There are many causes for back pain. Degenerative conditions involving the
intervertebral disc have been identified as one cause of low back pain (Cheung, Samartzis,
Karppinen & Luk, 2012; Freemont, 2009; Livshits, et al., 2011; Takatalo et al., 2011). The
prevalence of disc degeneration is estimated to be as high as 40% for individuals under the age

1

of 30 (Cheung, et al., 2009). By age 50, the prevalence rises to 60-90% (Cheung, et al, 2009;
Kalichman, Kim, Li, Guermazi & Hunter, 2010).
Back pain is not the only problematic symptom of lumbar degeneration. As degeneration
progresses, the combination of facet joint arthritis and disc height reduction can produce changes
that decrease the diameter of the canal and neuroforamen, which can contribute to spinal nerve
symptoms of limb pain, numbness, tingling and weakness (Genevay & Atlas, 2010).
The population of the United States is aging. Current estimates predict that by the year
2015, 27% of the population will be age 55 or older and 14.4% of the population will be aged 65
or older (U.S. Census Bureau, 2012). Therefore, the prevalence of degenerative conditions
affecting the spine is anticipated to increase as well. More persons will develop progressive
degenerative spinal changes and therefore be at risk for the development of the symptoms that
accompany degenerative spinal conditions.
Factors Affecting Outcome in Persons with Lumbar Degenerative Conditions
Patient characteristics and their influences on outcomes for persons experiencing lumbar
degenerative conditions have been studied. Several situational, psychological and physiological
patient characteristics have been shown to affect physical function outcomes for persons
experiencing lumbar degenerative conditions. The physiological factors (defined as biologic and
physical features possessed by the individual) genotype, obesity, smoking and age have been
shown to influence lumbar degeneration and low back pain. The situational factors (defined as
features that are outside the individual that may influence health status) employment status,
worker’s compensation claim and insurance type can affect the physical function outcome of
non-surgical treatments for individuals with lumbar degenerative conditions. The psychological
factor (defined as mental state or mood) depression has been associated with greater pain and

2

worse physical function in persons with lumbar degenerative conditions. Symptom experience
can influence outcome, and in general, the worse the symptom experience, the more negative the
impact on the outcome of physical function.
Purpose of the Study
The purpose of this cross-sectional observational study is to explore the contributions of
patient characteristics (including genotype), and symptoms to the outcome of physical function
in a population of individuals experiencing lumbar degenerative conditions. The goal is to begin
to develop a method to identify persons with lumbar degenerative conditions at increased risk of
experiencing decreased physical function, in order to tailor interventions or adjust treatment
approaches to improve outcomes for the entire population.
Specific Aims
Aim 1: To determine the contribution of physiological (BMI, sex, age, smoking status),
situational (employment status, worker’s compensation claim, insurance type), and psychological
(depression) characteristics in persons receiving non-surgical interventions for degenerative
lumbar conditions to symptoms and physical function.
Aim 2: Develop a predictive model for the outcome of physical function in persons receiving
non-surgical interventions for lumbar degenerative conditions, using symptoms (back and/or leg
pain, numbness, and weakness) and physiological, situational, and psychological patient
characteristics.
Exploratory Aim 3: Explore the impact of the physiological characteristic genotype (disc
structural genes and pain genes) on symptoms (back and/or leg pain, numbness, and weakness)
and on physical function in persons experiencing lumbar degenerative conditions.

3

The expected outcome from this research will be knowledge about the symptom
experience in persons experiencing degenerative lumbar conditions, the interaction of symptoms
with patient factors influencing the outcome of physical function in this population, and the
development of predictive models to identify populations at risk for worse physical function. As
a result of this proposed investigation, it is expected that predictions based on symptoms and
patient factors will result in improved outcomes for persons with lumbar degenerative
conditions.
Outcome of Interest
Physical Function Definition
Physical function--defined as an individual’s ability to fully perform in the various
physical roles in their lives, to accomplish ADLS, to work, carry out daily tasks for self and
significant others, to be mobile, and maintain leisure physical activities--is a requisite part of
overall quality of life (Rejeski & Mihalko, 2001; Ferrans, et al., 2005). Physical function is
foundational to the ability to operationalize roles (Lenz, et al., 1997), forming the basis for an
individual’s ability to accomplish the activity required to provide for basic needs, fulfill life
roles, and maintain health and well-being (Leidy, 1994; Hoffman, et al., 2009).
Decline in physical function with aging can negatively affect cognitive function
(Eggermont, Milberg, Lipsitz, Scherder & Leveille, 2009). Decline in physical function is also
associated with increased mortality and greater risk of disability (Cawthon et al., 2011; Gillum &
Obisesan, 2010).

4

Physical Function in Persons with Lumbar Degenerative Changes
Lumbar spine degenerative changes can cause alterations in physical function. Low back
pain is associated with reduced physical function in younger and older adults (Samartzis et al.,
2011; Chung-Wei, et al., 2011). The presence of chronic low back pain and leg pain is
associated with greater disability and less optimum health (Prins, van der Wurff & Groen, 2013).
Lumbar degenerative changes increase with age (Bogduk, 2012).
Lumbar Degeneration
Anatomic Changes
Manifestations of lumbar degeneration include decreased height of the intervertebral disc,
bulging of the outer layer (annulus) of the intervertebral disc, facet joint hypertrophy, thickening
of the ligamentum flavum, and stenosis of the central canal, lateral recesses, and neuroforamen
(Chokshi, Quencer & Smoker, 2010; Genevay & Atlas, 2010; Varlotta et al., 2011). As the
intervertebral disc degenerates, more stress is placed on the facet joints, contributing to arthritis,
joint space narrowing, erosion of the joint, hypertrophy and the development of bone spurs
(Kalichman & Hunter, 2007; Modic, 2007). Facet joint degenerative changes and decreased disc
height are associated with hypertrophy and buckling of the ligamentum flavum, which then
encroaches on the spinal canal (Altinkaya, Yildirim, Demir, Alkan & Sarica, 2011; Chokshi,
Quencer & Smoker, 2010; Genevay & Atlas, 2010). Lumbar facet joints and the posterior
annulus of the intervertebral disc are enervated (Falco et al., 2012; Moon et al., 2012; Van
Zundert, Vanelderen, Kessels & van Kleef, 2012). Lumbar intervertebral disc degeneration,
lumbar stenosis and facet joint degeneration have been associated with low back pain (Cheung et
al., 2009; Cohen & Raja, 2007; Kalichman, Kim, Lee, Guermazi & Hunter, 2010; Moon, et al.,
2012; van Kleef et al, 2010).

5

Economic Problem
Estimates of the cost of low back pain in the United States vary widely, but sources
suggest the costs range from $50-625 billion per year (Dagenais, S., Caro, J. & Haldeman, S.
2008; Healthy People 2020). Many treatments are available for lumbar spinal conditions,
ranging from physical therapy to surgery. Wide variations in the approach to diagnosis and
treatment exist. Complex surgery rates for lumbar stenosis are on the rise, and significant
geographical differences in surgical rates have been identified (Deyo, Mirza, Martin, Kreuter,
Goodman & Jarvik, 2010; Weinstein, Lurie, Olson, Bronner & Fisher, 2006).
The charges for lumbar fusion surgery in the U.S. increased nearly eight-fold between
1998 and 2008, rising from 4.3 billion to 33.9 billion over that decade (Rajaee, Bae, Kanim &
Delamarter, 2012). Rising surgical costs have been fueled by increased instrumentation,
biologics, and device usage (Deyo, Mirza, Martin, Kreuter, Goodman & Jarvik, 2010; Weinstein,
Lurie, Olson, Bronner & Fisher, 2006). At best, overall success for lumbar spinal surgical
procedures has been estimated to be fifty percent; 25% persons undergoing spinal surgery
experience no improvement at all (Block, Gatchel, Deardorff & Guyer, 2003). Costs associated
with non-surgical treatment for lumbar disc herniations are also substantial (Daffner, Hymanson
& Wang, 2010).
Diagnostic and Treatment Variation
There is considerable variability in the classification of low back pain, given the many
different sources of pain in the lumbar spine (Fairbank, et al 2011). Low back pain is felt to be a
heterogenous condition with clinically distinct subgroups and different pain generators (Fourney,
et al., 2011). Because of the lack of consensus on the source and classification of low back pain,
there is variability in the recommendations for treatment (Benoist, Boulo & Hayem, 2012;

6

Cheng, et al., 2011; Choma, Schuster, Norvell, Dettori & Chutkan, 2011; Pereira et al., 2012).
There is therefore a need to begin to classify subgroups of patients whose profiles suggest a
higher risk for impairment of physical function.
Genetics and Lumbar Degeneration
The role of genetics in the development of lumbar degenerative conditions has been of
great interest in recent years. Hereditary and biological mechanisms contributing to disc
degeneration have been identified (Zhang, Sun, Liu & Guo, 2008). Heritability is the variance in
phenotype attributable to genetic factors (Holliday & McBeth, 2011). The heritability of lumbar
intervertebral disc degeneration has been estimated to be 29-61% (Battie, Videman, Levalahti,
Gill & Kaprio, 2008; Kalichman & Hunter, 2008). Lumbar disc degeneration is now considered
to be a complex process with both genetic and environmental contributors, and investigators
have identified several candidate genes that may be involved in the lumbar degenerative process
(Hadjipavlou, Tzermiadianos, Bogduk & Zindrick, 2008). Genes related to the integrity of the
intervertebral disc and genes related to the breakdown of disc components are among those
implicated in the process of disc degeneration.
In summary, lumbar disc degeneration is one cause of low back pain. Once considered a
consequence of mechanical stress, disc degeneration is now thought to be a complex process
related in part, to genetic as well as environmental factors.
Genetics and the Experience of Low Back Pain
Pain, like lumbar disc degeneration, is an etiologically complex phenomenon, likely
influenced by genetic and environmental factors. In fact, much is known about the genetics of
pain. For example, twin studies have demonstrated the heritability of the symptom of back pain.
Estimates of the heritability of low back pain ranges from 30-68% (Battie, Videman, Levalahti,

7

Gill & Kaprio, 2007; Hartvigsen et al, 2009; MacGregor, Andrew, Sambrook & Spector, 2004).
In addition, several genes have been implicated in the variability of the experience of pain,
among them, genes that code for opioid receptors and catechol-o-methyltransferase. Variability
in these genes has been implicated in an increased experience of pain, increased susceptibility to
pain, and differences in analgesic requirements for pain states (Argoff, 2010; Dai, F. et al., 2010;
Kim & Schwartz, 2010; Kleiber, et al., 2007; Miaskowski, 2009).
In summary, the experience of pain as a symptom in general is now known to be related
in part, to genetic factors. There is also accumulating evidence that pain genetics influence pain
states specifically in degenerative lumbar spinal conditions. And, while more is known
regarding genetic influences on the degenerative process involving lumbar intervertebral discs
and the genetic influences on the symptom of low back pain, there is a need for studies
examining the combined effects of pain and disc degeneration genotype on the symptoms of
lumbar degeneration and the outcome of physical function in this population.
Patient Characteristics and Effects on Various Outcomes in Lumbar Degeneration
Patient characteristics and their influence on outcomes for persons experiencing lumbar
degenerative conditions have been studied. Several situational, psychological and physiological
patient characteristics have been shown to affect aspects of pain and functional outcomes of
persons experiencing lumbar degenerative conditions. These outcomes have included functional
status, ability to return to work, intensity of the experience of pain and the development of
chronic pain.
The relationships between patient situational characteristics and surgical spinal outcomes
are well-documented. Patients receiving Worker’s Compensation had worse functional status
after surgical and non-surgical treatments for spine conditions (Anderson, Subach, & Riew,

8

2009; Atlas, Chang, Kamman, Keller, Deyo, & Singer, 2000; Burnham, et al, 1996; Voorhies,
Jiang & Thomas, 2007; Yang, Lowe, de la Harpe & Richardson, 2010). Unemployment status
has a negative impact on post-treatment outcomes for persons undergoing surgical or nonsurgical treatments and those working pre-operatively were ten times more likely to be working
post-operatively after lumbar fusion surgery (Anderson, Schwaegler, Cizek & Leverson, 2006;
Burnham et al., 1996; Silverplats et al., 2010; Zieger, et al., 2011).
The psychological characteristic of depression can contribute to the development of
chronic low back pain, can be a predictor of new pain episodes, and is negatively correlated with
outcome and return to work after surgery for lumbar herniated disc (Carragee, Alamin, Miller &
Carragee, 2005; Jarvik, Hollingworth, Heagerty, Haynor, Boyko, & Deyo, 2005; Kohlboeck et
al, 2004; Pincus, Burton, Vogel & Field, 2002; Trief, Grant & Fredrickson, 2000). Greater levels
of depression are associated with more functional disability in persons with chronic low back
pain (Feirerra & Pereira, 2013). Persons with low back pain have been found to have higher
rates of depression than those without low back pain (Bener et al., 2013). Not only are
depression and low back pain significantly correlated, but depression and anxiety are predictors
of greater low back pain intensity (Mok & Lee, 2008; Tetsunaga et al., 2013).
Several physiological characteristics have been shown to contribute to the development
of low back pain. Obesity is a risk factor for low back pain (Heuch, Hagen, Heuch, Nygaard &
Zwart, 2012; Shiri, Karppinen, Leino-Arjas, Solovieva & Viikari-Juntura, 2010; Shiri, et al.,
2008). Obesity was one of the factors found to increase the costs associated with lumbar
interbody fusion (LaCaille, DeBerard, LaCaille, Masters, & Colledge, 2007). Smokers have a
higher incidence of back pain than non-smokers (Shiri, Karppinen, Lein-Arjas, Solovieva, &
Viikari-Juntura, 2010).

9

The Effect of Symptoms on Outcomes
Symptoms are subjective phenomena that indicate a change in health or normal function
(Dodd, et al., 2000; Fu, LeMone, & McDaniel, 2004; Farrar, Berlin, & Strom, 2003; Fu,
McDaniel, & Rhodes, 2007). Symptom experience can influence outcome, and in general, the
worse the symptom experience, the more negative the impact on outcome.
Lumbar degenerative conditions can be a cause of the symptoms of low back pain and
lower limb pain, numbness, tingling and weakness. Low back pain and other symptoms related
to lumbar degenerative conditions can reduce physical function. The consequences of symptoms
in general include impact on adjustment to illness, quality of life, functional status, psychological
state, survival, and disease progression (Armstrong, 2003). The consistent finding is that the
worse the symptom experience, the poorer the outcomes, across many health conditions.
(Edward, et al., 2007; Hammer, Howell, Bytzer, Horowitz, & Talley, 2003; Wilson, Robinson, &
Turk, 2009).
Identification of subgroups of patients who experience symptoms with greater severity
may alert nurses to persons at risk for poorer outcomes (Miaskowski, et al. 2006). Nurses can
help patients identify and understand the cause for their symptoms, thereby leading to prompt
intervention and more effective coping through behavior interventions (Heidrich, Egan,
Hengudomsub, & Randolph, 2006). Identification of those priority symptoms that exert a
negative effect on other symptoms enable nurses to target intervention on the priority symptom,
thereby reducing the severity of the other symptoms and improving outcomes (Hoffman, von
Eye, Given, Given, & Rothert, 2009). Lumbar degenerative conditions are an expensive and
highly prevalent health condition which can lead to diminished physical function. Optimizing

10

physical function, in the context of lumbar degenerative conditions, is an important nursing
concern.
Knowledge Gap
While multiple patient factors have been found to independently influence physical
function outcomes for persons experiencing lumbar degenerative conditions, little is known
about the interaction of individual patient characteristics, genotype, and symptoms and their
impact on physical functioning in persons with lumbar spinal conditions receiving non-surgical
care. The gap in knowledge regarding symptoms and patient factors as predictors of physical
function in this population limits caregiver’s ability to tailor interventions designed to improve or
preserve physical function. Identifying those at risk for poor outcomes would allow for adjusting
treatment approaches to improve outcomes in this population.
Identification of the relationships between these factors could assist in the accurate
prediction of those patients at risk for sub-optimal outcomes from a non-surgical approach to
treatment for lumbar spinal conditions. Alternate care models could then be developed to
improve the outcomes of at-risk populations. The long term goal is to develop a predictive
model for outcome in persons with lumbar spinal conditions being treated non-surgically based
on patient characteristics, genetics and symptoms. Nurses are unique among all health
professionals in their holistic focus in diagnosing and treating human responses to health
conditions. There is no literature that examines patient factors (including genotype) and
symptoms and their effects on the outcome of physical function in adults experiencing lumbar
degenerative conditions. This study addresses a serious gap in knowledge regarding the multiple
factors that contribute to worse physical function outcomes, thus providing important data for
personalizing care for patients experiencing lumbar degenerative conditions.

11

A better understanding of the role of genes involved in the experience of pain and the
genes involved in disc degeneration may help identify those at risk for not only disc
degeneration, but also at risk for greater pain and disability. Exploration of genetic and patient
factors can identify individuals at risk for poorer outcomes from spinal interventions. Early,
tailored interventions to control pain and prevent chronicity could be implemented when these
risk factors are known. As scientists learn more about the links between the genes involved in
disc degeneration and environmental factors, nursing interventions can be developed for
populations at risk, to reduce pain and disability.
In summary, this study aims to add to nursing science by examining simultaneously the
physiological, situational and psychological individual characteristics that affect physical
function for persons experiencing lumbar spinal degenerative conditions. Moreover, the
incorporation of symptoms in combination with individual characteristics and their influence on
physical function brings a uniquely nursing perspective to a condition that affects 80% of
persons in their lifetime. Last, the incorporation of genotyping as a relevant physiological
characteristic in this study is innovative and may lead to further insight into personalizing care
for persons experiencing lumbar degenerative conditions.

12

CHAPTER II
Conceptual Framework
The framework organizing the approach to this inquiry is The Theory of Unpleasant
Symptoms, (Lenz, Gift, Pugh, & Milligan, 1995; Lenz, Pugh, Milligan, Gift, & Suppe, 1997).
The Theory of Unpleasant Symptoms (TOUS) is a multi-dimensional, dynamic, middle-range
theory that is unique in its consideration of multiple symptoms occurring simultaneously that
catalyze each other. The TOUS is a middle-range theory and is therefore more specific than a
grand theory. Middle-range theories are less abstract and are focused more on specific
phenomena (Fawcett, 2005). Middle-range theories are more directly useable for nursing
practice application (Peterson & Bredow, 2009; Smith & Liehr, 2008).
The TOUS was developed after nursing clinicians, separately working on the symptoms
of dyspnea and fatigue, recognized similarities between their conceptualizations regarding the
context in which these symptoms occurred and the effect these symptoms had on performance
(Gift, 2009). Knowing that there were similar activities focused on the symptom of pain, they
set out to craft one model that could guide the understanding and management for many
symptoms. In the first iteration of the model, three categories of factors were believed to
influence the predisposition to or manifestation of an unpleasant symptom (Lenz, Suppe, Gift,
Pugh, & Milligan, 1995). These categories were: physiological, situational and psychological.
These factors were specifically conceptualized to begin to identify interventions to ameliorate or
reduce the impact of fatigue. The authors believed that by identifying the factors that contributed
to the symptoms, interventions aimed at modifying these factors would reduce the symptoms
(Lenz, Suppe, Gift, Pugh & Milligan, 1995).

Symptoms were conceptualized as having

variable duration, intensity, quality, and distress (Lenz, Suppe, Gift, Pugh, & Milligan, 1995).
13

Symptoms influence performance, which includes functional status, cognitive functioning, and
physical performance (Lenz, Suppe, Gift, Pugh, & Milligan, 1995). The original model
presented a linear depiction of the variables.
Work continued on the model over the next two years, and in 1997, the authors published
their updated Theory of Unpleasant Symptoms (Lenz, Pugh, Milligan, Gift, & Suppe, 1997).
The new model went from a linear, unidirectional depiction of variables to a sophisticated,
interactive, dynamic feedback loop incorporating antecedents, (or influencing factors), the
symptoms themselves (with recognition that many symptoms can be experienced at once, and
that they interact with and catalyze one another), and the outcome, performance (which in turn,
affects how symptoms are experienced and the influencing factors). The propositions of the
TOUS describe how each concept relates to the others. The antecedent factors may interact
together, antecedent factors interact in their influence on symptoms, symptoms may influence the
effect antecedent factors have on performance, antecedent factors and symptoms together
influence cognitive and physical performance and performance can have reciprocal effects on
symptoms and antecedent factors (Lenz, Pugh, Milligan, Gift & Suppe, 1997). The outcome of
performance includes both functional and cognitive features (Lenz, Pugh, Milligan, Gift &
Suppe, 1997). Physical function is the performance outcome of interest in this study. (See Figure
1 for the Theory of Unpleasant Symptoms with Study Variables).

14

Antecedent factors are described in the TOUS update (1997). Physiological factors
include normally functioning body systems, the presence of trauma, or the existence of
pathology (Lenz, Pugh, Milligan, Gift & Suppe, 1997). Psychological factors include mental
state or mood, affective reaction to illness, and uncertainty about the symptoms and their
meaning (Lenz, Pugh, Milligan, Gift & Suppe, 1997). Situational factors include marital and
employment status, access to health care, diet, exercise and social support (Lenz, Pugh, Milligan,
Gift & Suppe, 1997).
The defining attributes of physical function are fairly explicit in this model. Physical
function, or “functional performance”, includes physical activity, activities of daily living, social
activities, work, and “other role related tasks” (Lenz, Pugh, Milligan, Gift, & Suppe, 1997).
Greater or more severe symptoms can reduce functional performance, role performance, and
“physical performance capabilities” (Lenz, Pugh, Milligan, Gift, & Suppe, 1997). Decreased
levels of performance in this dynamic, reciprocal model can affect symptoms and the
physiologic, psychological, and situational antecedent factors (Lenz, Pugh, Milligan, Gift, &
Suppe, 1997).
The TOUS is circumscribed and limited in scope and addresses the phenomena of
symptoms, how they influence one another, how symptoms are influenced by antecedent factors,
and how these phenomena influence performance. Each concept in the TOUS interacts together
in a continuous feedback loop. The authors claim that the TOUS is parsimonious for proposing
that the same antecedent factors could influence many symptoms and that a single intervention
has the potential for alleviating more than one symptom (Lenz, Pugh, Milligan, Gift & Suppe,
1997). The concepts and propositions in the TOUS are stated concisely. Even though there are

15

multiple relationships between the concepts of the model, they are portrayed in an economical
way.
The nursing metaparadigm concepts addressed by the TOUS include: an aspect of health
(cognitive and physical function), human beings (their symptoms and the physiological and
psychological features they possess), and their environment (the situational factors that influence
their symptoms and function). Lenz, Suppe, Gift, Pugh and Milligan (1995) were clear that the
early focus on antecedent factors and their influence on symptoms were for the purpose of
identifying interventions. Interventions could then be developed to modify the antecedent
factors found to influence symptoms (Lenz, Suppe, Gift, Pugh & Milligan, 1995). Implicit in the
model is that the goal of nursing is to enhance function. In the TOUS update, Lenz, Pugh,
Milligan, Gift and Suppe (1997) describe how interventions can be individualized by using the
antecedent factors and patterns of symptoms unique to the individual. The authors do state that
by controlling one symptom, the effect of many symptoms and function may be enhanced (Lenz,
Pugh, Milligan, Gift & Suppe, 1997).
Symptoms are the central focus of the model. Symptoms can occur together because of a
single event, such as surgery, or one symptom can precede another. Although symptoms may be
different, most symptoms share the dimensions of intensity, quality, duration and distress (Lenz,
Pugh, Milligan, Gift & Suppe, 1997). Intensity refers to the amount, strength or severity of a
symptom. Quality is the way in which a symptom is manifested, and is reflected in the words
used by the individual to describe its nature. Symptom quality also includes the location of the
symptom. Quality aspects are felt to be specific to a given symptom, and this symptom feature
may be difficult for individuals because ability to recognize and describe a symptom may vary
(Lenz, Pugh, Milligan, Gift & Suppe, 1997). Symptom duration provides a time element to the

16

symptom experience and includes the frequency, timing and length of the symptom. Distress
reflects the degree to which an individual is bothered by a symptom. How much an individual is
bothered by a symptom can determine help-seeking. Individuals vary in their estimations of how
bothered they are by the same symptom. The symptom distress dimension contributes most to
quality of life (Lenz, Pugh, Milligan, Gift & Suppe, 1997). Symptoms are conceptualized to
catalyze each other, with the effect of greater impact of symptoms on function. Lenz, Pugh,
Milligan, Gift and Suppe (1997) assert that symptoms occurring simultaneously have a
multiplicative, rather than an additive effect on each other. Interventions to manage symptoms
should be based on the dimensions of the symptom.
While the specific activities of nurses are not explicitly portrayed in the TOUS, the
authors do state that the purpose for development of the model was to help nurses identify
individualized interventions through delineation of the antecedent factors and their effects on the
symptoms experienced by the patient (Lenz, Suppe, Gift, Pugh & Milligan, 1995; Lenz, Pugh,
Milligan, Gift & Suppe, 1997). The TOUS provides for a method to discern the dimensions of
symptoms and identify the antecedent factors that contribute to them across a range of clinical
conditions. This allows the nurse to tailor interventions appropriate to the situation. However,
Brant, Beck and Miaskowski (2010), in their comparison of middle-range theories addressing
symptoms, contend that there is no consideration for intervention in the TOUS, or for resolution
of a symptom.
The special contribution this middle-range theory makes is its recognition that most
individuals experience more than one symptom, and that the experience of symptoms may be
multiplicative, rather than additive (Lenz, Pugh, Milligan, Gift & Suppe, 1997). The
conceptualization of multiple symptoms occurring simultaneously represents the reality of

17

clinical care. The TOUS has even been used outside the discipline of nursing (Motl & McAuley,
2009).
The TOUS has been criticized for lack of clarity of what constitutes physiological,
situational and psychological antecedent factors (Brant, Beck & Miaskowski, 2010). A lack of
clear differentiation between antecedent factors and symptoms in the TOUS has also been noted.
In a qualitative study using the TOUS in a population of patients and care-givers with Alzheimer
Disease (AD), the authors found utility and fit in the model’s antecedent factors, multiple
simultaneous symptom experience, interaction between symptoms, interaction between
antecedent factors and symptoms and interaction between antecedent factors in AD (Hutchinson
& Wilson, 1998). However, they noted blurred boundaries and overlap between antecedent
factors and symptoms—there was lack of clarity regarding whether study variables like anxiety
and depression were psychological antecedent factors or symptoms. However, in the first
iteration of the model, the authors explicitly state that depression and fatigue are conceptualized
to be psychological antecedent factors (Lenz, Suppe, Gift, Pugh and Milligan, 1995).
Hutchinson and Wilson (1998) concluded that the TOUS was useful in describing and assessing
the complexity and relationships between multiple antecedent factors, symptoms and
performance outcomes in AD. In fact, most studies utilizing the TOUS conceptualize depression
as an antecedent factor (Corwin, Klein & Rickelman, 2002; Liu 2006; Redeker, Lev & Ruggiero,
2000; Rychnovsky, 2007; So et al., 2012). Only one study conceptualized depression as a
symptom (Motl & McAuley, 2009).
The model has been found to be useful in demonstrating that multiple symptoms
occurring together affects outcome (Gift, Jablonski, Stommel & Given 2004; Liu, 2006; Motl &
McCauley 2000; Myers 2009). Multiple antecedent factors have also been shown to affect the

18

experience of one symptom (Corwin, Klein & Rickelman 2002; Woods, Kozachik & Hall,
2010). Antecedent factors have also been shown to affect symptoms, which in turn, affects
function (Hoffman, von Eye, Gift, Given & Given, 2009). However, not all studies provide
support for the influence of antecedent factors on symptoms and the combined effect on function
(Redeker, Lev & Ruggiero, 2000).
The TOUS is a testable model. Many nursing and some non-nursing studies have tested
the propositions of the TOUS. Though not all of the propositions of the TOUS have been
supported, many studies have explored the influence of antecedent factors on symptoms, the
influence of symptoms on function and multiple symptoms occurring together influencing
outcome. In the 1997 update, Lenz, Pugh, Milligan, Gift and Suppe offer examples of
instruments that capture symptom dimensions. The McGill Pain Questionnaire and the Fatigue
Symptom Checklist are provided as examples of instruments used to measure symptom quality.
A Visual Pain Analog can measure symptom intensity. Both quantitative and qualitative
methods should be considered in the measurement of symptoms, and the authors recommend
“multidimensional, multifactorial measurement procedures” (Lenz, Pugh, Milligan, Gift &
Suppe, 1997).
Cognitive and physical performance is not the only outcomes examined using the TOUS.
Some authors have inserted quality of life, self-efficacy and depression as the outcome. Some
authors have placed self-efficacy as a mediator between symptoms and outcome, but have not
defined self-efficacy as a psychological antecedent factor.
In an analysis of the usefulness of the TOUS to guide the evolving understanding of
symptom burden in irritable bowel syndrome, the authors concluded that the TOUS had utility to
guide symptom research in this disease (Farrell & Savage, 2009). Myers (2009) compared the

19

TOUS and the Conceptual Model of Chemotherapy-Related Changes in Cognitive Function for
guiding research, and found the TOUS to be advantageous for its inclusion of multiple cooccurring symptoms and its interrelationships between antecedent factors and symptoms. The
TOUS was felt to be a useful way for nurses to place symptoms in the context of antecedent
factors that influence them in the study of patients undergoing bariatric surgery (Tyler & Pugh,
2009).
The TOUS has been used widely to guide research across a number of populations,
although no studies were identified involving patients with spinal conditions that utilized the
TOUS as an organizing framework. There were several studies using the framework in
populations experiencing fatigue and cancer (Corwin, Brownstead, Barton, Heckard & Morin,
2005; Corwin, Klein & Rickelman, 2002; Gift, Jablonski, Stommel & Given, 2004; Hoffman,
von Eye, Gift, Given & Given, 2009; Liu, 2006; Motl & McAuley, 2009; Redeker, Lev &
Ruggiero, 2000; Reishtein, 2004; Rychnovsky, 2007). Most studies have provided support for
the propositions of the TOUS.
There is evidence for psychological antecedent factors influencing symptoms. Corwin,
Brownstead, Barton, Heckard and Morin (2005) used post-partum depression as the outcome in a
study to explore the factors predictive of this condition. Those women who were fatigued at
post-partum day 14 also scored as significantly depressed at post-partum day 28.
Some evidence has been provided for the influence of antecedent factors on symptoms,
which in turn, affects function. Using a secondary analysis of baseline data from two
randomized controlled trials involving individuals undergoing chemotherapy for cancer, the
authors set out to test the hypothesis that physical functional status can be predicted through
patient factors, cancer-related fatigue, “other” symptoms, and perceived self-efficacy for fatigue

20

self-management in individuals with cancer (Hoffman, von Eye, Gift, Given & Given, 2009).
Fatigue was the most severe and prevalent symptom, and was correlated with cancer-related
fatigue severity. Younger age, female sex, and greater number of co-morbid conditions
predicted greater cancer-related fatigue severity. Greater cancer-related fatigue severity
predicted greater symptom severity, but the reverse was not demonstrated. Greater cancerrelated fatigue severity predicted lower perceived self-efficacy for fatigue self-management, and
greater perceived self-efficacy for fatigue self-management predicted greater physical functional
status.
However, not all studies have confirmed the influencing relationship between antecedent
factors on symptoms affecting function. Redeker, Lev and Ruggiero (2000) examined the
symptoms of insomnia and fatigue and the psychological factors of depression and anxiety to
determine their contribution to quality of life in a Chinese population undergoing chemotherapy.
They found that depression had the greatest effect on quality of life, with the symptoms of
insomnia and fatigue only accounting for 4% of the variance in quality of life. While they were
able to demonstrate that depression explained most of the variance in quality of life, they were
unable to demonstrate that psychological factors catalyzed symptoms to affect quality of life. In
their critique of the utility of the TOUS, they questioned how to account for changing
psychological factors like anxiety and depression, and called for more clarity regarding
psychological factors (Redeker, Lev & Ruggiero, 2000). In a response following Redeker, Lev
& Ruggiero’s published study, Pugh, Milligan and Lenz (2000) acknowledged that different
concepts could potentially fit as antecedent factors or symptoms, depending on the phenomenon
under study. However, they contended that the study by Redeker, Lev and Ruggiero (2000) was

21

not a true test of the model, because it was used simply to correlate relationships proposed from a
secondary analysis of data (Pugh, Milligan & Lenz, 2000).
Another study was unable to make the connection between antecedent factors, symptoms,
and outcome (quality of life). Consistent with the purpose of the TOUS, hospitalized heart
failure patients were studied to identify symptom clusters and factors contributing to the
experience of these symptoms (Jurgens, et al., 2009). Using quality of life as an outcome
measure, the authors discovered three symptom clusters explaining much of the variance in
quality of life in hospitalized heart failure patients. Shortness of breath, fatigue and sleep
problems as a cluster explained 46% of the variance in quality of life; depression, memory
problems and worry as a cluster explained 13% of the variance. The symptom cluster of
swelling, need to rest and dyspnea explained an additional nine percent of the variance in quality
of life (Jurgens, et al. 2009). They were unable to demonstrate that the factors of co-morbid
disease and age catalyzed the impact of these symptom clusters to affect quality of life.
There has been support for the influence of symptoms on physical function. Motl and
McAuley (2009) studied patients with Multiple Sclerosis and the temporal relationship between
symptoms and physical activity behavior six months later. They were able to demonstrate a
predominant symptom cluster of fatigue, depression and pain, which had a strong and negative
effect on physical activity behavior measured by accelerometry. They also explored whether the
symptom cluster had a direct effect on physical activity behavior, or whether this effect was
mediated by self-efficacy. They found that self-efficacy did not mediate this relationship.
Curiously, as they were testing model fit, they conceptualized functional limitation as a mediator
between the symptom cluster identified and physical activity behavior, instead of as an outcome,

22

as suggested by the TOUS. Functional limitation was found to be a significant mediator between
the symptom cluster and physical activity behavior (Motl & McAuley, 2009).
While the TOUS has been used in a variety of clinical conditions, by far, the most studied
phenomenon (both symptom and outcome) using the TOUS is fatigue. The psychological,
physiological and situational antecedent factors of depression, breast feeding and disturbed sleep,
respectively, all affected post-partum fatigue in military women (Rychnovsky, 2007). Postpartum fatigue was highly associated with the symptom of post-partum depression (Corwin,
Brownstead, Barton, Heckard & Morin, 2005). Physiological factors of cigarette smoking and
younger age, but not biological markers, (blood pressure, BMI, immune or inflammatory indices)
were found to be correlated with fatigue (Corwin, Klein & Rickelman, 2002). Fatigue was a
significant symptom in studies of cancer patients, patients with Chronic Obstructive Pulmonary
Disease (COPD) and hemodialysis patients (Gift, Jablonski, Stommel & Given, 2004; Liu, 2006;
Reishtein, 2004).
Few studies measured biomarkers as indicators of physiological antecedent factors. In
their exploration of predictors of fatigue in healthy young adults, Corwin, Klein and Rickelman
(2002) hypothesized that among other situational and psychological factors, physiologic
antecedent factors including serum cotinine levels, (a metabolite of nicotine) and c-reactive
protein and tumor-necrosis-alpha (inflammatory markers) would influence fatigue in a well
population. While these biomarkers were not found to be significant predictors of fatigue in this
population, the most important predictor was cigarette smoking. Corwin, Brownstead, Barton,
Heckard and Morin (2005) included serum cortisol level, (a marker of stress), as a physiological
predictor contributing to postpartum depression. While self-report of stress and fatigue were
correlated with post-partum depression, serum cortisol was not. Moreover, serum cortisol levels

23

were not correlated with perceived stress (Corwin, Brownstead, Barton, Heckard & Morin 2005).
McCann and Boore (2000) hypothesized that among other physiological factors, hemoglobin,
hematocrit, ferritin, urea, creatinine, albumin, phosphate and calcium levels were associated with
the symptom of fatigue. While they found a relationship between depression and fatigue, they
were unable to demonstrate an association between the biological markers and fatigue (McCann
& Boore, 2000).
Corwin, Klein and Rickelman (2002) introduced the concept of fixed and unfixed
antecedent factors, a conceptual approach also included in Corwin, Brownstead, Barton, Heckard
and Morin (2005). Fixed antecedent factors are those that cannot be changed, such as gender,
age, family or personal history of depression and post-partum status. Because of its association
with iron deficiency anemia, thyroid hormone deficiency and post-partum inflammatory status,
fatigue was considered an unfixed physiologic factor in the post-partum depression study
(Corwin, Brownstead, Barton, Heckard & Morin, 2005). BMI, resting blood pressure,
inflammatory and immune status were considered unfixed physiologic factors in the study
exploring predictors of fatigue in a well population (Corwin, Klein & Rickelman, 2002).
More studies utilizing the TOUS as an organizing framework and using performance as
an outcome focused on the physical function aspect. For example, one study focused on the
cognitive outcome of attentional function in women with breast cancer (Lee, 2005). Mood
disturbance and symptoms each were associated with attentional function, and while symptoms
were not found to mediate the relationship between mood disturbance and attentional function,
symptoms did mediate the relationship between mood disturbance and attentional function when
symptoms were rated at a medium level (not low or high) (Lee, 2005). Finally, Parks, Lenz,
Milligan and Han (1999) introduced the notion that the impact of symptoms on performance

24

could actually extend outside the focal individual to affect others. They were able to
demonstrate that infant development was higher when mothers were not persistently fatigued.
As expected, because of the concepts and relationships proposed in the TOUS, the
nursing studies using the TOUS as a framework were focused on the nature of the relationships
between antecedent factors, symptoms and outcomes. There were no studies testing nursing
interventions using the TOUS. It is likely that the nature of these relationships in the clinical
conditions studied has not been sufficiently explained yet to determine appropriate nursing
interventions.
All but one study using the TOUS as the organizing framework were authored by nurses.
However, the study examining the ability to predict future physical activity in patients with
Multiple Sclerosis using symptoms was authored by kinesiologists (Motl & McAuley, 2009).
In summary, the TOUS has been used extensively to study the influence of symptoms
and antecedent factors on outcome. Both physical performance and cognitive performance
outcomes have been studied, as well as quality of life. Most propositions of the TOUS have
been supported by research, with the most conflicting findings regarding the effect of antecedent
factors on symptoms, which in turn affects function.
The TOUS has been modified several ways, sometimes limiting the focus to the impact
of antecedent factors on symptoms. The TOUS has been used in a variety of clinical settings and
conditions, with fatigue being the most studied symptom. While there is ambiguity regarding
overlap of symptoms and antecedent factors, the theory is flexible and can be used in a variety of
clinical situations. There is no specific inclusion of nursing intervention in the TOUS, although
the implication is that identification of the salient antecedent factors and symptoms that affect
outcome will lead to interventions designed to improve function (Lenz, Suppe, Gift, Pugh &

25

Milligan 1995). Interventions can target antecedent factors and symptoms. Since symptoms are
conceptualized to be multiplicative, and all of the categories of concepts in the model
(antecedent factors, symptoms and outcome) are proposed to be interactive, one intervention has
the potential to affect more than one component of the model (Lenz, Pugh, Milligan, Gift &
Suppe, 1997).
Biologic indicators of physiologic antecedent factors have been included in a few studies,
and have not proved to have associations with the outcomes being studied. Biologic markers as
an indicator of physiological antecedent factor deserve further study.
Use of the TOUS in This Study
The TOUS was selected as the framework for the current study because of the multiple
antecedent factors found to contribute to the physical function outcomes for patients
experiencing lumbar spinal degenerative problems (See Figure 1). The TOUS accurately depicts
the reality that back pain and lumbar degenerative conditions are likely heterogeneous clinical
conditions, the result of genetic, physiologic, behavioral and situational influences (Fourney, et
al., 2011). The TOUS also allows for accurate depiction of the multiple symptoms experienced
by individuals with lumbar degenerative conditions, and the many antecedent factors that likely
contribute to the outcome of physical function in this population.
For the purposes of this research, physiological factors are defined as biologic and
physical features possessed by the individual. Physiological factors influencing the symptom
experience in this population are conceptualized to include genotype, body mass index (BMI),
sex, age, and smoking. Situational factors are defined as features that are outside the individual
that may influence health status and are conceptualized to include employment status, worker’s
compensation claim, and insurance type (commercial, Medicaid, Medicare, Tricare or none).

26

Psychological factors are defined as mental state or mood and are conceptualized to include
depression. Physiological, situational and psychological factors are conceptualized to influence
symptoms and physical function.
The physiological factors of BMI, sex, age and smoking have all been found to have
independent and varying effects on back pain and physical function. Genotype is now associated
with both the symptom of low back pain and lumbar disc degeneration. The psychological factor
depression can influence the symptom of low back pain and physical function in individuals with
lumbar degenerative conditions. The situational factors of litigation and worker’s compensation
influence physical function in individuals with lumbar degenerative conditions, and insurance
type has been found to influence health outcomes in general. Finally, symptom research in this
population is lacking and deserves further study.
Symptoms are defined as a perception of change in normal functioning in individuals
(Lenz, Pugh, Milligan, Gift & Suppe, 1997). For the purposes of this study, symptoms will
include the presence of back and/or leg pain, pain intensity (measured on a 10 cm visual analog
scale) and associated symptoms of leg numbness and weakness. Symptom duration and distress
will not be explored, but quality (numbness) of the sensory symptom will be included.
Physical function is the primary outcome variable for this inquiry. Physical function is
defined as an individual’s ability to perform in the various physical roles in their lives, to
accomplish ADLS, to work, to carry out daily tasks for self and significant others, to be mobile
and maintain leisure physical activities. Instruments used to measure physical function include
the Oswestry Disability Index (ODI) and the physical functioning subscale of the SF-36.
For Aim 1, the contributions of patient physiological, situational and psychological
factors to symptoms and to physical function will be explored, in order to demonstrate that

27

physical function in the population of persons with lumbar degenerative conditions is the result
of a constellation of factors. For Aim 2, the combination of patient factors and symptoms will be
explored to determine if profiles of specific variable combinations predict persons at greater risk
for poorer physical function outcomes. For Aim 3, biologic data will be used to determine if
genotype influences symptoms and physical function, or whether genotype, with other patient
factors, can contribute to the ability to predict persons at risk for poorer physical function.
Although patient factors are theorized to influence each other in the TOUS, these relationships
are beyond the scope of this study.
In summary, the TOUS has been useful to guide inquiry into the influence of patient
characteristics and symptoms in different clinical conditions. Several patient characteristics have
been shown to influence many different outcomes for persons experiencing lumbar degenerative
conditions. A few studies have explored the impact of back and/or leg pain on physical function
in persons experiencing lumbar degenerative conditions. However, studies are lacking that
explore the influence of patient characteristics and symptoms on the outcome of physical
function in this clinical condition. In Chapter 3, Review of the Literature, each patient
characteristic and symptom under study will be reviewed for their effects on physical function
and other outcomes for persons experiencing lumbar degenerative conditions.

28

CHAPTER III
Review of the Literature
The review of literature will first address relevant lumbar spinal anatomy and the
pathophysiological processes associated with degenerative changes that can lead to the
symptoms of low back pain and leg pain and numbness. There are studies that explore the
relationship between obesity, sex, smoking, OPRM1 and COMT genotypes and physical
function, and these will be reviewed in this chapter. While there are no studies that explore the
relationship between the genes implicated in the structural integrity of the disc and physical
function, it is known that disc degeneration can contribute to the development of the symptom of
low back pain. This study will include all persons with lumbar degenerative conditions, in order
to maintain the focus on symptoms and patient characteristics that may be common to all. In
reality, many different lumbar degeneration diagnostic categories co-exist in the same individual.
The physiological, situational and psychological antecedent factors that have been shown
to influence the symptoms and physical function will be reviewed. And, while many different
genes have been identified to contribute to lumbar disc degeneration, only a few disc structural
genes were included in this study. Genes involved in the degrading process of the disc have been
identified, but these were not included in this study. Many genes have been implicated in the
experience of pain. Only those encoding for COMT and OPRM-1 are included in this study.
Finally, the symptoms commonly experienced by individuals with lumbar degenerative
conditions will be reviewed, along with the available literature regarding the effects of
antecedent factors and symptoms on physical function.

29

Lumbar Spinal Anatomy and Degenerative Changes
The lumbar disc is situated between the vertebrae, and consists of a gelatinous inner core
called the nucleus pulposis, encased by concentric layers of diagonally oriented collagen fibers
called the annulus fibrosis. The nucleus contains proteoglycan molecules that hold water. The
nucleus functions to absorb and accommodate compression loads. The annulus consists of type I
and II collagen fibers, with cross-links of type IX collagen. The annulus holds the nucleus in
place and attaches the disc to the vertebral bodies (Smith & Fazzalari, 2006).
Degenerative disc changes progress over time (Williams, et al. 2011). Ideally, there is a
balance between synthesis and degradation of the constituents of the disc. Over time, however,
the cells capable of synthesizing proteogylcans diminish in number, causing the water content of
the nucleus to decline (Hadjipavlou, et al., 2008). This, in turn, causes the height of the disc to
diminish. Cytokines, normally in balance with disc regeneration factors, gradually increase,
contributing to degeneration (Hadjipavlou, et al., 2008). The disc structures become more
disorganized, and the ability of the disc to resist normal forces is diminished. Conditions
involving the intervertebral disc have been identified as a cause of low back pain (Cheung,
Samartzis, Karppinen & Luk, 2012; Freemont, 2009; Livshits, et al, 2011; Takatalo et al., 2011).
As degeneration progresses, the combination of facet joint arthritis and disc height reduction can
produce changes that decrease the diameter of the canal and neuroforamen, which can contribute
to spinal nerve symptoms of limb pain, numbness, tingling and weakness (Genevay & Atlas,
2010).
Lumbar disc degeneration and its accompanying symptoms are a multi-factorial health
condition, likely resulting from both genetic and environmental factors. Evidence supporting the

30

key physiological, situational, and psychological variables to be examined in this study are
summarized according to the TOUS model.
Physiological Factors
Several physiological variables influence symptoms associated with lumbar disc
degeneration. There is some evidence linking these physiological variables with physical
function in persons with lumbar degeneration. The physiological variables to be examined in
this study include body mass index (BMI), sex, age, smoking status and genotype, each
discussed in detail here.
Obesity
Obesity is a patient characteristic that is a strong risk factor for low back pain (Heuch,
Hagen, Heuch, Nygaard & Zwart, 2010; Heuch, Heuch, Hagen & Zwart, 2012; Shiri, Karppinen,
Leino-Arjas, Solovieva & Viikari-Juntura, 2010; Shiri, et al., 2008). BMI greater than 30 is a
risk factor for the development of low back pain in persons without baseline low back pain, even
when adjusted for age, work status, education, physical activity and smoking (Heuch, Heuch,
Hagen & Zwart, 2013).
Persons with overweight or obese BMI values are more likely to have disc degeneration
and more likely to have greater severity of disc degeneration at more levels (Samartzis,
Karppinen, Chan, Luk & Cheung, 2012). Being overweight at any age increases the risk of
lumbar disc degeneration, but persons who are overweight at an earlier age have a greater risk of
lumbar disc degeneration (Liuke, et al., 2005). Takatalo et al. (2013) demonstrated that higher
adiposity measures, including waist circumference and body fat percentage were associated with
lumbar disc degeneration in males, but not in females. In a study of Japanese persons over the
age of 50, the odds ratio of having lumbar disc degeneration was greater at nearly every lumbar

31

level for those with BMI greater than or equal to 25 (Hangai et al., 2008). Specifically, the odds
ratio was 2.98 (95% CI 1.52-6.05), 3.58 (95% CI 1.85-7.21), 2.32 (95% CI 1.18-4.72), and 3.34
(95% CI 1.70-6.81) for L2-3, L3-4, L4-5 and L5-S1 levels, respectively (Hangai et al, 2008). In
obese individuals, physical function outcomes have been worse for operative and non-operative
treatment for lumbar disc herniation (Rihn, et al., 2013).
In summary, in persons who are obese, there is a higher risk of disc degeneration, one
cause of low back pain. Moreover, in those obese at a younger age, there is a greater risk of disc
degeneration at more levels. Obesity is also directly associated with low back pain.
Sex and Age
While the experience of pain varies between females and males, it is not clear whether
lumbar disc degeneration differs in rate and severity between females and males. One systematic
review suggested that the rate of progression of lumbar disc degeneration was greater in females
ages 50-59; with disc degeneration in males progressing faster during ages 60-79 (Lee, Dettori,
Standaert, Brodt & Chapman, 2012). However, a cadaveric study failed to show any difference
in disc degeneration rates between females and males (Siemionow, An, Masuda, Andersson &
Cs-Szabo, 2011).
Females may experience greater pain levels and worse physical function than males with
lumbar stenosis. Kim et al. (2013) identified significantly worse pain VAS and ODI scores for
women than for men, even after controlling for BMI, age, and severity of disc degeneration and
stenosis.
The prevalence of disc degeneration does increase with age (Cheung, et al., 2009). The
prevalence of disc degeneration is estimated to be as high as 40% for individuals under the age

32

of 30 (Cheung, et al., 2009). By age 50, the prevalence rises to 60-90% (Cheung, et al, 2009;
Kalichman, Kim, Li, Guermazi & Hunter, 2010).
In summary, it is not clear whether the rate and prevalence of disc degeneration varies by
sex. Sex differences in estimations of pain in lumbar stenosis have been identified. Disc
degeneration increases with age.
Smoking
Smokers have a higher incidence of back pain than non-smokers (Karahan, Kav,
Abbasoglu & Dogan, 2009; Shiri, Karppinen, Lein-Arjas, Solovieva, & Viikari-Juntura, 2010).
Current smokers had the highest risk of low back pain, compared with former and never smokers
in one meta-analysis (Shiri, Karppinen, Leino-Arjas, Solovieva & Viikari-Juntura, 2010).
Specifically, the odds ratio for low back pain in current smokers in the past month was 1.30
(95% CI 1.16-1.45), for low back pain in the past 12 months, 1.33 (95% CI 1.26-1.41), for
seeking care for low back pain,1.49 (95% CI 1.38-1.60, for chronic low back pain, 1.79 (95% CI
1.27-2.50), and for disabling low back pain, 2.14 (95% CI 1.11-4.13) (Shiri, Karppinen, LeinoArjas, Solovieva, S. & Viikari-Juntura, 2010). Data from the Nurses’ Health Study reveal that
current smokers have a higher risk of lumbar disc herniation than former and never smokers, and
the risk increases with number of cigarettes smoked per day (Jhawar, Fuchs, Colditz & Stampfer,
2006). Among patients presenting for treatment for spine complaints, current smokers had
highest baseline Oswestry Disability Index (ODI) scores (a lumbar disease-specific instrument to
measure function), followed by former then never smokers (44.22, 38.11. 36.02, respectively)
(Prasarn, Horodyski, Behrend, Wright & Rechtine, 2012). Current smokers in treatment for
spine-related pain had higher pain visual analog scores (VAS) than nonsmokers, and those

33

smokers who quit during treatment experienced greater improvement in VAS pain scores
(Behrend, Prasarn, Coyne, Horodyski, Wright & Rechtine, 2012).
Along with aortic calcification and stenosis of the lumbar arteries, high cholesterol levels
and smoking were associated with low back pain and lumbar disc degeneration in a systematic
review (Kauppila, 2009). The relative risk for smokers compared to non-smokers to be
hospitalized for a lumbar disc degeneration-related cause in a Swedish prospective cohort study
was 1.27 (95% CI 1.15-1.39) (Wahlstrom, Burstrom, Nilsson & Jarvholm, 2012). Smokers had
18% greater mean lumbar disc degeneration scores than non-smokers (Battie et al., 1991).
Genotype
Lumbar disc degeneration has traditionally been considered to be the result of age, sex,
occupation, smoking and repetitive vibration. More recently, however, hereditary and biological
mechanisms contributing to disc degeneration have been identified (Zhang, Sun, Liu & Guo,
2008). Scientists have discovered several genes that contribute to disc degeneration and to disc
structural integrity. Lumbar disc degeneration is now considered to be a complex process with
both genetic and environmental contributors (Hadjipavlou, Tzermiadianos, Bogduk & Zindrick,
2008).
Selected Candidate Genes for Disc Structure
Genes that are associated with the structural components of the disc that help to maintain
its integrity include those that code for collagen (COL9A2 and COL9A3, with others), aggrecan
(ACAN), and vitamin D receptors (VDR) (Hadjipavlou, et al., 2008; Kao, Chan, Samartzis, Sham
& Song, 2011).

34

Collagen IX Alpha 2 and Alpha 3 (COL9A2 and COL9A3) Genes
The intervertebral disc contains an extracellular matrix of proteoglycans and collagen.
The inner portion of the disc, the nucleus pulposis, consists mainly of proteoglycans (about 50%)
with about 20% collagen II (Annunen, et al., 1999).

Both the proteoglycan and the collagen II

contain small amounts of collagen IX. Collagen IX contains three genetically distinct chains,
alpha 1, alpha 2, and alpha 3 (Diab, Wu & Eyre, 1996). Collagen IX is believed to function as a
link between collagens and non-collagenous proteins in tissues (Annunen, et al., 1999). These
cross links are believed to play an important role in protecting the disc from distension from
aggrecan and water by their interconnected network, thereby absorbing and distributing loads
(Aladin, et al., 2007; Diab, Wu & Eyre, 1996). The Collagen Type IX, Alpha-2 (COL 9A2) gene
codes for the alpha 2 chain and the collagen Type IX, Alpha-3 (COL9A3) codes for the alpha 3
chain (Kalichman & Hunter, 2008).
Polymorphisms in the COL9A2 and COL9A3 genes (location 1p34.2 and 20q13.33,
respectively) have been implicated in changes of the type IX collagen that make it more unstable,
making the disc more susceptible to mechanical stress (Hadjipavlou, Tzermiadianos, Bogduk &
Zindrick, 2008). An arginine (wild-type) to tryptophan (Trp3) change in the COL9A3 gene has
been associated with a higher risk of disc degeneration among obese individuals (Solovieva, et
al., 2002). A glutamine to tryptophan (Trp2) change in the COL9A2 gene has been associated
with disc degeneration (Annunen, et al., 1999; Jim, et al., 2005; Rathod, et al., 2012).
Results from a Japanese study of 84 patients who underwent surgery for herniated disc
seemed to suggest that those possessing the Trp2 allele were more likely to have developed disc
degeneration at an earlier age (Higashino et al, 2007). The Trp3 allele was not identified in this
Japanese population. Although there was no statistically significant association between the

35

Trp2 allele and disc degeneration for those over 40 years of age, the authors found that for those
under the age of 40, there was a six-fold greater chance of having disc degeneration for those
possessing the Trp2 allele (Higashino et al., 2007). Conflicting findings in a large (N=470 cases
and 658 controls) Japanese population were published by Seki et al. (2006). The Trp2 allele was
actually under-represented in those with lumbar intervertebral disc degeneration. Instead, the
authors found that a specific haplotype was over-represented in those with lumbar intervertebral
disc degeneration (Seki, et al., 2006).
The finding of polymorphisms of the COL9A2 and COL9A3 genes influencing the
development of disc degeneration is not consistent among different ethnic groups, however.
This is due in part to the frequency with which the Trp2 and Trp3 alleles are found in different
ethnic populations. Several genetic association studies have investigated the role of the COL9A2
and COL9A3 genes in disc degeneration, including Finnish, (Annunen et al., 1999) Japanese,
(Higashino, 2007; Seki, et al., 2006) southern Chinese (Jim et al., 2005) and Indian (Rathod et
al., 2012).
In summary, both COL9A2 and COL9A3 play a role in the integrity of the intervertebral
disc. Certain polymorphisms have been associated with more degenerative changes within the
disc, although their representation varies among ethnic groups.
Aggrecan (ACAN) Gene
Aggrecan is a large chondroitin sulfate proteogycan that functions to hold water content
within the disc, making it more resilient to compressive and mechanical forces (Solovieva et al,
2007; Watanabe, Yamada & Kimata, 1998). With age, the proteoglycan content of the disc
diminishes, and the disc becomes thinner and more fibrotic, resulting in the disorganization of
the disc components (Modic & Ross, 2007).

36

Variable numbers of tandem repeat (VNTR) polymorphisms in the aggrecan (ACAN)
gene, located on chromosome 15q26, have been linked to different levels of lumbar disc
degeneration. The variable number of tandem repeats results in different length aggrecan
proteins possessing differing numbers of attachment sites for chondroitin sulfate. Shorter alleles
have been found to be associated with greater degrees of disc degeneration and development of
disc degeneration at an earlier age (Eser, et al., 2010; Kawaguchi, et al., 1999). However, these
results have not been consistently replicated. For example, one study found that individuals
homozygous for 26 VNTRs experience a higher risk of lumbar disc degeneration and that 25 and
28 VNTRs may actually be protective (Solovieva, et al., 2007).
In summary, ACAN plays a role in disc integrity by its ability to attach proteoglycan
molecules, keeping the disc hydrated. Findings thus far suggest that greater variable numbers of
tandem repeats that encode for longer ACAN molecules provide for more proteoglycan
attachment sites, and may be protective for the disc.
Vitamin D Receptor (VDR ) Gene
Vitamin D receptor gene (VDR) is associated with osteoporosis and osteoarthritis
(Kalichman & Hunter, 2008; Kawaguchi, et al., 2002). The exact influence VDR variants have
on intervertebral disc degeneration is not known, but may play a role in the structure of cartilage
cells (Balmain, Hauchecorn, Pike, Cuisiner-Gleizes & Mathieu 1993; Yuan, et al., 2010). The
location of vitamin D receptor gene is near to the genes for insulin-like growth factor and type II
collagen, and may be a marker for other genes that influence disc degeneration (Kawaguchi et
al., 2002; Kalichman & Hunter 2008).
Single nucleotide polymorphisms of the VDR gene (location 12q13.11) have been
associated with higher incidence of lumbar disc degeneration. In TaqI and FokI polymorphisms

37

tt, Ff, and ff genotypes have been found to be associated with more severe grades of disc
degeneration (Eser, et al. 2010; Videman et al., 1998). Yuan, et al. (2010) was unable to
demonstrate the VDR TaqI tt genotype in a population of Chinese individuals, but there was a
significant increased risk for disc degeneration in persons with the VDR-apa aa genotype.
In summary, many candidate genes have been studied for their effect on lumbar
intervertebral disc degeneration. There is a beginning understanding of the influence of many
genes on disc degeneration, but the mechanism and degree of contribution of each candidate
gene has not been well established to date. The studies on the candidate genes and their effect on
intervertebral disc degeneration differ in methodology, making it difficult to compare findings
across studies. The methods for determination of degree of disc degeneration also vary between
studies. It is becoming clear that findings differ across ethnic groups. More studies must be
undertaken before clarity in the genetic contribution to intervertebral disc degeneration is
achieved.
Selected Candidate Genes for Pain
While many genes have been associated with increased susceptibility to pain, the pain
genotype variables included for this study are opioid receptor mu-1 and catechol-Omethyltransferase (OPRM-1 and COMT).
Opioid Receptor, mu-1 (OPRM1) Gene
Opioid receptor sites play a role in pain. Genetic differences in opioid receptor sites have
been found to play a role in the experience of pain. Differences in mu-opioid receptors influence
pain perception and post-operative analgesic requirements in many studies (Chou, et al., 2006;
DeCapraris, et al., 2011; Henker, et al., 2012; Tan, et al., 2009). The receptors are activated by
both endogenous opioids and opioid drugs (Mura et al., 2013).

38

The SNP A118G allele has been widely studied, with conflicting findings regarding pain
thresholds and opioid requirements for various pain states. The A118G allele is expressed
differently in ethnic subgroups.
There is evidence that the A118G allele is associated with increased opioid requirements
in various pain states, including post-operative, migraine and cancer pain (Chou, et al., 2006;
Gong, et al., 2013; Menon, et al., 2012; Sia, et al., 2013). There is also evidence that pain
threshold may be higher in persons with the A118G OPRM1 allele, although one study was able
to validate this finding only for Caucasians (Hastie et al., 2012; Huang, et al., 2008). However,
the A118G allele has also been associated with higher pain ratings in women, and had no effect
on cortical pain processing in individuals with chronic back pain compared to healthy controls
(Fillingim, et al., 2005; Vossen, Kenis, Rutten, van Os, Hermens & Lousberg, 2010).
In persons with lumbar disc herniation, pain levels during the subsequent year varied by
sex and OPRM-1 genetic differences, irrespective of treatment type (operative and nonoperative) (Olsen, et al., 2012). The single nucleotide polymorphism (SNP) A118G was
associated with less pain in men, but was associated with slower recovery and greater pain levels
in women, in both operative and non-operative treatment groups. One meta-analysis failed to
validate differences in pain level and analgesic requirements based on variation in OPRM-1
genotype (Walter & Lotsch, 2009).
In summary, there is evidence for increased pain threshold and increased opioid
requirements in persons with the A118G allele, but these findings have not been consistent, and
some studies have failed to demonstrate the A118G allele affects cortical pain processing, pain
threshold, or opioid requirements.

39

Catechol-o-Methyltransferase (COMT) Gene
Catechol-o-methyltransferase (COMT) is involved in the metabolism of
neurotransmitters, inactivating catecholamines. COMT may have an influence in the function of
mu-opioid receptors, which are regulated by neurotransmitters (Zubieta, et al., 2003). Zubieta et
al., (2003) studied opioid receptor site activity and pain responses in a small sample of
individuals to determine if different polymorphisms were associated with different levels of
opioid receptor site activation and different pain levels. They were able to demonstrate that
individuals homozygous for the met/met allele in the COMT gene demonstrated lower µ-opioid
system responses and had higher reported levels of pain (Zubieta et al., 2003). The volume of
hypertonic saline necessary to reach a preset level of pain intensity was also lower in met/met
individuals. Individuals with high COMT activity (val/val) had higher mu-opioid system
activation. COMT genotype is associated with processing of pain in the brain, demonstrated by
Positron-Emission Tomography (PET) scanning in the Zubieta et al (2003) study and by
functional Magnetic Resonance Imaging (MRI) (Schmahl, et al., 2012).
Specific COMT polymorphisms have been associated with increased pain perception and
the development of chronic pain states (Diatchenko et al., 2005; Henker et al., 2012; Orrey et al.,
2012;). The studies involving COMT have focused on haplotypes and single-nucleotide
polymorhpisms. Diatchenko et al. (2005) demonstrated that nearly 11% of the variability in
sensitivity to experimental pain in females could be attributed to three distinct COMT haplotypes
based upon genotype at four SNPs. Persons with haplotype GCGG had the lowest responsiveness
to experimental pain, designated as the low pain sensitivity (LPS) haplotype. Individuals with
haplotype ATCA had intermediate pain responsiveness, designated as APS haplotype. The

40

greatest pain responsiveness was observed in individuals heterozygous for ATCA and ACCG
haplotypes, designated the high pain sensitivity (HPS) haplotype.
Single-nucleotide polymorphisms variants of COMT have been studied, with mixed
results. Studies investigating the outcomes of surgical and non-surgical treatment for low back
pain suggest that COMT polymorphism plays a role in pain levels and outcome, although sample
sizes were small, and study methods differed (Dai, et al., 2010; Omair, Lie, Reikeras, Holden &
Brox, 2012). By far the most studied is the Val158Met variant. Val 158 homozygous individuals
have increased COMT activity compared to Met homozygous individuals, with heterozygotes
possessing intermediate activity (Dai et al., 2010; Lotta et al., 1995; Lachman et al., 1996).
COMT activity in general has shown an inverse correlation with pain sensitivity (Dai, et
al., 2010). However, studies examining the association of the Val158Met SNP with pain and
functional outcomes have produced mixed results. Omair et al., (2012) found Val158Met
heterozygotes with discogenic low back pain randomized to surgical and non-surgical treatment
experienced a greater pain improvement after treatment than either Met or Val homozygotes,
although the effect was small. In contrast, no significant association was found between the
Val158Met polymorphism and improvement in post-operative ODI scores in a population of
individuals after lumbar fusion surgery for discogenic pain (Dai et al., 2010).
In summary, while the effects of COMT are known with regard to the effects on
neurotransmitters, its effects on the experience of pain remain unclear. While some associations
between COMT SNPs and haplotypes and the experience of pain have been observed, the
findings have been inconsistent. Moreover, the studies have varied widely in method,
population, and outcome measures used. Overall, the observed associations of both OPRM1 and
COMT genotypes and pain are small, supporting the notion that prediction of treatment outcome

41

in persons with lumbar degenerative conditions is likely related to a constellation of patient
characteristics. There is some evidence that links OPRM1 and COMT genotypes to the symptom
of pain and to physical function in populations with lumbar degenerative conditions.
In summary, many physiological factors have been associated with lumbar degenerative
conditions and the symptoms associated with lumbar degenerative conditions. It is hypothesized
that genotype, BMI, sex, age and smoking physiological factors have the potential to interact
with symptoms to affect physical function in individuals with lumbar degenerative conditions.
Certain situational factors may also interact with symptoms to affect physical function in
individuals with lumbar degenerative conditions.
Situational Factors
Evidence suggests that patient situational factors influence the outcome of lumbar
degenerative conditions. The situational variables included for this study are employment status,
worker’s compensation claim, and insurance type.
Employment Status
Being unemployed has a negative impact on post-treatment outcomes for persons
undergoing treatment for lumbar degenerative conditions. Those working pre-operatively were
ten times more likely to be working post-operatively after lumbar fusion surgery (Anderson,
Schwaegler, Cizek & Leverson, 2006; Burnham et al., 1996; Silverplats et al., 2010; Zieger, et
al., 2011). For patients undergoing lumbar surgery, length of time off work preoperatively was a
strong predictor of outcome in visual analog pain scores and function as measured by the ODI.
Patients off work for 13 weeks or less had more favorable outcomes for pain and physical
function than those who were off work for longer than 13 weeks, regardless of the surgical
procedure (Rohan et al., 2009). While the exact reason for this observation is not known, in

42

general, the longer an individual is off work related to a spine cause, the less likely that
individual is to return to work, and employment prior to surgical treatment was associated with
better physical function and less pain postoperatively (Guyer, et al., 2008; Nguyen, Randolph,
Talmage, Succup, & Travis, 2011).
Similar findings were reported for patients receiving intensive non-surgical treatment for
chronic low back pain. Out of all patient characteristics studied, working prior to treatment was
the variable most strongly associated with improved physical function scores on the ODI after
treatment (van Hooff, Spruitt, O’Dowd, van Lankveld, Fairbank & van Limbeek, 2013). In
summary, being employed prior to treatment for lumbar degenerative conditions and associated
low back pain is associated with better physical function after treatment.
Workers Compensation
Patients receiving worker’s compensation had worse functional status after surgical and
non-surgical treatments for spine conditions (Anderson, Subach, & Riew, 2009; Atlas, Chang,
Kamman, Keller, Deyo, & Singer, 2000; Burnham, et al, 1996; Voorhies, Jiang & Thomas, 2007;
Yang, Lowe, de la Harpe & Richardson, 2010). In one meta-analysis of worker’s compensation
and outcome after any surgical procedure, patients receiving worker’s compensation had worse
outcomes after surgery measured by a disease-specific outcome instrument, a general functional
score, a general health outcome score, a patient satisfaction score or a pain score (Harris,
Mulford, Solomon, van Gelder & Young, 2005). The summary odds ratio for an unsatisfactory
outcome after surgery in persons receiving worker’s compensation was 3.79 (95% CI 3.28-4.37)
( Harris, Mulford, Solomon, van Gelder & Young, 2005).
Similarly, patients receiving worker’s compensation after a low back injury were less
likely to return to work than those not receiving worker’s compensation (Crook, Milner, Schultz

43

& Stringer, 2002). Worker’s compensation and litigation are both associated with worse ODI
scores in patients presenting for treatment for complaints of spine and/or limb pain (Prasarn,
Horodyski, Behrend, Wright & Rechtine, 2012).
Insurance Type
Insurance type can be associated with less optimal health outcomes. Patients with
indigent care plans tend to have reduced access to standard of care and less optimal treatment
outcomes across a variety of health conditions (Greenstein, Moskowitz, Gelijns & Egorova,
2012; Kruper, et al., 2011; McClelland, Guo & Okuyemi, 2011; Yorio, Yan, Xie & Gerber,
2012). Over a several year period, uninsured patients had more complications and longer
intensive care stays after neurosurgery than Medicaid patients, and Medicaid patients had more
complications and longer intensive care stays than Medicare patients in a major Midwestern
medical center (El-sayed et al. 2012). However, in Nationwide Inpatient Sample, outcomes after
surgery for spinal metastasis did not differ among uninsured, Medicaid, or Medicare patients,
after adjusting for acuity of presentation (Dasenbrock et al., 2012).
A medline search revealed no studies examining the relationship of insurance type with
low back pain, lumbar degenerative conditions, or physical function in this population.
However, insurance plans differ substantially in the type and extent of covered treatments. Many
health plans restrict access to diagnostic imaging and treatments, including physical therapy and
spinal injections. Lack of access to these interventions may influence symptom control and
physical function in persons with lumbar degenerative conditions.
In summary, being off work and having a worker’s compensation claim have been
associated with worse outcomes for individuals experiencing lumbar degenerative conditions.

44

While no studies addressed the influence of insurance type on outcomes for individuals with
lumbar degenerative conditions, insurance type can affect outcomes of health care in general.
Psychological Factors
Psychological factors also play an important role in the outcome of treatment for many
spinal conditions. The psychological variable included for this study is depression.
Depression
Depression can contribute to the development of chronic low back pain and is negatively
correlated with outcome and return to work after surgery for lumbar herniated disc (Carragee,
Alamin, Miller & Carragee, 2005; Kohlboeck et al, 2004; Pincus, Burton, Vogel & Field, 2002;
Trief, Grant & Fredrickson, 2000). In an asymptomatic cohort of veterans, depression was a
more reliable predictor of future back pain episodes than baseline MRI findings (Jarvik,
Hollingworth, Heagerty, Haynor, Boyko, & Deyo, 2005). Depressed patients have worse
functional scores and higher pain visual analog ratings than non-depressed patients with similar
musculoskeletal conditions (George, et al., 2011; Kaptan, Yelcin & Kasimcan, 2012).
In summary, physiological, situational and psychological factors have been shown to
significantly influence the outcome of treatment for lumbar degenerative conditions. However,
studies assessing the combined effect of multiple physiological, situational and psychological
factors on the outcome of treatment for lumbar degenerative conditions are lacking. Recognizing
that outcomes for this population are likely due to a multifactorial process, there is a need for
more research addressing the combined effect of these factors.
Symptoms in Persons with Lumbar Degeneration
Symptoms are subjective phenomena that indicate a change in health or normal function
(Dodd, et al., 2001; Fu, LeMone, & McDaniel, 2004; Farrar, Berlin, & Strom, 2003; Fu,

45

McDaniel, & Rhodes, 2007). Fu, McDaniel and Rhodes (2007), define symptom occurrence as
the frequency the symptom is experienced over a period of time, and symptom distress as the
discomfort or suffering accompanying the symptom. Symptoms possess the dimension of
timing, frequency, intensity, duration, and meaning, (Armstrong, 2003). In general, the worse
the symptom experience, the more negative the impact on outcome.
Symptoms of lumbar degenerative conditions include low back pain and limb numbness,
pain, and weakness. Narrowing of the central spinal canal (stenosis) from ligamentum flavum
hypertrophy, bone spurs, and bulging of the outer annulus can contribute to neurogenic
claudication, which refers to lower limb symptoms of pain, weakness, sensory alteration and
fatigue (Genevay & Atlas, 2010). Stenosis of the lateral aspects of the spinal canal or the
neuroforamen can cause radicular symptoms, which refers to leg pain and sensory alteration that
corresponds to the particular nerve root affected, with weakness in the corresponding myotome
(Genevay & Atlas, 2010). Since the lumbar facet joints and the posterior annulus of the
intervertebral disc are enervated, degenerative changes in these structures can contribute to low
back pain (Falco et al., 2012; Moon et al., 2012; Van Zundert, Vanelderen, Kessels & van Kleef,
2012).
The intensity of symptoms has been associated with worse outcomes. Greater pain and
anxiety after discharge for severe burn injuries predicted increased fatigue, increased pain, and
decreased physical function; even at the two-year follow up (Edward, et al., 2007). Wilson,
Robinson, and Turk (2009) clustered fibromyalgia symptoms based on physical or
cognitive/psychological categories, and low, moderate, or high intensity. Subjects with the more
intense symptoms in both the physical and cognitive/psychological categories used more health
care resources, had the worst physical function, and least favorable work characteristics.

46

There is limited literature examining the relationship of symptoms to outcome in persons
with degenerative lumbar spinal conditions, although persons with both back and leg pain tend to
do worse than persons with back pain alone, across many outcome measures. Symptom location
in both the back and leg and greater symptom intensity predicted greater disability in lumbar
spinal stenosis patients (Lin, Lin, & Huang, 2006). Individuals experiencing both leg and back
pain experienced greater pain irritability and activity limitation and missed more work than
individuals with back pain alone in a large Danish study of over 2,600 patients (Kongstead, Kent,
Albert, Jensen & Manniche, 2012). Persons defining their pre-operative pain with more intense
adjectives from the McGill Sensory and Affective Scores experienced worse outcomes after
surgery for herniated disc (Voorhies, Jiang & Thomas, 2007).
Physical Function
Physical Function is a concept often used in health care, yet its definition remains unclear
and its use is inconsistent (Leidy, 1994). Often, the meaning of physical function is implied.
Sometimes used interchangeably with quality of life, functional status, and health status, it has
been measured by many different methods. Physical function has conceptually been used as a
predictor, mediator, and outcome. Optimization of physical function is a key focus for health
care professionals, and it is a requisite part of overall quality of life (Rejeski & Mihalko, 2001,
Ferrans, et al., 2005). Physical function has been described as foundational to the ability to
operationalize roles (Lenz, et al, 1997). Physical function forms the basis for an individual’s
ability to accomplish the activity required to provide for basic needs, fulfill life roles, and
maintain health and well-being (Hoffman, et al., 2009). In their update to the TOUS, Lenz,
Pugh, Milligan, Gift and Suppe (1997) conceptualize the functional performance outcome of the
model to include physical activity, activities of daily living, social activities, and role

47

performance (which includes work). In this study, physical function is operationally defined by
the use of the ODI and the physical function subscale of the SF-36. An important distinction
should be noted between physical function as measured by a disease-specific lumbar instrument
and the physical function subscale of the SF- 36. Physical function as measured by the ODI may
represent a portion of global physical function of an individual. The ODI represents the portion
of physical function that may attributable to lumbar spine influences.
The World Health Organization (WHO, 2002) has defined disability as consisting of
“impairments, activity limitations and participation restrictions”. Human functioning is divided
into three levels: the individual body part, the whole person, and the person in a social context
(WHO, 2002). Activity is one component of functioning according to the WHO, and is defined
as “the execution of a task or action by an individual”. Participation is another component of
functioning. Participation “is involvement in a life situation” (WHO, 2002). The WHO also
describes environmental factors that impact functioning, similar to the physiological, situational
and psychological antecedent factors in the TOUS. These include the “physical, social and
attitudinal environment in which people live and conduct their lives” (WHO, 2002). For this
study, physical function is defined as an individual’s ability to fully perform in the various
physical roles in their lives, to accomplish ADLS, to work, carry out daily tasks for self and
significant others, to be mobile, and maintain leisure physical activities.
Low back pain symptom aggravation by movement is associated with worse ODI scores
(Cai, Pua & Lim, 2007). Low back pain has a negative effect on physical activity and was
associated with measures of disability in a Turkish population (Soysal, Kara & Arda, 2012).
Physical function is worse for individuals with low back pain accompanied by leg pain than for
those individuals with back pain alone (Kongsted, Kent, Albert, Jensen & Manniche, 2012;

48

Konstantinou, et al., 2013; Prins, van der Wurff & Groen, 2013). Individuals rating their back
and leg pain as equal experienced greater interference with physical function as measured by
ODI scores than those rating back pain or leg pain as greater (Sigmundsson, Jonsson &
Stromqvist, 2013). Hirano et al., (2014) found that back pain and knee pain had stronger
associations with reduced physical function in an elderly population than leg pain or leg
numbness. In a cohort of individuals with lumbar spinal stenosis, back and leg pain severity as
measured by VAS was negatively associated with physical function scores on the ODI, even
when adjusted for age and degree of canal stenosis (Kim, et al., 2013). Pain sensitivity,
measured by the Pain Sensitivity Questionnaire, was associated with the severity of pain
measured by the VAS (Kim, et al., 2013). Kongsted, Kent, Albert, Jensen and Manniche (2012)
also found that individuals with back and leg pain with signs of nerve root irritation (depressed
reflexes, weakness, sensory alteration and positive neurotension signs) had worse estimations of
pain and physical function than individuals with back pain alone and individuals with back pain
and leg pain without signs of nerve root irritation.
In summary, multiple patient factors have been found to independently influence
outcomes, including physical function, for persons experiencing lumbar degenerative conditions.
These factors include physiological, situational, and psychological variables and symptoms.
Many of these patient factors have been studied separately for their influence on outcomes in
individuals experiencing lumbar degenerative conditions. More symptom research in this
population is needed. While there are studies examining the relationships between certain
genotypes and their influence on disc degeneration, pain, and functional outcomes in persons
with lumbar degenerative conditions, there are no studies that attempt to identify a profile of
these combined factors and their influence on physical function. A gap of knowledge exists

49

related to how these patient factors and symptoms interact to affect physical function for persons
with degenerative lumbar spinal conditions. Therefore, the aims of this study will address these
factors, symptoms, and genotype to begin to identify their combined effects on physical function
in a population of adults experiencing lumbar degenerative conditions.

50

CHAPTER IV
Methods
Research Design
A cross-sectional, descriptive study design was employed to address the proposed aims.
A randomized sample consisting of individuals referred to a tertiary spine service outpatient
clinic at multi-specialty neuroscience center was used. The following Aims were used to guide
the study. Aim 1 focused on determining the contribution of physiological (BMI, sex, age,
smoking status), situational (employment status, worker’s compensation claim, insurance type),
and psychological (depression) characteristics in persons receiving non-surgical interventions for
degenerative lumbar conditions to symptoms and physical function. Aim 2 sought to develop a
predictive model for the outcome of physical function in persons receiving non-surgical
interventions for lumbar degenerative conditions, using symptoms (back and/or leg pain,
numbness and weakness) and physiological, situational, and psychological patient
characteristics. The exploratory aim was to examine the impact of the physiological
characteristic genotype (disc structural genes and pain genes) on symptoms (back and/or leg
pain, numbness, and weakness) and on physical function in persons experiencing lumbar
degenerative conditions.
Subjects for Aims 1 and 2 were randomly chosen from a database of approximately 1,300
individuals with completed baseline ODI and SF-36 instruments from the tertiary spine service
outpatient clinic from 2009-2012. Patients are referred to the tertiary spine service from primary
care providers and other specialty providers. The spine center is a regional source for specialty
spine care, treating patients for degenerative and trauma-related spine problems. For Aim 3, a

51

randomly selected subset of the study sample for Aims 1 and 2 was used to explore the impact of
genotype on symptoms and physical function.
Sample.
The study sample consisted of persons referred to the spine service at the Hauenstein
Neuroscience Center at Mercy Health Saint Mary’s from February 2009 through early 2012, with
symptoms of back and/or leg pain. As part of the intake process, every patient with a new
encounter at the spine service completed SF-36 and ODI questionnaires, placed in the patient
chart as part of the medical record. All of the raw scores from the completed SF-36 and ODI
questionnaires were also entered into a separate excel data sheet and contained in a password
protected computer file on the hospital system hard drive. The spine service at the Hauenstein
Neuroscience Center has a data base that includes completed baseline ODI and SF-36
questionnaire responses from approximately 1,300 patients. This password protected file is
stored on the hospital hard drive, under the heading “Groups”. A computer program for random
numbers was applied to the excel sheet containing patients with completed ODI and SF-36
questionnaires to arrange individuals in a random order. Medical records were reviewed
proceeding from the beginning of this randomly arranged list, until an adequate sample of
individuals with completed questionnaires and complete physiological, situational and
psychological data were identified.
Inclusion criteria were: 1) aged 18 years or older, 2) back and/or leg complaints of pain,
numbness, and/or weakness, 3) completed SF-36 and ODI information at first clinic visit, 4)
complete information on selected patient factors and symptoms, including a completed anatomic
pain drawing, 5) English-speaking. All eligible persons with lumbar degenerative conditions
were included. Exclusion criteria were: 1) spinal cancer (primary or metastatic), 2) myelopathy

52

or cauda equina syndrome, 3) major psychiatric disorder (personality disorder, schizophrenia and
bipolar illness), 4) spinal fracture, 5) spinal infection, 6) being scheduled for surgery, 7) pain in
the neck and upper extremities, 8) lumbar surgery within the last year, and 9) current pregnancy.
Prisoners, considered a vulnerable population in research, were not treated in the outpatient spine
service.
The target sample size for Aims 1 and 2 was 154. This determination was based on
Tabachnick and Fidell’s (2006) recommendations for sample size using multiple regression.
Tabachnik and Fidell recommend 8(k) + 50 as a general rule for multiple regression, with k =
number of independent variables. Considering BMI, sex, age, smoking status, employment
status, workers compensation claim, insurance type, depression, pain visual analog score, back
pain, leg pain, numbness and weakness as separate independent variables, as in Aim 2, the
required sample size was 154.
All but the genotype data were obtained from the medical record. The raw scores for SF36 and ODI questionnaires were obtained from the separate excel data sheet from the password
protected file on the hospital hard drive. A randomly selected subset of the study participants
with completed ODI and SF-36 questionnaires and complete physiological, situational and
psychological data were contacted regarding genotyping. Aim 3 subjects were selected by
applying a computer program for randomization to the excel sheet containing the patients with
completed ODI and SF-36 questionnaires and all physiological, situational and psychological
data, to arrange these individuals in a random order. Working from the top of this list, Aim 3
subjects were contacted sequentially by phone to participate in genotyping. Since Aim 3 is
exploratory, a smaller sample size of 30 subjects was used as a target. Since the subjects for Aim

53

3 were randomly selected from the study sample for Aims 1 and 2, identical eligibility criteria
were used.
Setting.
The spine service is located in the Hauenstein Neuroscience Center at Mercy Health Saint
Mary’s, a 343-bed urban teaching hospital in Grand Rapids, Michigan. Persons with spinal
symptoms are referred to the outpatient spine service by primary care providers and other
specialists. These providers are mainly from Kent County, Michigan, but also include those from
several outlying counties. There were more than 4,000 patient visits to the spine service in fiscal
year 2011. The providers in the spine service are a contracted physiatrist and an employed nurse
practitioner. The providers work collaboratively with on-site contracted neurosurgeons,
independent provider pain specialists and employed physical therapists specially trained in the
management of spinal disorders. Mercy Health Saint Mary’s is part of a larger Catholic health
system, Trinity Health. Because of the Catholic mission of Mercy Health Saint Mary’s and
Trinity Health, the spine service provides care to the uninsured and underinsured. Complete ODI
and SF-36 intake data are available for more than 1,300 patients currently in the spine service
database.
Instruments and Measures.
All data for the proposed research were extracted from the medical record, except for
genotype data. The specific instruments used to measure each variable are discussed below.
(See Figure 2 for Neurosurgery/Spine Health History (intake questionnaire), Figure 3 for the SF36, Appendix A for the Oswestry Disability Index and Appendix B for the Data Collection Tool,
used to extract patient characteristic data from the medical record).

54

Patient Factors.
Patient factors examined in this study included patient characteristics in the categories of
physiological factors (genotype, BMI, sex, age, smoking status), situational factors (employment
status, worker’s compensation claim and insurance type) and psychological factors (depression).
Physiological factors.
Body Mass Index.
BMI was calculated from the height and weight recorded in the spine service at the initial
visit. Each patient is weighed at the initial visit. The height is reported by the patient. This
information is recorded for each patient in the spine service on the last page of the
Neurosurgery/Spine Health History (intake questionnaire). BMI is a continuous quantitative
variable. Since height is recorded as reported by the patient and not measured directly, BMI may
not be accurate, and this may be a limitation of the study.
Sex.
The sex of each patient is recorded at the time of the initial visit. This information is
listed on the first page of the intake questionnaire. Sex is a categorical variable. Male or female
was recorded for sex.
Age.
The birth date of each participant was extracted from the medical record. The birth date
of each patient is used in the medical record as a patient identifier and validated with the patient,
insurance sources and the referring provider’s medical record by clinic administrative staff. The
patient’s birth date was recorded. Age is a discrete continuous variable.

55

Smoking status.
Since smokers have a higher incidence of back pain than non-smokers (Karahan, Kav,
Abbasoglu & Dogan, 2009; Shiri, Karppinen, Lein-Arjas, Solovieva, & Viikari-Juntura, 2010),
participant smoking status was extracted from the medical record. Each patient’s smoking status
was recorded at the time of the initial visit. This information is listed on the fourth page of the
spine service intake questionnaire. The individual’s current smoking status was recorded as
yes/no. Since smoking status is self-report, this may be a study limitation. Smoking status is a
categorical variable.
Genotype.
Saliva samples were collected by the primary investigator at the spine service at the
Hauenstein Neuroscience Center as a DNA source for genotyping. Once full physiological,
situational and psychological data, symptoms and outcome measures were identified for the
desired number of study subjects, a random subset 30 of these subjects was identified and
contacted for genotyping. Known variants within two candidate genes for pain experience
(OPRM-1 and COMT) and within four candidate genes for disc structural integrity (COL9A2,
COL9A3, ACAN and VDR) were genotyped. Genotyping data are categorical. Physiological,
situational and psychological factor data, symptom and physical function outcome data were
collected from retrospective chart review from the first visit at the spine service, with baseline
data collected from 2009-2012. Saliva samples for genotyping were collected in February, 2014.
This time lapse between data collection times should not be a limitation of the study because
genotype does not change over time (See Procedures section at the end of this chapter for
specific genotyping procedures).

56

Situational factors.
Situational factors conceptualized in the TOUS include marital and employment status,
access to health care, diet, exercise and social support (Lenz, Pugh, Milligan, Gift & Suppe,
1997). The situational factors conceptualized to interact with symptoms to affect physical
function in this study included employment status, worker’s compensation claim and insurance
type.
Employment status.
Participant employment status was obtained from the medical record at the time of the
first visit to the spine service. Employment status is recorded by the patient at the time of the
initial visit on the spine service intake questionnaire (pg. 4, Appendix A). Employment status
was recorded as employed/not employed. If an individual has recorded their status as retired or
disabled, this was recorded as not employed. Employment status is a categorical variable.
Employment status is self-report, and may be a study limitation.
If the individual had a worker’s compensation claim related to the reason for their spine
service visit, this was recorded and validated by clinic administrative staff prior to the time of the
initial visit and noted in the payer information in the medical record. This information is also
listed on the fourth page of the patient’s intake questionnaire. If the patient presented to the
spine service for care as a worker’s compensation claim, this was recorded as “yes”. If the
patient presented to the spine service for care unrelated to a worker’s compensation claim, this
was recorded as “no”. Since worker’s compensation information is validated by administrative
staff, the accuracy is not dependent on patient report. Worker’s compensation claim is
categorical data.

57

Insurance type.
Participant insurance type data was extracted from the medical record at the time of the
first visit to the spine service. Insurance type is validated for each patient at each visit by clinic
administrative staff and recorded on the patient’s chart. Insurance type was recorded as
commercial, Medicaid, Medicare, Tricare, or none. If the patient had more than one insurance
policy, the primary insurance was recorded. Since insurance type is validated by administrative
staff, the accuracy is not dependent on patient report. Insurance type is categorical data.
In summary, all physiological (except genotype) and situational factors were identified
from the medical record at the time of the first visit to the spine service. These variables included
BMI, sex, age, smoking status, work status, worker’s compensation claim and insurance type.
See Appendix D for the Data Collection Tool used to record the physiological and situational
patient characteristics obtained from the medical record. Some data were dependent on patient
self-report, and may represent a study limitation. Genotype data were collected in some cases as
many as four years after the other data.
Psychological factors.
Psychological factors conceptualized in the TOUS include mental state or mood,
affective reaction to illness, and uncertainty about the symptoms and their meaning (Lenz, Pugh,
Milligan, Gift & Suppe, 1997). The psychological factor conceptualized to interact with
symptoms to affect physical function in this study is depression.
Depression.
The presence of various psychological problems was recorded for each patient at the time
of the first visit as part of the past medical history. The medical history section is on page two of
the intake questionnaire. The medical history section allows patients to check a box next to the

58

medical problem, if present. Depression is specifically included in this list. The review of
systems section on the intake questionnaire also includes a list of psychological problems,
including depression, anxiety, bipolar, and “other”. The review of systems list instructs patients
to check a box next to the psychological problem, if present. The review of systems is on the
second page of the intake questionnaire. Medical records from the referring provider are also
received before the first clinic visit. Medical records from the primary care provider include
information on the individual’s past medical history. If the referring provider indicated a history
of depression, it was considered to be present. Depression was recorded as yes/no, using the
medical records from the referring provider. Depression is a categorical variable. In summary,
depression was considered to be present if the referring provider’s notes indicated depression as
part of the medical history. See Appendix D for the Data Collection Tool.
The study is limited because depression was not measured directly. However, scores
from the mental health subscale of the SF-36 were recorded, and the study population average
mental health scores were compared to population norm values as well as published population
values for lumbar degenerative conditions. This was done in order to compare the study
population to published norms for mental health.
Symptoms.
Pain.
The intake questionnaire includes a horizontal pain visual analog scale, (VAS). Patients
are instructed on the intake questionnaire to circle the number, (0-10) that best corresponds to
their current pain level. Patients are asked to record their current pain on an 11 point horizontal
line from 0 (no pain at all) to 10 (the worst pain you can imagine). The VAS score circled was

59

recorded. If more than one number was indicated by the participant, the average score was
recorded, to the nearest .5. The VAS is on page one of the intake questionnaire, Appendix A.
The VAS is a pain intensity measure (Jensen, Karoly & Braver, 1986). The VAS is brief
and easy to administer and score, produces interval-level data, and has been used across a wide
variety of clinical conditions, including acute and chronic pain states (McGuire, 1997).
Although there is some concern over the ability of persons to conceptualize pain in a linear
fashion, the tool is considered reliable and valid (McGuire, 1997).
Validity of the VAS has been explored by comparing it to other methods of reporting
pain intensity. Pearson correlation coefficients for the VAS compared with McGill Pain
Questionnaire sensory, affective and evaluative scales has been reported as 0.49, 0.42, and 0.57,
respectively, in a population of cancer patients (Ahles, Ruckdeschel & Blanchard, 1984).
Correlation coefficient for the VAS and a verbal rating scale in a cancer population has been
reported as 0.81 (Ohnhaus & Adler, 1975). Similarly, the correlation coefficient between the
VAS and a numeric pain scale in a cancer population has been reported as 0.92 (Ahles,
Ruckdeschel & Blanchard, 1984). Correlation coefficients comparing the horizontal VAS with
the vertical VAS, a numeric pain rating score, and a simple descriptive score for pain in a
population with rheumatic diseases were reported at 0.907, 0.616, and 0.726, respectively
(Downie, Leatham, Rhind, Wright, Branco & Anderson, 1978). The VAS has demonstrated
greater sensitivity than a simple descriptive scale (Scot & Huskisson, 1974; Downie, Leatham,
Rhind, Wright, Branco & Anderson, 1978).
Test-retest reliability comparing scores on days one, three and five with days two, four
and six was 0.78 in the same population of cancer patients (Ahles, Ruckdeschel & Blanchard,
1984). The VAS was one of the five most utilized instruments out of eleven pain scales studied

60

in a systematic review of pain instruments for use in chronic low back pain (Chapman et al.,
2011). The VAS was determined to be reliable and responsive in this population, but the authors
did not report reliability statistics. Chapman, et al. (2011) did not identify floor or ceiling effects
with the VAS. The VAS was highly correlated with a verbal rating score for pain in a population
of 85 patients with chronic pain (r = 0.906, p < 0.001) (Cork, et al., 2004).
Cut points for pain intensity have been studied in populations of cancer patients and
patients having undergone amputation of a lower limb who were also experiencing low back
pain. Pain levels 1-4 correspond to mild pain, 5-6 correspond to moderate pain, and 7-10
correspond to severe pain (Jensen, Smith, Ehde & Robinsin, 2001; Kathy, Harris, Hadi & Chow,
2007; Serlin, Mendoza, Nakamura, Edwards & Cleeland, 1995).
New spine patients are asked to complete an anatomic symptom diagram on the intake
questionnaire. An anatomic diagram of the human body, with both anterior and posterior views,
appears on page three of the spine service intake questionnaire. Patients are instructed to place
symbols on the anatomic drawing where their pain is located. A pain diagram overlay was used
to record the precise location of the patient’s pain as described by Werneke, Hart and Cook
(1999) and Cleland, Childs, Palmer and Eberhart (2006). The overlay assigns numbers 1 through
6, corresponding to the anatomic location of the pain from the low back, through the buttock, and
into the leg. Every number correlating to the patient’s location of pain on the anatomic diagram
was recorded for data analysis. Pain below the gluteal fold was considered lower limb pain,
consistent with the Quebec Task Force guidelines described in Atlas, Deyo, Patrick, Convery,
Keller and Singer (1996) and Werneke and Hart (2004). Pain indicated in areas 1 and 2 was
considered back pain for data analysis. Pain in areas 3, 4, 5 and 6 was considered leg pain for

61

data analysis, consistent with Quebec Task Force Guidelines (Atlas, Deyo, Patrick, Convery,
Keller & Singer, 1996; Werneke & Hart, 2004).
Numbness.
Lower limb numbness can be a symptom associated with irritation of a lumbar spinal
nerve root from degenerative changes described previously in Chapter 3. As part of the spine
service intake questionnaire, patients are asked to complete an anatomic symptom diagram for
pain and numbness. An anatomic diagram of the human body, with both anterior and posterior
views, appears on page three of the intake questionnaire. Along with the anatomic drawing, there
are explicit instructions for the patient, showing the symbols to use for the symptoms pain and
numbness. The patient is instructed to place the appropriate symbol for pain and/or numbness at
the location on the body part where the symptom is experienced. The presence of numbness was
also determined by review of the dictated note from the spine service provider at the patient’s
initial clinic visit. The presence or absence of the symptom leg numbness in locations 3, 4, 5 or 6
was recorded as yes/no, respectively. Numbness was considered as a separate symptom in the
data analysis. Since numbness in the low back is non-anatomic for a nerve root distribution,
numbness in the low back was not recorded for data analysis. Numbness is a categorical
variable. Numbness was determined from both the spine service provider’s office dictation and
from patient report, which strengthens the validity of this measure.
Weakness.
Motor strength is evaluated for each patient at the initial visit to the spine service. Motor
strength is assessed by the provider on the first visit to the spine service during the physical exam
and documented in the provider’s dictation of the visit. Weakness was recorded as yes/no. The
presence of weakness was obtained from the medical record. Weakness is a categorical variable.

62

In summary, symptom information was obtained from the medical record. Pain
information was obtained from the VAS and the symptom diagram completed by the patient on
the initial visit to the spine service. Information about numbness was obtained from the
symptom diagram and provider documentation in the medical record. Information about
weakness was obtained from the provider documentation in the medical record. See Table 1 for
a list of study variables and sources.
Table 1
Study Variables
Variable Category
Physiological
Characteristics

Situational
Characteristics

Psychological
Characteristic
Symptoms

Variable Name
BMI

Variable Source
Medical Record

Variable Type
Quantitative

Age
Sex
Smoking Status
Employment Status

Medical Record
Medical Record
Medical Record
Medical Record

Quantitative
Categorical
Categorical
Categorical

Worker’s Compensation
Claim
Insurance Type
Depression

Medical Record

Categorical

Medical Record
Medical Record

Categorical
Categorical

Pain VAS
Back Pain

Medical Record
Medical Record Pain
Diagram (with overlay
to measure)
Medical Record Pain
Diagram (with overlay
to measure)
Medical Record Pain
Diagram and Provider
Dictated Notes
Medical Record
Provider Dictated Notes
Saliva Sample

Quantitative
Categorical

Spine Service Password
Protected Excel File
Spine Service Password
Protected Excel File

Quantitative

Leg Pain

Numbness

Weakness
Genotype
(Physiological
Characteristic)
Physical Function
Outcome Variables

ODI
SF-36 Physical Function
Subscale

63

Categorical

Categorical

Categorical
Categorical

Quantitative

Physical Function.
Physical function status was measured by scores on the ODI and the physical function
subscale of the SF-36. ODI and SF-36 raw data for Spine Service patients have been entered
onto an excel spreadsheet in the spine service office. The data are contained in a passwordprotected file. Only the primary investigator had access to the password. Once complete data on
patient characteristics, symptoms, and physical function were identified for an adequate number
of participants, data were cleaned, and raw data were scored according to ODI and SF-36 scoring
instructions.
Oswestry Disability Index (ODI).
The ODI was used to measure physical function (Fairbank, Couper, Davies & O’Brien,
1980). The ODI is a 10-item disease-specific instrument for the lumbar spine population and is
widely used as an outcome measure for patients with lumbar degenerative conditions (Roland &
Fairbank, 2000). Although the authors of the original version hold the copyright, they do not
require permission for its use (Roland & Fairbank, 2000). See Appendix C for the ODI.
The ODI takes approximately five minutes to complete and one minute to score. Two
items address pain and the remaining eight items focus on how activities of daily living are
affected by pain (Monticone, et al, 2009). Each item has six response levels, 0-5. The total
numeric score is doubled and expressed as a percentage. The sum of the 10 scores is expressed as
a percentage of the maximum scores, ranging from 0-100. Lower scores reflect better function.
The ODI has been found to be reliable, with intra-class correlations reported to be 0.840.94 (Davidson & Keating, 2002; Fritz & Irrgang, 2001). The ODI has high correlation (.77),
with another lumbar functional instrument, the Roland-Morris Disability Questionnaire (RMDQ)
(Fairbank & Pynsent, 2000). Internal consistency is demonstrated by Cronbach’s Alpha ranging

64

from 0.71-0.87 (Roland & Fairbank, 2000). Reproducibility at 24-hour intervals was reported as
r = 0.99, at four days as r = 0.91, and at one week as r = 0.83 (Roland & Fairbank, 2000).
Because the ODI measures a clinical condition that can vary from day to day, the reduction in
reproducibility scores may reflect natural fluctuations in the individual’s spine condition. Testretest reliability has been reported to range from 0.83-0.99 (Fairbank & Pynsent, 2000).
Evidence also exists for the validity of the ODI. Comparing ODI change scores to
individual’s global rating of change, the ODI ranked best out of all the outcome measures tested,
r = -0.64, p < 0.01 (Taylor, Taylor, Foy & Fogg, 1999). A systematic review of multiple spine
outcome measures found the ODI to be valid and highly correlated with the RMDQ, but the
correlation coefficient and significance were not reported in this review article (Chapman et al.,
2011).
The ODI was found to be comparable to other lumbar specific physical function
instruments, including the RMDQ, Quebec Back Pain Disability Scale, the Waddell Disability
Index and the physical function subscale of the SF-36, in responsiveness to change (Davidson &
Keating, 2002). The responsiveness of the ODI has been demonstrated in individuals with acute
and chronic low back pain. The correlations between positive ODI change scores and
individual’s estimations of improvements were 0.66 and 0.49 for the acute low back pain group
and the chronic low back pain group, respectively (Grotle, Brox & Vollestad, 2004). In another
large (N = 970) study comparing a disease-specific measure (the ODI), to the SF-36 in
individuals with back and leg pain, the two instruments were similar in responsiveness (Receiver
Operating Characteristic Curve = 0.723 and 0.721 for the ODI and the physical functioning
subscale of the SF-36, respectively) (Walsh, Hanscom, Lurie & Weinstein, 2003).

65

Normative data with weighted mean scores on the ODI have been reported as 10.19 for
normal populations, 26.63 in persons with spondylolisthesis, 36.65 in persons with neurogenic
claudication, 43.3 in persons with chronic back pain, 44.65 in persons with sciatica, 44.83 in
persons with fibromyalgia, and 48.04 in persons with spinal metastases (Roland & Fairbank,
2000). Scores from 0-20% reflect minimal disability, scores from 20-40% reflect moderate
disability, scores from 40-60% reflect severe disability, scores from 60-80% reflect “crippled”
state, and scores from 80-100% indicate the person is “bedbound or exaggerating” (Fairbank,
Couper, Davies & O’Brien, 1980).
Medical co-morbidities can affect ODI scores. Baseline survey results from a large data
set (N = 26, 290) were regressed with co-morbidities and patient characteristics. Although the
investigators report that ODI scores decreased at baseline from an average of 62.4 to 42.0 for
individuals with no and ≥ 7 co-morbidities, respectively (this would actually represent an
improvement on the ODI), the regression analysis results table showed that poor self-rated health
and the presence of worker’s compensation had the most negative impact on ODI scores, with
depression and smoking also having a significantly negative effect (Slover, Abdu, Hanscom,
Lurie & Weinstein, 2006).
An expert panel convened to discuss a special issue of Spine devoted to measurement
recommended that whenever possible, a condition-specific instrument for back pain should be
used, specifically, either the ODI or the RMDQ (Roland & Fairbank, 2000). Two systematic
reviews of lumbar-specific outcome instruments found the ODI and the RMDQ to be the most
comprehensively validated functional measures for responsiveness (including improvement and
deterioration in status), reliability and validity Chapman, et al., 2010; Cleland, Gillani, Bienen &

66

Sadosky, 2010). The total ODI score (the sum of the 10 scores expressed as a percentage of the
maximum score) was used as an outcome measure.
Physical function subscale of the SF-36.
The Short Form-36 (SF-36) is a multi-purpose health survey that yields two component
summary scores (i.e. the physical component summary and mental component summary) and
eight subscale scores, including physical function, role-physical, bodily pain, general health,
vitality, social function, role-emotional, and mental health. The SF-36 was developed to monitor
health status of individuals with chronic health and psychiatric conditions over time (Tarlov,
Ware, Greenfield, Nelson, Perrin & Zubkoff, 1989). Higher scores in each subscale or the total
survey reflect better function. The physical function subscale scores are standardized such that
the general U.S. population average scores are 50, with a standard deviation of 10 (Beaton &
Schemitsch, 2003). See Appendix B for the SF-36.
The SF-36 has been used extensively across a wide range of clinical conditions and
populations. It is widely used in orthopaedics and spine surgery. Many studies examining
physical function as an outcome use the physical component summary scores and the physical
functioning subscale scores.
The SF-36 physical function subscale has been demonstrated to be reliable. In a
population with low back pain followed over a 6 week period, intra-class correlations (ICC) were
0.83 and 0.91 for the SF-36 physical function scale for groups estimating they were “unchanged”
and “about the same”, respectively (Davidson & Keating, 2002). Patrick, Deyo, Atlas, Singer,
Chapin & Keller (1995) reported a similar ICC of 0.89. The SF-36 was specifically used to test
psychometrics in individuals with rheumatoid arthritis and osteoarthritis participating in a
placebo-controlled drug trial after a 3-14 day washout period (Kosinski, Keller, Hatoum, Kong &

67

Ware, 1999). Internal consistency for all the subscales ranged from 91.4%-97.1%, and item
discriminant validity ranged from 96.9%-100.0%. For the physical function subscale, item
internal consistency ranged from 0.37-0.80, and reliability ranged from 0.89-0.91. Although the
authors found the distribution of scores for the physical function subscale (and others) positively
skewed in this population, floor and ceiling effects were observed only for the subscales of role
physical and role emotional (Kosinski, Keller, Hatoun, Kong & Ware, 1999).
There is evidence for the validity of the physical function subscale of the SF-36. The
physical function subscale is similar to the RMDQ (Patrick, Deyo, Atlas, Singer, Chapin &
Keller, 1995; Davidson & Keating, 2002). Large effect sizes (≥0.80) were found after
orthopedic surgery in the SF-36 subscales of physical function, role physical and bodily pain and
small (0.20-0.49) to moderate (0.50-0.79) effect sizes were found in the subscales measuring
mental and social aspects (Busija, Osborn, Nilsdotter, Buchbinder & Roos, 2008).
The standard version of the SF-36 has been designed for administration at four week
intervals. There are normative data available for many different health conditions. The
instrument has been translated into 121 languages. Shorter versions of the instrument have been
developed, including the SF-12 and SF-8 (Ware, 2003). The SF-36 and its scoring software are
copy-righted and must be purchased.
The presence of co-morbidities and other patient characteristics can change SF-36 scores.
In patients with no co-morbidities, average change scores on the physical function subscale were
16.8, but with four co-morbidities, the average change scores decreased to 6.9 (Slover, Abdu,
Hanscom & Weinstein, 2006). Physical function subscale change scores were most negatively
impacted by headaches, poor self-rated health and age (Slover, Abdu, Hanscom & Weinstein,
2008). See Appendix A for ODI and SF-36 instruments. Physical function subscale scores were

68

used as an outcome measure, in addition to the ODI total score. Mental health subscale scores
were also recorded, in order to compare study participant average with known population norms
and normative scores for individuals with lumbar conditions.
Medications.
Because the use of analgesic medications can affect the experience of pain, analgesic use
was also recorded on the data collection instrument. Five categories of medications were
recorded, including non-steroidal anti-inflammatories, steroids, narcotics, other analgesics, and
anti-convulsants such as gabapentin and pregabalin, often used to treat neuropathic pain.
Procedures
The dissertation proposal was approved by the student’s Dissertation Committee. Grant
funding was awarded by the Saint Mary’s Foundation for resources to accomplish the
exploratory Aim 3, involving genotyping. The study was approved by expedited review by the
Mercy Health Saint Mary’s Institutional Review Board (IRB # 13-1816-01-SM). A Reliance
Agreement exists between Mercy Health Saint Mary’s and Michigan State University.
Eligibility criteria were: 1) aged 18 years or older, 2) back and/or leg complaints of pain,
numbness, and/or weakness, 3) completed SF-36 and ODI information at first clinic visit, 4)
complete information on selected patient factors and symptoms, including a completed anatomic
pain drawing, 5) English-speaking. All eligible persons with lumbar degenerative conditions
were included. Exclusion criteria were: 1) spinal cancer (primary or metastatic), 2) myelopathy
or cauda equina syndrome, 3) major psychiatric disorder (personality disorder, schizophrenia and
bipolar illness), 4) spinal fracture, 5) spinal infection, 6) being scheduled for surgery, 7) pain in
the neck and upper extremities, 8) lumbar surgery within the last year, and 9) current pregnancy.

69

Recruitment Procedures for Aims 1 and 2.
Since the data needed to address Aims 1 and 2 were available from the medical record in
the spine service, no recruitment of subjects for these aims was required. These subjects were
identified by the primary researcher from the database of approximately 1,300 completed ODI
and SF-36 questionnaires in the spine service. A computer program for random numbers was
applied to the excel sheet containing patients with completed ODI and SF-36 questionnaires to
arrange individuals in a random order. The medical records were reviewed proceeding from the
beginning of this randomly arranged list for all of the data required to address Aims 1 and 2. If
the necessary data on patient characteristics and symptoms were incomplete in the medical
record, the next subject on the randomly arranged list was selected. This process was followed
until complete data for patient characteristics and symptoms were collected for 163 subjects on
the data collection sheet, more than the minimum number required for statistical analysis. This
was done to assure sufficient data. Once complete data on patient characteristics and symptoms
for 163 participants were identified the ODI and SF-36 physical function and mental health
subscale raw data was cleaned and then scored according to guidelines. All data for each patient
characteristic and outcome were then entered into an excel sheet for data analysis, with coded
patient identifiers.
Recruitment Procedures for Aim 3.
When 163 subjects with complete data on patient characteristics were identified from the
database of completed outcome measures, a subset of 30 individuals was selected from this
population for the exploratory genotyping aim using the same computerized method to randomly
arrange the 163 subjects from Aims 1 and 2. These individuals were contacted by phone by the
primary investigator. Using a script, potential participants in the genotyping aim were informed

70

of the study and purpose and were invited to provide saliva samples. A minimum of two attempts
were made to contact each potential subject for Aim 3, using mobile or home phone numbers.
Subjects who traveled to the spine service to provide a saliva sample were provided a $10 gift
card to a local retailer.
Subjects who agreed to provide a saliva sample were given a telephone number to contact
the primary investigator to cancel or reschedule the appointment time, if necessary. Subjects
calling to cancel were given the opportunity to reschedule. If the subject declined to reschedule,
the next subject identified on the random list was contacted. The primary investigator continued
to contact potential subjects from the randomized list for genotyping until 30 subjects agreed to
provide saliva samples. The investigator telephoned 105 potential subjects in order to arrange
saliva collection from the desired 30 participants for Aim 3. Each time a participant cancelled a
saliva collection appointment, the next potential subject on the random list was contacted. If a
subject failed to show for saliva collection, one attempt was made by telephone to re-schedule.
This process continued over the course of two weeks. Saliva samples were successfully
collected from 28 participants, but two participants either cancelled, or declined after arriving on
the last day of saliva collection.
When the primary investigator met with subjects to collect the saliva sample, an eight
page informed consent form was used to describe the dissertation study, (including the aims and
significance) and the potential risks to participants. Subjects were assured of confidentiality.
HIPAA authorization was included in the informed consent. Subjects signed the informed
consent form prior to providing a saliva sample. The investigator was present to answer
questions and provide clarification if needed. Procedures for collection of biological samples
were reviewed, using the instructions provided with the Oragene OG-500 saliva collection kits.

71

Consents were stored in a locked cabinet at the spine service, accessible only to the primary
investigator. If participants failed to show for a scheduled time to provide a saliva sample, a
phone contact was made to inquire about rescheduling. If the participant declined to reschedule,
the next subject identified by the table of random numbers was contacted.
Data Collection Procedures.
Saliva samples were obtained from subjects at the spine service. Although DNA yields
are lower in saliva than blood, saliva DNA yields are sufficient for Taqman assays (Abraham et
al., 2012). Saliva as the source for DNA was chosen because it was easier to collect and did not
involve a venipuncture procedure. Saliva samples are stable at room temperature and were
stored in the Clinical Trial Unit at Saint Mary’s until transport to Michigan State University
Genomics Core Facility, after a Materials Transfer Agreement was signed by both Mercy Health
Saint Mary’s legal representatives and the Michigan State University Technologies office. Only
one type of saliva collection method was used, reducing potential variation in protein
composition in saliva (Mohamed, Campbell, Cooper-White, Dimeski & Punyadeera, 2012;
Golatowski, et al., 2013). A data use agreement has been signed by the investigator, Michigan
State University and Mercy Health Saint Mary’s.
Genotyping Procedures.
Saliva samples were stored at Mercy Health Saint Mary’s Clinical Trials Unit in a locked
specimen storage room until data collection for Aim 3 was complete. The investigator accessed
this storage room only by admittance by Clinical Trials Unit staff.

Saliva samples then were

transported to the Michigan State University Genomics Core Facility by the investigator directly
to Michigan State Core Genomics lab staff for processing and genotyping.

72

Saliva samples were processed using the Oragene DNA OG-500 kits (DNAGenotek,
Ontario, Canada). An average 35-40µg of high quality DNA/1ml can be extracted from saliva.
After establishing DNA sample yield and purity through spectrophotometry and PCR
amplification, DNA samples were split and stored at –20C for immediate access and at –80C
for back-up. The stability of Puregene and Oragene-purified genomic DNA is verified to at least
11 years (Puregene Product Information, www.gentra.com; Oragene Product Information,
www.dnagenotek.com/). Subsequent genotyping was conducted using the Taqman® PCR
platform (Applied Biosystems, Carlsbad, CA) for all variants except the ACAN VNTR
polymorphism and the COL9A2 polymorphism. The ACAN VNTR was genotyped according to
the methods described by Eser and colleagues (2010). COL9A2 (rs2228564) was genotyped
using direct sequencing. See Table 2 for specific candidate genes and Single Nucleotide
Polymorphisms (SNPs).
A study manual was created to include all procedures, scripts for patient contacts, and
data collection tools for all study variables. The primary investigator was responsible for all data
collection procedures, thereby ensuring consistency. See Appendix D for data collection tool.

73

Table 2
Genes Selected for Genotyping with SNPs (Single-Nucleotide Polymorphisms) Tested
Gene Acronym Locus
COL9A2
1p34.2
COL9A3
20q13.33
ACAN
15q26.1
VDR
12q13.11
OPRM1
6q25.2
COMT
22q11.21

SNP rs#

Major/Minor Allele

MAF

rs2228564

A/C/G/T

0.385 (C)

rs61734651a

C/T

0.080 (T)

Exon 12
VNTR

variable

variable

rs731236b

C/T

0.264 (C)

rs1799971

A/G

0.348 (C)

rs4680c

A/G

0.389 (A)

a

also referred to asTrp3 allele in some studies.
also referred to as VDR Taq1 allele in some studies.
c
also referred to as val158met in some studies.
MAF = minor allele frequency
b

Data Management.
SF-36 and ODI raw data for spine service patients have been entered onto an excel
spreadsheet on a spine service computer. The data are contained in a secure, password-protected
file on the hospital hard drive. Only the primary investigator had access to the password.
Identifiable patient data was not stored on a personal laptop. Identifiable patient data was not
stored on a flash drive. Identifiable patient data was coded prior to statistical analysis. ODI and
SF-36 raw data was converted into subscale scores with proprietary scoring instructions from the
user’s manuals.
The medical record was reviewed for patient characteristics and symptoms, and this data
was recorded onto the patient characteristics and symptoms data collection instrument. These
activities occurred within the spine service office by the primary investigator. All procedures are
detailed in a study procedures manual, which contains information on the study, scripting for
initial phone contact with eligible persons, all study data collection instruments, and steps in the
74

study process. Data cleaning and entering from the ODI and SF-36 was completed by the
primary investigator.
For Aim 3, saliva samples were collected in person from consenting subjects by the
primary investigator. Saliva was collected according to instructions provided by the
manufacturer. Saliva sample containers were labeled with the coded number assigned to the
subject and no patient identifiers were used on the saliva sample container. The collected saliva
samples were stored at room temperature in the Clinical Trials Unit at Mercy Health Saint
Mary’s until all samples were collected. The saliva samples were then transferred to the Core
Genomics Facility at Michigan State University by the investigator after a Material Transfer
Agreement was signed between the two institutions.
Data Analysis.
Summary statistics were created to describe the population. Descriptive statistics were
used to describe the population on the characteristics of physiological factors (genotype, BMI,
age, sex, smoking status), situational factors (employment status, workers compensation claim,
and insurance status), and psychological factors (depression). Assumptions for normality, lack of
extreme outliers, multicollinearity, homoscedasticity, and linearity were checked.
Aim 1 Specific Strategies: To determine the contribution of physiological (BMI, sex,
age, smoking status), situational (employment status, worker’s compensation claim,
insurance type), and psychological (depression) factors in persons receiving non-surgical
interventions for degenerative lumbar conditions to symptoms and physical function.
In the specific Aim 1 analysis, the independent variables include BMI, sex, age, smoking
status, employment status, workers compensation claim, insurance type, and depression.
Symptoms, including pain (location in back only or back and leg) numbness in the leg, and

75

weakness were treated as both dependent variables and independent variables in separate
statistical testing. Physical function is considered the dependent variable, as measured by the
scores on the ODI and the physical function subscale of the SF-36.
In the specific Aim 1 analysis, descriptive statistics were used to describe the population
on the characteristics of physiological factors, situational factors, psychological factors, and
symptoms. Although depression was not measured directly, SF-36 mental health subscale scores
for the study population were compared with population norms. Multivariate methods (multiple
regression) was used to examine the contribution of patient characteristics to symptoms and to
physical function. Multiple regression was also used to examine the contribution of patient
characteristics and symptoms to scores on the ODI and physical function subscale of the SF-36.
Aim 2 Specific Strategies: Develop a predictive model for the outcome of physical
function in persons receiving non-surgical interventions for lumbar degenerative
conditions, using symptoms (back and/or leg pain, numbness, and weakness) and
physiological, situational, and psychological patient factors.
In the specific Aim 2 analysis, multivariate methods were used to identify how patient
characteristics and symptoms combine to predict physical function.
Exploratory Aim 3 Specific Strategies: Explore the impact of the physiological
factor genotype (disc structural genes and pain genes) on symptoms (back and/or leg pain,
numbness, and weakness) and on physical function in persons experiencing lumbar
degenerative conditions.
In the specific exploratory Aim 3 analysis, genotype data on COMT, OPRM-1, ACAN,
VDR, COL9A2 and COL9A3 was analyzed. Descriptive statistics were used to describe the allele
and genotype frequencies for each polymorphism. Multivariate statistics (multiple regression)

76

were used to examine the contribution of genotype to symptom experience and genotype to
physical function.
Limitations.
Limitations of this study include the descriptive, cross-sectional design and the use of
secondary data. This limited the ability to establish a temporal relationship between the
predictors and the outcome. This study is also limited by the use of a convenience sample,
limiting generalizability of the findings. The presence of depression was based on review of the
medical record in this study and not measured directly. The validity of this variable and the
interpretation of its significance in this study were therefore limited. Future prospective studies
with this population should be planned measuring depression directly with reliable and valid
instruments. Data on patient characteristics, symptoms and physical function predates
genotyping data by as many as four years, which should not affect interpretation of results
because genotype does not change over time.
This dissertation study organizes and examines salient antecedent physiological,
situational and psychological factors and symptoms and how they interact to influence physical
function in individuals with lumbar degenerative conditions. Considering genotype to be a
physiological antecedent factor is consistent with the concepts and propositions of the TOUS,
and represents an innovative incorporation of biomarkers with patient characteristics and
symptoms and their influence on outcome in this population.

77

Human Subjects.
Human subjects characteristics and involvement.
The study sample includes persons referred to the Spine Service at the Hauenstein
Neuroscience Center at Mercy Health Saint Mary’s from February 2009 through early 2012, with
back and/or leg pain. All English-speaking patients aged 18 or older, referred to the spine service
with back and/or leg pain and complete data for outcome measures and patient characteristics
were eligible to be included in the study population. Women, men and minorities had equal
chance of being represented in the study population, as did disadvantaged patients, since the
mission of Mercy Health Saint Mary’s includes care to the underserved. The Hauenstein
Neuroscience Center (including the spine service) does not provide care to children (persons
under the age of 18).
The spine service at the Hauenstein Neuroscience Center has a data base that includes
completed ODI and SF-36 questionnaire responses from approximately 1,300 patients. Inclusion
criteria are: 1) aged 18 years or older, 2) back and/or leg complaints of pain, numbness, and/or
weakness, 3) completed SF-36 and ODI information at first clinic visit, 4) complete information
on study variables patient characteristics and symptoms, including a completed anatomic pain
drawing, and 5) English-speaking. A specific radiographic diagnosis prior to presentation at the
spine service is not required. Exclusion criteria are: 1) spinal cancer (primary or metastatic), 2)
myelopathy or cauda equina syndrome, 3) major psychiatric disorder (personality disorder,
schizophrenia, and bipolar illness), 4) spinal fracture, 5) spinal infection, 6) being scheduled for
surgery, 7) pain in neck and upper extremities, 8) lumbar surgery within the last year, and 9)
pregnant status. Prisoners, considered a vulnerable population in research, are not treated in the
outpatient spine service.

78

Sources of material.
ODI and SF-36 data will be obtained from a password-protected file that contains
baseline scores from more than 1,300 individuals treated in the spine service. Using a computer
program to randomize the file containing baseline scores for ODI and SF-36 questionnaires, 163
subjects with complete data on outcome measures were identified from the database, slightly
more than the target number of 154. The medical records of these individuals were reviewed for
patient characteristics and symptoms. Subjects for genotyping were identified from the 163
study participants from Aims 1 and 2 by applying the same computer program to randomize the
list of 163. DNA for genotyping was isolated from saliva samples collected from consenting
subjects.
Potential risks.
Potential risks to subjects were minimal. The procedure of collecting saliva samples
causes little to no discomfort and has a minimal possibility of infection. It is not anticipated that
information generated through this research will affect the insurability of subjects. Insurance
companies will not have access to this research data. Participants were informed that the genetic
analyses performed during this study are not a form of treatment, diagnosis, or prediction of
lumbar spinal degeneration. Therefore the results of the genetic studies were not reported to the
participants nor were they placed in the subject’s medical record.
Participation in this study may cause anxiety related to increased awareness of the genetic
contributions to lumbar degenerative conditions. Basic education and reassurance regarding the
multi-factorial nature of lumbar spinal degeneration was provided by the primary investigator, if
necessary. Referrals for genetic and psychological counseling were not made.

79

Protection against risk.
Several strategies to protect human subjects were implemented. Saliva samples were
collected by the primary investigator. In the event of psychological or emotional distress related
to an increased awareness of lumbar spinal degeneration heritability, subjects had access to basic
education regarding the multi-factorial nature of lumbar spinal degeneration from the PI. In
addition, several safeguards to ensure privacy of data were undertaken. Coded ID numbers were
used on the saliva collection containers, DNA sample vial and genotype reports. Any flash
drives with subject information were coded, to avoid identification of subjects. The code key
linking names and ID numbers were kept separately from other data in a password protected file
on the hospital hard drive only. All paper records were maintained in locked files in a locked
research office. In addition, published reports of results will not include subject identifiers.
Because the clinical usefulness of the candidate lumbar spinal degeneration genotype data
remains experimental, results of the genotyping were not disclosed to subjects. Subjects were
advised that they could withdraw their genotype data from the study analysis at any time without
penalty. Following completion of this study, DNA samples were destroyed, in compliance with
the Data Use Agreement between Mercy Health Saint Mary’s and Michigan State University.
See Appendix E for Data Use Agreement.
Participants maintained the right to withdraw from the study at any time without affecting
their care at the spine service. Confidentiality of all findings, including genotyping results, was
maintained.
Potential benefits of the proposed research to subjects and others.
While there were no anticipated direct benefits to subjects for participating in this study,
the findings may enable health care providers to better predict persons at relatively high risk for

80

worse physical function outcome from a combination of individual characteristics (including
genotype) and symptoms. Findings may also provide researchers with a better understanding of
the genetic mechanisms contributing to decreased physical function in persons with lumbar
degenerative changes. This understanding may lead to the development of improved
interventions in the future. This chapter outlined the methods, variables, subjects, setting,
sources of data, procedures, potential risks, human subjects protection and anticipated benefits.
Chapter will discuss results.

81

CHAPTER V
Results and Interpretation
Organization of Results Chapter
The results chapter will be organized into sections that describe demographic
information, patient characteristics and physical function status for study participants. Data
analysis and findings will then be discussed, organized by study aim. Because of the large
number of variables used in the multiple regression models, for the sake of brevity, only the full
and final models are shown in the tables.
The population of subjects included in Aims 1 and Aims 2 will be referred to as the
sample. The population of subjects included in Aim 3 will be referred to as genotyped subjects.
First, demographic descriptions of the study population will be discussed. Patient
characteristics and symptoms will be described. Physical function status (ODI scores and
physical function subscale scores for the SF-36) for the study population will be discussed.
Next, data analysis and findings for Aim 1 will be reviewed. Aim 2 data analysis and findings
will then be discussed. Other data analysis relevant to the review of the literature, the study
population and the Aims of the study will also be reviewed.
The demographic description of genotyped subjects will progress in the same manner.
Any significant differences between the two populations will be discussed. Last, Aim 3 data
analysis and findings will be described.
Medical Records Reviewed
Patient characteristics data (BMI, sex, age, smoking, employment status, worker’s
compensation claim, insurance type and depression) were obtained from the medical record.
Likewise, symptom data was obtained from the medical record from the subjects first visit to the

82

spine service during the time period 2009-2012. Specifically, the intake questionnaire was
reviewed as well as the provider’s dictated report of the initial clinic visit. Since many of the
subjects had been evaluated in the spine service between 2009 and 2012, most of the medical
records were retrieved from an off-site storage facility. The randomized list of subjects was sent
to the storage facility by the medical records staff.
In order to ensure complete data for analysis for Aims 1 and 2 for the desired 154
subjects, medical records for 275 individuals were requested from the medical records staff at the
Hauenstein Neuroscience Center at Mercy Health Saint Mary’s. Out of the 275 medical records
requested, 64 records (23%) could not be located by the medical records staff of the Hauenstein
Neuroscience Center or by the staff of the storage facility where past medical records were kept.
Another 48 medical records (17.5%) were excluded because they did not meet inclusion criteria,
leaving 163 useable medical records (59%) for the study.
The reasons for exclusion were pain complaints not related to the lumbar spine,
insufficient data on symptoms, patient characteristics, or outcome measures, and mental illness.
Twenty- eight records (10%) were excluded because the clinical complaints were related to the
cervical spine. Seven medical records (2.5%) were excluded because the pain visual analog
scale was not completed. Six medical records (2.2%) were excluded because the pain diagram
was not completed. Three medical records (1%) were excluded because of a diagnosis of bipolar
illness or multiple personality disorder. Three other medical records (1%) were excluded
because of incomplete data on an outcome measure and a patient characteristic, lumbar surgery
within the previous year and age less than 18. One medical record (.03%) was quarantined and
unavailable because of ongoing litigation.

83

There were therefore 163 medical records that met inclusion criteria and were included in
the study. This number exceeded the desired study population of 154. However, since all of the
medical records contained useable data, it was decided to include all of them in the study.
SF-36 physical function and mental health item responses were checked for missing and
out of range responses. One subject’s SF-36 physical function subscale responses were all
missing but one, so this was treated as a missing variable. No subject had more than 3 missing
responses for the SF-36 physical function subscale. These subscales were able to be scored,
using the Half-Scale Rule, which states that if at least half of the subscale items have been
answered, the subscale can be scored and used (Ware et al., 2007). The two mental health
subscale responses requiring reverse coding were recoded according to scoring instructions
(Ware et al, 2007). Physical function and mental health subscale scores were transformed into zscores, and then to norm-based scores using the formulas from the User’s Manual for the SF36v2 Health Survey (Ware, et al., 2007). ODI scores were expressed as a percentage, according
to scoring instructions for the ODI.
Demographic Information and Patient Characteristics for the Sample
The mean BMI for study subjects was 30.4 (S.D. = 8.41). Forty-eight subjects (29.4%)
were considered overweight, with a BMI between 25 and 29.9. Forty-eight subjects (29.4%)
were considered obese, with a BMI between 30 and 39.9. Twenty subjects (12.3%) had a BMI
of 40 or greater, considered to be class III, or high risk obesity (National Institutes of Health,
2014). In summary, 71.1% of the subjects (n = 116) in the study population were overweight,
obese, or Class III obese. See Table 3 for study population N and percent of overweight, obese
and class III obese individuals.

84

Table 3
Sample N and % of Overweight, Obese and Class III Obese (N = 163)
BMI

N

%

Overweight (BMI >/= 25-29.9)

48

29.4%

Obese (BMI >/= 30-39.9)

48

29.4%

Class III Obese (BMI >/= 40)

20

12.3%

Nearly 58% of study subjects were female (n = 91). The mean age of study subjects was
54 (S.D. = 16.85). The youngest was 22 and the oldest was 93. The majority of subjects were
younger than 65 (119 subjects, or 73%). Forty-four subjects (27%) were aged 65 or older.
Twenty- eight percent of all study subjects (46) were current smokers/tobacco users. Among
men, 35% (24) were smokers. Twenty-three percent (22) of women were smokers.
More than 56% (92) of the study subjects were not working at the time of presentation to
the spine service. Among men, 52% (36) were not working. Sixty percent of women were not
working (56). Among the 119 subjects younger than 65, 46% (55) were not working.
The majority of study subjects were covered by commercial insurance (57%, or 93
subjects). Medicare insurance accounted for coverage for nearly 21% of subjects (36), followed
by 20 covered by Medicaid (12%), eight with no insurance (5%), three with auto (<2%), three
with worker’s compensation (<2%) and two with tricare (1%). Among the study subjects who
were working, 87.32% (62) were covered by commercial insurance, 4.23% (3) had Medicare
insurance, 4.23% (3) had Medicaid insurance, 2.82% (2) had no insurance and 1.4% (1) had auto
insurance. See Table 4 for insurance coverage for the entire sample and for working subjects in
the sample.

85

Table 4
Insurance Coverage for the Sample and for Working Subjects in the Sample
Sample

Working Subjects in the Sample

N and % (N = 163)

N and % (N = 71)

3 (1.84)

0

Commercial Insurance

93 (57.06)

62 (87.32)

Medicare

34 (20.86)

3 (4.23)

Medicaid

20 (12.27)

3 (4.23)

No Insurance

8 (4.91)

2 (2.82)

Auto

3 (1.84)

1 (1.4)

Tricare

2 (1.23)

0

Insurance Type
Worker’s Compensation

Thirty-one percent (51subjects) had been diagnosed with depression, according to the
clinical diagnosis from the medical record. Of the women, 38% (36 subjects) were depressed,
while 22% of the men (15 subjects) were depressed. This difference was statistically significant
(x2 = 5.075, p. = .024).
Mental health subscale scores and depression were significantly related for the sample.
Both parametric and non-parametric tests of association were used, because of one extreme
outlier (t-statistic = 4.180, p. = .000; Mann-Whitney U test p. = .000). This finding helps
strengthen the reliability of the variable depression for this study. See Table 5 for sample patient
physiological, situational and psychological characteristics, N and percent for categorical
variables. See Table 6 for sample patient physiological characteristics, range, minimum,
maximum, mean and SD for BMI and age.

86

Table 5
Sample Patient Physiological, Situational and Psychological Characteristics, N and % for
Categorical Variables (N = 163)
Category
Physiological

Characteristic
Smoking

Sex

Situational

Work Status

Insurance

Variable

N

%

Non-smoking

117 71.8

Smoking

46

28.2

Female

91

57.7

Male

69

42.3

Working

71

43.6

Non-working

92

56.4

Commercial

93

57.1

Medicare

34

20.9

Medicaid

20

12.3

Auto

3

1.8

Worker’s Compensation

3

1.8

Tricare

2

1.2

No Insurance

8

4.9

Depressed (per medical

51

31.3

Type

Psychological

Depression

record)
Not Depressed

112 68.7

Table 6
Sample Patient Physiological Characteristics, Range, Minimum, Maximum, Mean and SD
for BMI and Age (N = 163)
Category

Characteristic

Physiological

Sample
Range

Mean

SD

BMI

15.81-71.04

30.44

8.41

Age

22-93

54.07

16.85

87

Symptoms for the Study Population
The mean pain VAS for the sample was 6.83 (S.D. = 2.2). One individual had a pain
VAS of 0. Women and men were similar with regard to the mean pain VAS. The mean pain
VAS for women was 6.85 (S.D. = 1.99) and the mean pain VAS for men was 6.79 (S.D. = 2.46),
not a statistically significant difference (p. = .861). Twenty-seven subjects (16.6%) reported
weakness. Sixty-five subjects (40%) reported numbness. One subject reported no pain. Thirtynine subjects (24%) reported back pain only. One hundred subjects (61%) reported back and leg
pain. Twenty-three subjects (14%) reported leg pain only. See Table 7 for sample pain VAS
mean, maximum, minimum and sex differences. See Table 8 for sample categorical symptoms
of pain location, weakness and numbness.
Table 7
Sample Symptom Continuous Variable: Pain VAS (N = 163)
Symptom
Pain VAS Study Sample

Range Min.

Max. Sample Mean

SD

10

0

10

6.83

2.2

Females

8

2

10

6.85

1.99

Males

10

0

10

6.79

2.46

88

Table 8
Sample Symptom Categorical Variables: Pain Location, Weakness and Numbness
(N = 163)
Symptom

N

%

Back Pain Only

39

23.9

Leg Pain Only

23

14.1

Pain Location

Back and Leg Pain 100 61.3
Weakness
No Weakness

136 83.4

Weakness

27

16.6

No Numbness

98

60.1

Numbness

65

39.9

Numbness

Outcome Measures for the Sample
The mean ODI score for the sample was 50.96 (S.D. = 19.3) (range: 0-90). For the
ODI, lower scores reflect better physical function. Scores between 40 and 60 reflect severe
disability, (Fairbank, Couper, Davies & O’Brien, 1980). The mean ODI score for the sample is
higher (worse) than published normative scores for individuals with spinal metastases, which is
48.04, implying that the study sample perceived themselves as having worse physical function
than a population of subjects with spinal metastatic cancer (Roland & Fairbank, 2000).
Likewise, the mean physical function subscale score from the SF-36 for the sample was
32.57 (S. D. = 11.98) (range: 14.94-57.03). For the SF-36 subscales, higher scores reflect better
function. Published healthy population norm score is 54.76 (S.D. = 6.04) (Ware et al., 2007).
The sample mean physical function subscale score is lower than published mean score for
individuals with back pain and sciatica (46.78, S.D. = 11.14). Moreover, the sample mean

89

physical function subscale score is worse than mean population scores of the 25 th percentile for
individuals with back pain and sciatica (40.87, S.D. = 11.14) (Ware et al., 2007). The sample
score compares similarly to the mean population scores from the 25 th percentile of those with
diabetes (32.68, S.D. = 11.18), and is only slightly better than the 25th percentile for those with
cancer (30.64, S.D. = 11.52) (Ware et al., 2007). See Table 9 for sample population physical
function outcome scores for the ODI and the SF-36.
Table 9
Sample Physical Function Scores: Range, Minimum, Maximum, Mean and SD
Sample
Outcome Measure

N

Range

Mean

SD

ODI

162

0-90

50.96

19.29

SF-36 Physical Function

161

14.94-57.03

32.58

11.62

Subscale

The sample mean mental health subscale score from the SF-36 was 44.73 (S.D. = 14.4).
Published healthy population norm score is 53.43 (S.D. = 8.38) (Ware et al., 2007). The mean
score for a population with depression is 36.70 (S.D. = 11.08) (Ware et al., 2007). Therefore,
while the sample mean mental health subscale score was higher than for those with depression, it
was worse than the population norm. In fact, the sample mean mental health subscale score was
worse than the mean score for individuals with back pain and sciatica (47.46, S.D. = 10.78), but
slightly better than for those in the 25th percentile with back pain and sciatica (41.71, S.D. =
10.78) (Ware et al, 2007).
There were no statistically significant differences between men and women with regard
to ODI, physical function subscale or mental health subscale scores using independent t-tests (p.
= .765, .218, and .836, respectively).

90

In summary, the study population consisted of individuals who tended to be overweight.
More than half of the study population was not working at the time of data collection. With a
mean pain VAS just under 7 (on a 0-10 scale), the study population was experiencing severe
pain. Finally, the perceived physical function of the study population was severely limited. A
discussion of the analyses for Aims 1 and 2 will be presented in the following section.
Aim 1 Analysis
For Aim 1, to determine the contribution of physiological (BMI, sex, age, smoking
status), situational (employment status, worker’s compensation claim, insurance type), and
psychological (depression) factors to symptoms and physical function, the relationships between
patient characteristics and symptoms were examined first. Next, the relationship between patient
characteristics and outcome measures (ODI and physical function subscale of the SF-36) were
examined.
The relationship between patient characteristics and pain VAS.
Multiple regression was used to examine the contribution of BMI, sex, age, smoking,
employment status and depression to pain VAS. Sex, age and depression were not significant in
predicting pain VAS and were eliminated from the model early. BMI and employment status
became insignificant in the model as well. Smoking was the only significant predictor of pain
VAS. Smoking was associated with higher pain VAS scores but explained only 8.6% of the
variance (F = 15.066, p. = .000). See Table 10 for coefficients, significance and R2 for the full
and the final models for predicting pain VAS using patient characteristics (BMI, sex, age,
smoking, employment status and depression).

91

Table 10
Coefficients and Observed Levels of Significance for the Full and Final Multiple Regression
Models for Predicting Pain VAS Using Patient Characteristics (BMI, Sex, Age, Smoking,
Employment Status and Depression) (N = 163)
Unstandardized
Coefficients
Model

Independent
Variables

Full

(Constant)
BMI
sex
age
smoking
employment
status
depression

Final

(Constant)
smoking

Beta

Std.
Error

5.337

.943

.029

.020

.130

Standardized
Coefficients
Beta

t

Sign.

R2

5.661

.000

0.118

.113

1.476

.142

.345

.029

.377

.707

.005

.011

.038

.467

.641

1.408

.397

.289

3.542

.001a

-.469

.353

-.106

-1.328

.186

.175

.373

.037

.470

.639

6.423

.195

32.940

.000

1.425

.367

3.881

.000a

.293

0.086

Dependent Variable = Pain Visual Analog Scale (VAS)
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001
Insurance types (commercial insurance, Medicare, Medicaid, worker’s compensation and
no insurance) were evaluated for effect on pain VAS using multiple regression, keeping in the
final three patient characteristics of BMI, employment status and smoking from the previous
model. Tricare was not included in this model, because only two subjects in the study population
had this type of insurance. In the final model, smoking, having Medicaid insurance and not
having insurance were all associated with higher pain VAS, explaining 13% of the variance (F =
7.907, p. = .000). See Table 11 for the coeffecients, significance and R2 for the full and the final
92

regression models for predicting pain VAS using patient characteristics (BMI, employment
status and smoking) and insurance type.
Table 11
Coefficients and Observed Levels of Significance for the Full and Final Multiple Regression
Models for Predicting Pain VAS Using Patient Characteristics (BMI, employment status
and smoking) and Insurance Type (N = 163)
Unstandardized
Coefficients
Model
Full

Final

Independent
Variables

Beta

Std.
Error

(Constant)

6.293

1.089

BMI
smoking

.019
1.080

.021
.410

depression
employment
status

.190
-.504

workers
compensation
commercial
insurance

Standardized
Coefficients
Beta

t

Sign.

R2

5.778

.000

.156

.074
.222

.924
2.635

.357
.009a

.372
.396

.040
-.114

.512
-1.272

.609
.205

-2.249

1.533

-.138

-1.467

.144

-.302

.988

-.068

-.305

.760

medicare
medicaid

-.405
.378

1.020
1.069

-.075
.057

-.397
.353

.692
.724

noins
(Constant)

1.076
6.319

1.242
.196

.106

.867
32.318

.387
.000

.130

b

smoking
medicaid

1.018
1.096

.388
.519

.209
.164

2.624
2.112

.010
.036b

no insurance

1.731

.782

.171

2.213

.028b

Dependent Variable = Pain Visual Analog Scale (VAS)
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001

The relationship between patient characteristics and weakness and numbness.
Logistic regression was used to examine the relationship between patient characteristics
(BMI, sex, age, smoking, employment status and depression) and the symptoms of weakness and

93

numbness. Using a backward step-wise approach, age was the only statistically significant
variable in the final logistic regression model for patient characteristics and weakness (Table 12).
However, it had no discriminatory value, predicting weakness in every subject (Table 13). Using
a backward step-wise approach, employment status was statistically significant in the final
logistic regression model for patient characteristics and numbness (Table 14). However, its
predictive value was only 60% (Table 15). Therefore, it was concluded that there were no
patient characteristics with predictive value for the symptoms of weakness and numbness.
To determine the effects of patient characteristics on the symptom of pain location, the
patient’s documented location for pain, 1-6 was separated into the three clinically relevant
categories of back pain, leg pain, and back pain with leg pain, consistent with Quebec Task
Force Guidelines (Atlas, Deyo, Patrick, Convery, Keller & Singer, 1996; Werneke & Hart, 2004)
and according to methods described in Cleland, Childs, Palmer and Eberhart (2006). Pain in
areas 1 and 2 was considered back pain, and pain in areas 3-6 was considered leg pain. There
was no statistically significant relationship between pain VAS and pain location with ANOVA
(F = .273; p. = .761).

94

Table 12
Logistic Regression for Predicting Weakness Using Patient Characteristics, Full and Final
Models (N = 163)

Step

Independent
Variables

1
(Full)

BMI
sex

.018
-.448

.027
.447

age
smoking

.016
-1.244

employment
status
depression

6 (Final)

Beta

S.E.

Wald

df

Sign.

Exp(B)

.437
1.006

1
1

.508
.316

1.018
.639

.015
.675

1.184
3.390

1
1

.277
.066

1.016
.288

-.143

.476

.090

1

.764

.867

.138

.501

.076

1

.783

1.148

Constant

-2.537

3.734

1

.053

.079

smoking

-1.308

1.31
3
.640

4.183

1

.041

.270

Constant

-1.355

.229

35.002

1

.000

.258

Dependent Variable: Weakness
S.E.: standard error
Wald: Wald statistic
df: degrees of freedom
Exp(B): odds ratio
Table 13
Classification Table for Full and Final Logistic Regression for Predicting Weakness Using
Patient Characteristics (N = 163)

Predicted
weakness
Observed
no weakness

no weakness

weakness

Percentage
Correct

136

0

100.0

Step 1
(Full)

weakness

weakness
Overall Percentage

27

0

.0
83.4

Step 6
(Final)

weakness

136
27

0
0

100.0
.0

no weakness
weakness

Overall Percentage

83.4

95

Table 14
Logistic Regression for Predicting Numbness Using Patient Characteristics, Full and Final
Models (N = 163)

Step

Independent
Variables

Beta

S.E.

Wald

df

Sign.

Exp(B)

1
(Full)

BMI
sex

.012
-.119

.020
.346

.387
.118

1
1

.534
.731

1.012
.888

age
smoking

.010
.565

.011
.402

.875
1.977

1
1

.350
.160

1.010
1.759

employment
status
depression

.998

.359

7.714

1

.005

2.714

.455

.373

1.489

1

.222

1.576

Constant
employment
status

-2.033
.804

.970
.327

4.393
6.058

1
1

.036
.014

.131
2.234

Constant

-.776

.224

11.953

1

.001

.460

6
(Final)

Dependent Variable: Numbness
S.E.= standard error
Wald = Wald statistic
df = degrees of freedom
Exp(B): odds ratio
Table 15
Classification Table for Full and Final Logistic Regression for Predicting Numbness Using
Patient Characteristics (N = 163)

Step 1
(Full)

Step 6
(Final)

Predicted
Extremity numbness
no extremity
extremity
numbness
numbness
84
14

Observed
Extremity
no extremity
numbness
numbness
extremity numbness
Overall Percentage
Extremity
no extremity
numbness
numbness
extremity numbness
Overall Percentage

96

48

17

63

35

29

36

Percentage
Correct
85.7
26.2
62.0
64.3
55.4
60.7

The relationship between patient characteristics and pain location.
Discriminant analysis was used to determine the effects of patient characteristics (BMI,
age, sex, smoking, employment status and depression) on pain location. Only age and smoking
had significant associations with pain location (F = 6.643, p. = .002 and F = 5.331, p. = .000,
respectively). However, using the squared canonical correlations, age alone only accounted for
8% variance, and age and smoking together only accounted for 5% of the variance. See Table 16
for discriminant analysis using BMI, sex, age, smoking, employment status and depression to
predict pain location (back pain only, back and leg pain, leg pain only). Chi-square was used to
determine if sex and pain location, depression and pain location and worker’s compensation were
related, but there were no statistically significant relationships identified (x2 = 1.943, p. = .584;
x2 = 3.042, p. = .385 and x2 = 1.925, p. = .588, respectively). See Table 17 for chi-square tests
using the categorical patient characteristics of sex, depression and worker’s compensation as
predictors of pain location. Only three subjects had worker’s compensation insurance. With
analysis of variance (ANOVA), there was no statistically significant relationship between BMI
and pain location (F = .726, p. = .583). See Table 18 for analysis of variance for patient
continuous variable BMI as a predictor for pain location. None of the patient characteristic
variables predicted pain location. It was likely that that the sample was too small to model
sufficiently using these variables.

97

Table 16
Discriminant Analysis Patient Characteristics (BMI, Age, Sex, Smoking, Employment
Status and Depression) as Predictors of Pain Location (Back Pain Only, Back and Leg
Pain, Leg Pain Only) (N = 162)
Patient

Wilk’s

FSign.

Canonical

Squared Canonical

Correlation

Correlation

Eigenvalue

Characteristic

Lambda

statistic

1 Age

.923

6.643

.002b

.084

.278

.08

.878

5.331

.000c

.051

.221

.05

2 Age and
Smoking

Dependent Variable: Pain Location
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001
Table 17
Chi-square Categorical Patient Characteristics (Sex, Depression, Worker’s Compensation)
as Predictors of Pain Location (Back Pain Only, Back and Leg Pain, Leg Pain Only) (N =
162)
Patient
Chi-square

df

Sign.

Sex

1.943

3

.584

Depression

3.042

3

.385

1.925

3

.588

Characteristic

Worker’s
Compensation

Dependent Variable: Pain Location
df = degrees of freedom
Table 18
Analysis of Variance for Continuous Patient Characteristic (BMI) as Predictor of Pain
Location (Back Pain Only, Back Pain and Leg Pain, Leg Pain Only) (N = 162)
Patient
Characteristic
(BMI)
Between
Groups
Within Groups
Total

Sum of
Squares

df

Mean
Square

F

Sign.

155.015

3

51.672

.726

.538

11315.630
11470.645

159
162

71.167

Dependent Variable: Pain Location
df = Degrees of Freedom
Sign.: Significance
Groups: Back Pain Only, Back Pain and Leg Pain, Leg Pain Only

98

The relationship between patient characteristics and outcome measures.
Because there were two outcome measures, the ODI and the physical function subscale
from the SF-36, separate statistical tests were used to explore the influence of patient
physiological, situational and psychological characteristics on physical function. These tests
will be reported separately.
Backward multiple regression was used to examine the effects of patient characteristics
BMI, sex, age, smoking, employment status and depression on ODI scores. BMI, smoking and
employment status were significant and explained 15.4% of the variance in ODI scores (F =
9.621, p. = .000). Being employed was associated with a lower (better) ODI score, but higher
BMI and smoking were associated with worse ODI scores. See Table 19 for the full and final
backward regression models for predicting ODI using patient characteristics (BMI,sex, age,
smoking, employment status and depression).
Insurance types (commercial insurance, Medicare, Medicaid, worker’s compensation and
tricare) were evaluated for effect on ODI with multiple regression, in combination with the
patient characteristics of BMI, smoking and employment status, which were significant in the
first model. In the final model, higher BMI and smoking were associated with higher (worse)
ODI scores, and having commercial insurance or Medicare were associated with lower (better)
ODI scores (F = 8.597, p. = .000), explaining 18% of the variance. See Table 20 for the
coefficients, significance and R2 for the full and the final regression models for predicting ODI
scores using patient characteristics (BMI, employment status and smoking) and insurance type.

99

Table 19
Coefficients and Observed Levels of Significance for the Full and Final Backward
Regression Models for Predicting ODI Score Using Patient Characteristics (N = 162)
Unstandardized
Coefficients
Model
Full

Final

Independent
Variables
(Constant)

Beta

Std.
Error

31.921

8.082

Standardized
Coefficients
Beta

t

Sign.

R2

3.950

.000

16.3

b

BMI
sex

.392
1.140

.171
2.959

.172
.029

2.300
.385

.023
.701

age
smoking

.073
13.198

.091
3.397

.064
.309

.800
3.885

.425
.000c

employment
status
depression

-5.013

3.035

-.129

-1.652

.101

2.925

3.192

.071

.916

.361

(Constant)
BMI

36.975
.424

5.607
.168

.186

6.594
2.533

.000
.012b

smoking
employment
status

12.652
-5.821

3.167
2.879

.297
-.150

3.995
-2.022

.000c
.045b

15.4

Dependent Variable = Oswestry Disability Index (ODI)
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001

Using backward multiple regression BMI, sex, age, smoking, employment status and
depression were examined for their effect on SF-36 physical function subscale scores. BMI, age,
employment status and depression were significant in the final model, explaining 17.7% of the
variance (F = 8.399, p. = .000). BMI, depression and age were associated with lower (worse)
physical function subscale scores, while being employed was associated with higher (better)
physical function subscale scores. See Table 21 for the coefficients, significance and R2 for the
full and final backward regression models for predicting SF-36 physical function subscale scores
using patient characteristics (BMI, sex, age, smoking, employment status and depression).

100

Table 20
Coefficients and Observed Levels of Significance for the Full and Final Multiple Regression
Models for Predicting ODI Scores Using Patient Characteristics (BMI, Employment Status
and Smoking) and Insurance Type (N = 162)

Model
Full

Final

Independent
Variables
(Constant)
BMI
smoking
employment
status
workers
compensation
commercial
insurance
medicare
tricare
medicaid
(Constant)
BMI
smoking
commercial
insurance
medicare

Unstandardized
Coefficients
Std.
Beta
Error
49.142
8.345
.307
.173
9.689
3.467
-4.146
3.391

Standardized
Coefficients
Beta
.134
.227
-.107

t
5.889
1.776
2.795
-1.223

Sign.
.000
.078a
.006b
.223

-9.991

11.623

-.070

-.860

.391

-10.623

5.964

-.274

-1.781

.077a

-10.508
-7.771
.319
46.828
.317
10.015
-11.429

6.468
13.877
6.691
6.975
.170
3.407
3.809

-.222
-.045
.005
.139
.235
-.294

-1.625
-.560
.048
6.714
1.865
2.940
-3.000

.106
.576
.962
.000
.064a
.004b
.003b

-8.889

4.664

-.188

-1.906

.058a

Dependent Variable = Oswestry Disability Index (ODI)
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001

101

R2
.193

.180

Table 21
Coefficients and Observed Levels of Significance for the Full and Final Backward
Regression Models for Predicting SF-36 Physical Function Subscale Score Using Patient
Characteristics (N = 161)
Unstandardized
Coefficients
Model
Full

Final

Independent
Variables
(Constant)

Beta

Std.
Error

54.119

4.784

Standardized
Coefficients
Beta

t

Sign.

R2

11.312

.000

.196

c

BMI
sex

-.376
-1.654

.101
1.753

-.274
-.071

-3.729
-.943

.000
.347

age
smoking

-.157
-3.594

.054
2.033

-.229
-.139

-2.908
-1.768

.004b
.079a

depression
employment
status

-2.846
2.901

1.892
1.791

-.114
.124

-1.504
1.619

.135
.107

(Constant)
BMI

50.299
-.369

4.351
.101

-.269

11.559
-3.651

.000
.000c

age
depression

-.130
-3.496

.052
1.856

-.189
-.140

-2.493
-1.884

.014b
.061a

employment
status

3.727

1.744

.160

2.137

.034b

.177

Dependent Variable = SF-36 physical function subscale
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001
Insurance types (commercial insurance, Medicare, Medicaid, worker’s compensation and
tricare) were evaluated for effect on the physical function subscale of the SF-36 using multiple
regression, in combination with the patient characteristics of BMI, age, employment status and
depression. Since smoking was nearly significant in the first multiple regression, smoking was
also added. In the final model, BMI, age, smoking and having Medicaid insurance were all
associated with worse physical function subscale scores and explained 18.2% of the variance (F
= 8.689, p. = .000). See Table 22 for the coefficients, significance and R2 for the full and the

102

final regression models for predicting SF-36 physical function subscale scores using patient
characteristics (BMI, depression, age, employment status and smoking) and insurance type.
Table 22
Coefficients and Observed Levels of Significance for the Full and Final Multiple Regression
Models for Predicting SF-36 Physical Function Subscale Scores Using Patient
Characteristics (BMI, Depression, Age, Employment Status and Smoking) and Insurance
Type (N = 161)
Unstandardized
Coefficients
Model
Full

Final

Independent
Variables
(Constant)

Beta

Std.
Error

53.866

5.725

Standardized
Coefficients
Beta

t

Sign.

R2

9.408

.000

.207

b

BMI
age

-.342
-.201

.106
.073

-.249
-.292

-3.243
-2.757

.001
.007b

smoking
employment
status

-2.300

2.111

-.089

-1.089

.278

2.005

2.081

.086

.963

.337

depression
commercial
insurance

-2.648

1.900

-.106

-1.394

.166

1.339

3.673

.057

.365

.716

medicare
workers
compensation

1.925

4.581

.068

.420

.675

-1.955

7.015

-.023

-.279

.781

tricare
medicaid

-5.003
-3.730

8.394
4.047

-.048
-.106

-.596
-.922

.552
.358

(Constant)
BMI

55.603
-.361

4.286
.102

-.263

12.972
-3.536

.000
.001b

age
smoking

-.191
-3.456

.052
1.994

-.278
-.134

-3.639
-1.733

.000c
.085a

medicaid

-5.839

2.748

-.166

-2.125

.035b

Dependent Variable = SF-36 physical function subscale
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001

103

.182

The relationship between pain location and physical function.
One-way analysis of variance was used to determine if the presence of back pain and leg
pain together was related to worse physical function consistent with the review of literature. The
groups were: back pain only, back pain with leg pain, and leg pain only. There was a significant
between groups difference (F = 3.582, p. = .030), indicating that the group means were different.
To determine which means were significantly different, a multiple comparison test was used.
The presence of back and leg pain together was associated with higher (worse) scores on the
ODI, compared to back pain or leg pain alone, using Least Significant Difference (LSD) for
multiple comparisons (p. = .025). See Table 23 for one-way analysis of variance for the between
groups difference for pain location and ODI scores. Table 24 presents multiple comparisons
correction for the between groups difference and ODI scores.
Table 23
One-way ANOVA for Between Groups Difference for Pain Location and ODI
Scores (N = 162)
Sum of
Squares

df

Mean
Square

F

Sign.

3.582

.030

Between
Groups

2486.399

2

1243.200

Within Groups
Total

54830.023
57316.422

158
160

347.025

Dependent Variable: ODI Scores
df: degrees of freedom
Sign.: Significance
Groups: Back pain only, back pain and leg pain, leg pain only

104

Table 24
Multiple Comparison Test to Determine Which Means Differed for Pain Location and ODI
Scores (N = 162)
95% Confidence Interval
Multiple
Comparison Test
Least Significant
Difference

Pain Location

back pain only
back and leg
pain

leg pain only

Pain Location
back and leg
pain

Sign.

Lower
Bound

Upper
Bound

.058

-13.67

.22

leg pain only
back pain only

.524
.058

-6.64
-.22

12.99
13.67

leg pain only
back pain only

.025a
.524

1.23
-12.99

18.56
6.64

back and leg
pain

.025a

-18.56

-1.23

Dependent Variable: ODI Scores
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001
In contrast, there was no association between location of pain and physical function
subscale scores from the SF-36. There was no significant between groups difference between
pain location (back pain only, back pain and leg pain, leg pain only) and physical function
subscale scores (F = 1.503, p. = .226). See Table 25 for one-way ANOVA for between groups
difference for pain location and SF-36 physical function subscale scores.
There were physiological and situational factors that were found to contribute to
symptoms and physical function in this study. The physiological factor smoking contributed to
higher pain VAS. The situational factor Medicaid insurance also contributed to higher pain
VAS. There were no physiological, situational or psychological factors that contributed to the
symptoms of numbness or weakness. The physiological factors of higher BMI and smoking
contributed to worse scores for physical function on both the ODI and the physical function
subscale of the SF-36. The situational factors of having Medicare and Commercial insurance
were associated with better physical function scores on the ODI. These findings are consistent
105

with previous findings that smoking is associated with worse low back pain, although the
mechanism is unclear (Karahan, Kav, Abbasoglu & Dogan, 2009; Shiri, Karppinen, Lein-Arjas,
Solovieva, & Viikari-Juntura, 2010). Medicaid insurance has been associated with worse health
outcomes in general, and the key factor may be reduced or restricted coverage for treatments that
reduce pain, such as physical therapy and spinal injections (Greenstein, Moskowitz, Gelijns &
Egorova, 2012; Kruper, et al., 2011; McClelland, Guo & Okuyemi, 2011; Yorio, Yan, Xie &
Gerber, 2012). BMI and smoking were associated with worse physical function scores, and this
may be related to restricted pulmonary function and decreased mobility related to weight. These
findings are consistent with previous findings (Prasarn, Horodyski, Behrend, Wright & Rechtine,
2012; Rihn et al, 2013).

Table 25
One-way ANOVA for Between Groups Difference for Pain Location and SF36 Physical Function Subscale Scores (N = 161)
Sum of
Squares

df

Mean
Square

F

Sign.

1.503

.226

Between
Groups
Within Groups

398.184

2

199.092

20801.534

157

132.494

Total

21199.718

159

Dependent Variable: SF-36 physical function subscale scores
df: degrees of freedom
F: F-statistic
Groups: Back pain only, back pain and leg pain, leg pain only

Aim 2 Analysis
For Aim 2, to develop a predictive model for the outcome of physical function in persons
receiving non-surgical interventions for lumbar degenerative conditions using symptoms (back
and/or leg pain, pain VAS, numbness, and weakness) and physiological, situational, and
106

psychological patient factors, general linear modeling was used. Using the significant predictors
for ODI score from the Aim 1 analysis (smoking and BMI), these were added into an analysis of
variance with symptoms. Weakness and pain location were not significant and were dropped
from the model. With general linear modeling, BMI, pain VAS, smoking and extremity
numbness were kept in the final model. These variables were entered into a backward step-wise
multiple regression with ODI score the dependent variable. BMI, smoking, pain VAS and
extremity numbness were significantly associated with ODI scores (F = 20.679, p. = .000) and
explained almost 35% of the variance. See Table 26 for coefficients, observed level of
significance and R2 for the final backward step-wise multiple regression for predicting ODI
scores using patient characteristics (BMI, smoking, pain VAS) and symptoms (extremity
numbness).
Similar findings were supported for physical function subscale scores. Once again, using
the significant predictors for SF-36 physical function scores from the Aim 1 analysis, (BMI, age
and smoking) were added with symptoms into an analysis of variance. With general linear
modeling, BMI, age and pain VAS were kept in the final model. These variables were entered
into a backwards step-wise multiple regression with SF-36 physical function subscale score the
dependent variable. BMI, age and pain VAS were significantly negatively associated with
physical function subscale scores (F = 18.019, p. = .000), explaining almost 26% of the variance.
See Table 27 for coefficients, observed level of significance and R2 for the full and the final
backward step-wise multiple regression for predicting SF-36 physical function subscale scores
using patient characteristics (BMI and age) and symptoms (pain VAS).

107

Table 26
Coefficients and Observed Level of Significance for the Final Backward Step-wise Multiple
Regression for Predicting ODI Scores Using Patient Characteristics (BMI, Smoking, Pain
VAS and Symptoms (Extremity Numbness) (N = 162)
Unstandardized
Coefficients
Model

Independent
Variables

Final

(Constant)

B

Std.
Error

10.153

5.795

Standardized
Coefficients
Beta

t

Sign.

R2

1.752

.082

.347

a

BMI
smoking

.279
6.818

.146
2.853

.125
.163

1.907
2.390

.058
.018b

Pain VAS
extremity
numbness

4.196
4.992

.607
2.516

.472
.129

6.912
1.984

.000c
.049b

Dependent variable: ODI scores
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001

There were physiological and situational factors with symptoms that had predictive value
for physical function in this study. Higher BMI, smoking, higher pain VAS and extremity
numbness explained 35% of the variance in ODI scores, while higher BMI, older age and higher
pain VAS explained 26% of the variance in SF-36 physical function subscale scores.

108

Table 27
Coefficients and Observed Level of Significance for the Full and Final Backward Step-wise
Multiple Regression for Predicting SF-36 Physical Function Subscale Scores Using Patient
Characteristics (BMI and Age) and Symptoms (Pain VAS)
(N = 161)
Unstandardized
Coefficients
Full

Final

Standardized
Coefficients

Model
(Constant)

B
64.600

Std. Error
4.554

Beta

t
14.184

Sign.
.000

BMI
smoking

-.347
-1.646

.095
1.916

-.255
-.064

-3.657
-.859

.000c
.391

pain VAS
age

-1.857
-.155

.392
.049

-.344
-.227

-4.734
-3.161

.000c
.002b

(Constant)
BMI

64.010
-.341

4.499
.095

-.251

14.229
-3.606

.000
.000c

pain VAS
age

-1.953
-.144

.375
.047

-.362
-.210

-5.202
-3.045

.000c
.003b

R2
.261

.257

Dependent Variable: SF-36 physical function subscale score
a
indicates p-values < 0.10, b indicates p-values < 0.05, c indicates p-values < 0.001
Summary of Findings for Aims 1 and 2
There were patient physiological and situational characteristics with significant
associations with symptoms. Specifically, smoking, having Medicaid insurance, or not having
insurance was significantly associated with higher pain VAS ratings.
There were patient physiological and situational characteristics that were also associated
with physical function. Specifically, higher BMI and smoking, along with older age and having
Medicaid insurance were all significantly associated with worse SF-36 physical function
subscale scores. Higher BMI and smoking were associated with worse ODI scores, while having
Medicare or Commercial insurance were associated with better ODI scores. With older age,
physical function may be declining and degenerative spinal changes increase over time (Cheung,
et al., 2009).

109

There were patient physiological and situational characteristics that, along with
symptoms, were able to explain a portion of the variance in physical function scores. In
particular, higher BMI, smoking, higher pain VAS and numbness accounted for 35% of the
variance in ODI scores. Higher BMI, older age and higher pain VAS accounted for 26% of the
variance in SF-36 physical function subscale scores.
Aim 3 Study Subjects
To select study subjects for Aim 3, the study population of 163 randomly selected
medical records with complete patient characteristics, symptom and outcome measures entered
on an excel sheet were subjected to computerized randomization. Working from the beginning
of the list after randomization, individuals were contacted by telephone. One hundred five
individuals were called before 30 agreed to provide saliva samples, representing 29% of
individuals called. The main reason for the small percentage of consenting subjects was inability
to contact individuals by phone. For many, the phone number had been disconnected. For a few
subjects, transportation was a barrier. One subject arrived, but could not find parking and left
without being tested. One subject could not find child care. Each time a subject cancelled or
failed to show for saliva collection, another subject was contacted. At the end of a two-week
period of saliva collection, 2 subjects cancelled, leaving a total of 28 saliva samples collected.
The reduction of desired study subjects in Aim 3 was approved by the investigator’s Dissertation
Committee.
Demographics and Patient Characteristics for Genotyped Subjects
The mean BMI for the genotyped subjects was similar to the mean study population BMI
at 30.01. There was no statistically significant difference in the BMI of the study population and
the genotyped subjects using t-test (p. = .766). Males comprised 53.6% (n = 15) of the

110

genotyped subjects. The average age of genotyped subjects was 53.36 (S.D. = 15.2), not
statistically different from the study population as a whole (p. = .806). Ninety-three percent (n =
26) of the genotyped subjects were non-smokers compared to 71.8% of the study sample subjects
and this was significantly statistically different (x2 = 7.415; p. = .006). Similar to the study
population, 57% (n = 16) of the genotyped subjects was working. This was not statistically
different from the study population as a whole (x2 = 2.538; p. = .143). Among the genotyped
subjects, 68% (n = 19) were covered by commercial insurance, 18% (n = 5) had Medicare, 7% (n
= 2) had Medicaid, 3.5% (n = 1) had worker’s compensation, and 3.5% (n =1) had no insurance.
See Table 28 for genotyped subjects, N and % for categorical variables.
Eighteen percent (n = 5) of genotyped subjects had been diagnosed with depression (by
review of the medical record) compared to 31.3% of the study population, but this was not
statistically different (p. = .118). The mean pain VAS among the genotyped subjects was 6.34
compared to the mean pain VAS of 6.83 for the study sample (S.D. = 1.96), not a statistically
significant difference (p. = .200). See Table 28 for genotyped subjects, N and % for categorical
variables. See Table 29 for genotyped subjects patient characteristics, range, minimum,
maximum, mean and SD for continuous variables. See Table 30 for genotyped subjects’ pain
VAS mean, maximum, minimum.

111

Table 28
Genotyped Subjects Patient Characteristics, N and % for Categorical Variables
(N = 28)
Category
Physiological

Characteristic
Smoking

Sex

Situational

Work Status

Insurance

Variable

N

%

Non-smoking

26

93

Smoking

2

7

Female

13 46.4

Male

15 53.6

Working

16

57

Non-working

12

43

Commercial

19

68

Medicare

5

18

Medicaid

2

7

Worker’s Compensation

1

3.5

No Insurance

1

3.5

Depressed (per medical

5

18

23

82

Type

Auto

Psychological

Depression

record)
Not Depressed

112

Table 29
Genotyped Subjects Patient Characteristics, Range, Minimum, Maximum, Mean and SD
for Continuous Variables (N = 28)
Category

Characteristic

Range

Sample Mean

SD

BMI

20.68-71.04

30.01

9.75

Age

22-80

53.6

15.239

Physiological

Table 30
Genotyped Subjects Symptom Continuous Variable: Pain VAS (N = 28)
Symptom

Range

Mean

SD

Pain VAS for Genotyped Subjects

1.5-10

6.34

1.963

Outcome Measures for the Genotyped Subjects
The ODI score was missing for one of the genotyped subjects. The mean ODI score for
the genotyped subjects was 45.56 (S.D. = 17.88). The mean physical function subscale score for
the genotyped subjects was 35.76 (S.D. = 11.05), and the mean mental health subscale score for
the genotyped subjects was 52.52 (S.D. = 19.47). There was no statistically significant
difference between the study population and the genotyped subjects for ODI and PF scores (p. =
.111 and p. = .096, respectively). However, the difference between mental health subscale scores
between the two populations was statistically significant (p. = .001). Thus, the mental health
scores were significantly better for the genotyped subjects than for the study population as a
whole. The study population mean mental health subscale score was 44.73 and published health
population norm score is 53.43. See Table 31 for physical function outcome scores for the ODI
and the SF-36 for the genotyped subjects.

113

Table 31
Genotyped Subjects Physical Function Scores: Range, Minimum, Maximum, Mean and
SD (N = 28)
Outcome Measure

N

Range

Mean

SD

ODI

27

4-74

45.56

17.85

SF-36 Physical Function

28

14.94-57.03

35.76

11.05

Subscale

In summary, the genotyped subjects did not differ significantly in the areas of BMI, sex,
age and employment status. Fewer genotyped subjects smoked and fewer were depressed than in
the study population. The mean pain VAS was slightly lower in the genotyped subjects, but this
was not statistically significant. The insurance types were slightly differently represented among
the genotyped subjects, with slightly more subjects with commercial insurance than in the study
sample.
Although the genotyped subjects had slightly better ODI scores than the study population
as a whole, this was not statistically significant. Physical function subscale scores were similar,
but this was not statistically significant. The genotyped subjects had statistically significantly
better mental health subscale scores than the study population as a whole.
Genotyping Results
Genotyping was performed by staff at the Core Genomics Lab at Michigan State
University. The 28 saliva samples were tested for COL9A2 (rs2228564), COL9A3 (rs61734651),
OPRM1 (rs1799971), COMT (rs4680), VDR (rs731236) and for VNTR for ACAN. All 28 saliva
samples yielded valid testing results for the genes being tested, with the exception of ACAN.
Two saliva samples did not amplify for ACAN VNTR testing, and were not able to be
successfully genotyped, leaving 26 valid results for ACAN.

114

The genotyping results for COL9A2 revealed that 19 out of 28 individuals were
homozygous for the A/A allele, 8 were heterozygotes with the A/G allele, and one was
homozygous for the G/G allele. Therefore, 9 subjects possessed the Trp2 */G allele associated
with a higher rate of disc degeneration. The genotyping results for COL9A3 revealed limited
diversity, with 27 of the subjects homozygous for the C/C allele, and one heterozygous C/T
subject. Therefore, one subject possessed the Trp3 */T allele associated with a higher rate of
disc degeneration. Testing for OPRM1 revealed little diversity as well, with 24 subjects
homozygous for the A/A allele, one homozygous for the G/G allele, and three heterozygotes.
Results for COMT revealed greater diversity in genotype, with 12 A/A homozygotes, 5 G/G
homozygotes and 11 heterozygotes. Fourteen subjects were heterozygous C/T for VDR, 13 were
homozyogous T/T, and one was homozygous C/C.
The VNTRs for ACAN varied from 24 to 30 repeats, with seven different alleles
identified. In addition, seven different genotypes were identified, including 24/27, 27/29, 28/28,
28/30, 29/29, 30/30 and 30/33. Most subjects were homozygous, but four subjects were
heterozygous for ACAN VNTR. In summary, there was very little diversity represented in the
COL9A3, OPRM1 and ACAN genotypes. The other genes tested exhibited greater diversity in
genotype. Diversity in genotype has been shown to vary by ethnic group. However, the small
sample size of genotyped subjects did not allow for sufficient data to compare study genotype
representation with known populations. Also, the ethnicity of study subjects was not explored in
this study.
Aim 3 Analysis
For Aim 3, to explore the impact of the physiological factor genotype on symptoms, first
the genotypes for COL9A2, COL9A3, OPRM1, COMT, VDR and ACAN VNTR from the 28

115

subjects with saliva samples were each analyzed for their effects on symptoms. Next, the
relationship between genotype and outcome measures was examined.
Relationship between genotype and symptoms.
Genotypes COL9A2, COL9A3, OPRM1, COMT and VDR were examined for their effects
on pain VAS using one-way analysis of variance. None of the genotypes were found to have a
significant effect on pain VAS. However, OPRM1 exhibited a trend toward higher pain VAS in
individuals who were A/A, with lower pain scores for those A/G, and the lowest scores for G/G
individuals, although this was not statistically significant (p. = .201). When analyzed as a
dichotomous variable, OPRM1 continued the trend toward a significant association with pain
VAS, but was still not statistically significant (p. = .108). See Table 32 for analysis of variance
results of the genotypes COL9A2, COL9A3, OPRM1, COMT and VDR with pain VAS.
A scatter plot to determine any trends in the relationship between ACAN VNTR alleles
and pain VAS was analyzed. There was no observable linear trend between ACAN VNTR alleles
and pain VAS. There was no significant correlation between ACAN VNTR alleles and pain VAS
(r2 = -.047, p. = .821).
To explore the relationship between genotype and pain location, chi-square tests were
used for COL9A2, COL9A3, OPRM1, COMT, VDR and ACAN VNTR alleles and back pain,
back pain and leg pain and leg pain only. There was insufficient evidence to conclude that there
was a relationship between genotypes COL9A2, COL9A3, OPRM1, COMT, VDR and ACAN
VNTR alleles and pain location.

116

Table 32
One-way ANOVA for Between Groups Difference (Genotypes) COL9A2, COL9A3, OPRM1,
COMT and VDR and Pain VAS (N = 28)
Genotype

COL9A2
rs2228564

COL9A3
rs61734651

OPRM1
rs1799971

COMT
rs4680

VDR
rs731236
OPRM1
rs1799971
Dichotomous
A/A, */G

Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total

Sum of
Squares

df

Mean
Square

6.704

2

3.352

97.322

25

3.893

104.027

27

.453

1

.453

103.574

26

3.984

104.027

27

12.527

2

6.263

91.500

25

3.660

104.027

27

10.274

2

5.137

93.753

25

3.750

104.027

27

9.791

2

4.895

94.236

25

3.769

104.027

27

10.026

1

10.006

94.021

26

3.616

104.027

27

Dependent Variable: Pain VAS
df: Degrees of Freedom
Sign: Significance
COL9A2 Groups: A/A, A/G, G/G
COL9A3 Groups: C/C, C/T
OPRM1 Groups: A/A, A/G, G/G
COMT Groups: A/A, A/G, G/G
VDR Groups: C/C, C/T. T/T

117

F

Sign.

.861

.435

.114

.739

1.711

.201

1.370

.273

1.299

.291

2.767

.108

Because of the lack of diversity represented in the genotypes COL9A2 and OPRM1, these
genotypes were also tested with pain location as dichotomous variables (A/A and */G). There
was insufficient evidence to conclude that there was a relationship between COL9A2 and
OPRM1 and pain location, testing these genotypes as dichotomous variables. See Table 33 for
chi-square tests for relationships between genotype and pain location (back pain only, back pain
and leg pain, leg pain only). Similarly, using chi-square testing, there were no statistically
significant relationships between genotypes and the symptoms of numbness and weakness.
Table 33
Chi-square Tests for Genotype as Predictors of Pain Location (Back Pain Only, Back Pain
and Leg Pain, Leg Pain Only) (N = 28)
Genotype
COL9A2
rs2228564
COL9A3
rs61734651
OPRM1
rs1799971
COMT
rs4680
VDR
rs731236
ACAN VNTR alleles
COL9A2
rs2228564
Dichotomous (AA/*G)
OPRM1
rs1799971
Dichotomous (AA/*G)

Chi-square

df

Sign.

2.574

4

.631

.899

2

.638

5.279

4

.260

2.133

4

.711

4.335

4

.363

17.011

12

.149

1.981

2

.371

1.254

2

.534

Dependent Variable: Pain location (Back pain only, back pain and leg pain, leg pain
df: Degrees of Freedom
Sign.: Significance
ACAN VNTR alleles: 24/27, 27/29, 28/28, 28/30, 29/29, 30/30, 30/33

118

only)

Relationship between genotype and outcome measures.
Because there were two outcome measures, the ODI and the physical function subscale
from the SF-36, separate statistical tests were used to explore the influence of patient genotype
on physical function. These tests will be reported separately.
First the genotypes COL9A2, COL9A3, OPRM1, COMT and VDR were analyzed using
one-way analysis of variance with ODI scores. COL9A2, COL9A3, COMT and VDR genotypes
were not significantly associated with ODI scores. OPRM1 genotype was significantly
associated with ODI scores. A/A OPRM1 genotype was associated with higher (worse) ODI
scores (F = 3.643, p. = .042). See Table 34 for one-way analysis of variance for genotype and
ODI scores.
A correlation was computed to test the relationship between ACAN VNTR alleles and
ODI scores. There was no significant correlation between ACAN VNTR alleles and ODI scores
(r2 = -.007, p. = .974). No significant linear trend between ACAN VNTR allele and ODI scores
was identified on a scatter plot.
Finally, the genotypes COL9A2, COL9A3, OPRM1, COMT and VDR were analyzed
using one-way analysis of variance with SF-36 physical function subscale scores. There were no
genotypes that were significantly associated with SF-36 physical function subscale scores.
Because of the lack of diversity represented in the genotype OPRM1, this genotype was also
tested as a dichotomous variable (A/A and */G) with SF-36 physical function subscale score.
Treated as a dichotomous variable, OPRM1 was significantly associated with SF-36 physical
function subscale scores, with A/A genotypes associated with lower (worse) physical function
scores (F = 4.511, p. = .043), similar to the findings with ODI scores. See Table 35 for one-way
analysis of variance for genotypes and SF-36 physical function subscale scores.

119

Table 34
One-way ANOVA for Between Groups Difference (Genotypes) COL9A2, COL9A3, OPRM1,
COMT and VDR and ODI Scores (N = 27)
Genotype
COL9A2
rs2228564

COL9A3
rs61734651

OPRM1
rs1799971

COMT
rs4680

VDR
rs731236

Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total

Sum of
Squares

df

Mean Square

F

223.167

2

111.583

.332

8059.500

24

335.813

8282.667

26

32.051

1

32.051

8250.615

25

330.025

8282.667

26

1929.043

2

964.522

6353.623

24

264.734

8282.667

26

584.800

2

292.400

7697.867

24

320.744

8282.667

26

367.590

2

183.795

7915.077

24

329.795

8282.667

26

Sign.

.721

.097
.758

3.643
.042

.912
.415

.557
.580

Dependent Variable: ODI scores
df: Degrees of Freedom
Sign.: Significance
COL9A2 Groups: A/A, A/G, G/G
COL9A3 Groups: C/C, C/T
OPRM1 Groups: A/A, A/G, G/G
COMT Groups: A/A, A/G, G/G
VDR Groups: C/C, C/T. T/T

A correlation was computed to test the relationship between ACAN VNTR alleles and SF36 physical function subscale scores. There was no significant correlation between ACAN
VNTR alleles and SF-36 physical function subscale scores (r2= .104, p. = .613). A scatter plot

120

did not reveal a linear relationship between ACAN VNTR alleles and SF-36 physical function
subscale scores.
In summary, there were no significant relationships identified between the genotypes of
the 28 subjects from Exploratory Aim 3 and symptoms, although there was a trend toward a
significant relationship between OPRM1 genotype and pain VAS. OPRM1 genotype was found
to have a significant relationship with ODI scores, and when treated as a dichotomous variable
(A/A and */G), OPRM1 was significantly associated with SF-36 physical function subscale
scores. Since the genotyped sample was exploratory, a small sample size was used. Although
the only significant finding was the association between OPRM1 and pain VAS, it would be
premature to dismiss potential associations with the other genotypes, because of the small sample
size used. A discussion of these results will be found in Chapter VI.

121

Table 35
One-way ANOVA for Between Groups Difference (Genotypes) COL9A2, COL9A3, OPRM1,
COMT and VDR and SF-36 Physical Function Subscale Scores (N = 28)
Genotype

COL9A2
rs2228564

COL9A3
rs61734651

OPRM1
rs1799971

COMT
rs4680

VDR
rs731236

OPRM1
rs1799971
Dichotomous
(A/A, */G)

Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total
Between
Groups
Within
Groups
Total

Sum of
Squares

df

Mean
Square

F

188.550

2

94.275

.758

3109.511

25

124.380

3298.062

27

77.475

1

77.475

3220.586

26

123.869

3298.062

27

487.963

2

243.982

2810.098

25

112.404

3298.062

27

97.990

2

48.995

3200.071

25

128.003

3298.062

27

416.883

2

208.441

2881.179

25

115.247

3298.062

27

487.594

1

487.594

2810.468

26

108.095

3298.062

27

Dependent Variable: SF-36 Physical Function Subscale Scores
df: Degrees of Freedom
Sign.: Significance
COL9A2 Groups: A/A, A/G, G/G
COL9A3 Groups: C/C, C/T
OPRM1 Groups: A/A, A/G, G/G
COMT Groups: A/A, A/G, G/G
VDR Groups: C/C, C/T. T/T
OPRM1 Dichotomous Groups: A/A, */G

122

Sign.

.479

.625
.436

2.171
.135

.383
.686

1.809
.185

4.511
.043

CHAPTER VI
Discussion and Implications
The focus of this study was on patient physiological, situational and psychological
characteristics and symptoms, and their combined effects on physical function for persons with
lumbar degenerative conditions. A discussion of the results with interpretation and how they
support or differ from existing research and limitations of this study will be presented in this
chapter by Aims. Last, this final chapter will present contribution to science and implications for
nursing practice, research, and policy.
Discussion of Sample Patient Characteristics
The sample patient characteristics included BMI, sex, age, smoking, employment status,
worker’s compensation claim, insurance type and depression. The symptoms studied included
pain VAS, pain location (low back pain only, leg pain only and back and leg pain combined),
extremity weakness and extremity numbness. Outcome measures included scores on the ODI
and the physical function subscale of the SF-36. Additionally, mental health subscale scores
from the SF-36 were also examined, to compare the study population to population norm scores.
All patient data except genotype was obtained from the medical records randomly chosen from a
database of patients seeking care from the spine service from 2009-2012. Genotype data was
obtained from a randomly selected subset of this study population in February, 2014.
Discussion of Sample Physiological Characteristics.
The mean BMI of the sample was 30.44 (S.D. = 8.41), considered obese (National
Institutes of Health, 2014).

In fact, nearly 30% (n = 48) of subjects in the sample were

overweight, nearly 30% (n = 48) were considered obese and over 12% (n = 20) were considered
class III obese. Nearly 42% (n = 116) of the entire sample had a BMI of 30 or greater, compared

123

to the state of Michigan average obesity rate of 31.1% in 2012 (CDC, 2014).

The study

population was therefore heavier than the Michigan average.
Fifty-eight percent (n = 91) of the sample was women. There were 72 men (n = 42) in
the sample. Males and females were approximately equally represented in the study. The sample
mean age was 54 (S.D. = 16.85). Seventy-three percent (n = 119) were younger than 65. The
age range of the sample was 22-93.
More than 28% (n = 46) of the sample smoked. Thirty-five percent (n = 24) of men were
smokers, while 23% (n = 22) of the women smoked. This compares to the smoking rate of
American adults, which is 18.1% (CDC, 2014). Among American adult males, 20.5% are
smokers. Among American adult women, 15.8% are smokers. Thus, the sample smoking rate
was higher than the American average rates for adults, both men and women.
Discussion of Sample Situational Characteristics.
More than half of the study subjects were not working at the time of their presentation to
the spine service (56%, n = 92). Among men, 52% (36) were not working. Sixty percent of
women were not working (56). And, among those 119 subjects of working age (less than 65),
46% (55) were not working. It is undetermined whether work status in this population was
directly related to a spinal cause.
Only 3 individuals from the sample had worker’s compensation claims and all three
subjects with worker’s compensation claims were not working. This represented only 1.84% of
the sample, thus making it difficult to make conclusions regarding its influence on symptoms and
physical function in Aim 1. Additionally, because of the small proportion of the sample with
worker’s compensation, this variable did not play a significant role in the predictive modeling of
Aim 2.

124

The majority of the sample (n = 147, approximately 90%) was covered by three types of
insurance plans: commercial, Medicare and Medicaid. Commercial insurance covered 57% (93)
of the sample, followed by Medicare (20.9%, or 34 subjects). Twenty subjects (12.3 %) had
Medicaid insurance, and 8 subjects (4.9%) had no insurance. Three subjects (1.8%) had
worker’s compensation, 3 subjects (1.8%) had auto insurance, and two (1.2%) had tricare, an
insurance plan for those in the armed service. As would be expected, of those working, the
percentage of those covered by commercial insurance rose to 87%, (n = 62) with 4% (n = 3)
covered by Medicare, 4% (n = 3) covered by Medicaid, 3% (n = 2) with no insurance and 1
person with auto insurance. Most subjects were covered by some form of insurance. Given the
nature of the clinical condition of this population (low back pain), the low proportion of subjects
covered by worker’s compensation was unexpected.
Discussion of Sample Psychological Characteristic.
More than 31% (n = 51) of the sample had a clinical diagnosis of depression obtained
from review of the medical record received from the referring physician at the time of the first
visit to the spine service. More women (38%, n = 36) were diagnosed with depression than men
(22%, n = 15), a difference that was statistically significant. SF-36 mental health subscale scores
and the diagnosis of depression were related (t-statistic = 4.180, p. = .000; Mann-Whitney U test
p. = .000). And, SF-36 mental health subscale scores were significantly lower for those with the
diagnosis of depression. The mean sample mental health score was 44.73 (S.D. = 14.4), lower
than healthy population norm score (53.43 S.D. = 8.38) (Ware et al., 2007). The sample mean
mental health subscale score was higher than the published mean score for those with depression,
which is 36.7 (S.D. = 11.08) (Ware et al., 2007). Thus, the sample mean mental health was
worse than a healthy population, but not as low as scores for a population with depression. It is

125

possible that the study sample possessed more associated factors that may have impacted mental
health subscale scores. These factors may include the higher rates of smoking, more severe
estimations of pain, and higher rates of obesity than average. Medical co-morbidities were not
explored in this study, and it is possible that the burden of medical co-morbidities could have
contributed to depression in the subjects in this study.
In summary, the sample (n = 163) consisted of proportionately more obese individuals
than the state of Michigan average. Slightly more females than males were represented in the
sample. The age range of subjects in the sample was 22-93, with a mean age of 54 (S.D. =
16.85). The majority of subjects were younger than 65 (73%, n = 119). There were
proportionately more smokers in the sample than the American average. This was true for both
men and women.
Even though the majority of subjects in the sample were younger than 65 (n = 119), the
majority of subjects were not working (56%, n = 92). In fact, of those who were younger than 65
(n = 119), 46% (n = 55) were not working. Only 3 of the 163 subjects were covered by worker’s
compensation, which made it difficult to fully assess the influence this type of insurance had on
symptoms and physical function. By far, the most common insurance for the sample was
commercial, followed by Medicare and Medicaid. There were more subjects without insurance
than there were subjects with worker’s compensation, auto, or tricare.
Nearly one third of the sample had a diagnosis of depression obtained from review of the
medical record. There were more depressed females than depressed males. The sample mean
SF-36 mental health subscale score was lower than a healthy population norm score, but not as
low as the mean score for those with depression (Ware et al., 2007). Thus, the sample mean
mental health was worse than a healthy population. SF-36 mental health subscale scores were

126

significantly related to the diagnosis of depression, helping to strengthen the validity of this
variable.
Discussion of Symptoms of Sample
The mean pain VAS for the sample was 6.83 on a 0-10 scale (S.D. = 2.2). Males and
females were similar with regard to pain VAS, with males reporting a mean pain VAS of 6.79
(S.D. = 2.46) and females reporting a mean pain VAS of 6.85 (S.D. = 1.99), a difference that was
not statistically significantly different. Thus, the mean pain VAS for the sample approached
severe pain, according to cut points identified by Jensen, Smith, Ehde & Robinsin, (2001),
Kathy, Harris, Hadi and Chow (2007), and Serlin, Mendoza, Nakamura, Edwards and Cleeland,
(1995). It is possible that the associated factors in this population, such as higher than average
BMI, higher than average smoking rates and worse than average estimation of mental health may
play a role in the subjects’ estimation of pain. Pain treatments for the subjects were not known,
and were beyond the scope of this study.
Twenty-seven subjects reported weakness (16.6%). Sixty-five subjects reported
numbness (40%). One hundred subjects (60%) reported back pain and leg pain. Thirty-nine
subjects (24%) reported back pain only. Twenty-three subjects (14%) reported leg pain only.
Subjective numbness was reported by more subjects than weakness.
More subjects reported back and leg pain together than those reporting either pain
location alone. There are studies that have examined the pain diagrams of subjects with specific
pathologies (lumbar radiculopathy, sacro-iliac joint pain and facet joint pain). However, this
study included all individuals who presented to the spine service for care, regardless of the
specific anatomic diagnosis. The presence of back and leg pain together was associated with
worse physical function than back pain alone or leg pain alone, consistent with previous findings

127

(Kongstead, Kent, Albert, Jensen & Manniche, 2012). Future studies should explore further the
associations between pain location and physical function. This will be discussed in the
Implications for Research section.
Discussion of Sample Outcome Measures
For ODI scores, higher values reflect worse physical function. The sample mean ODI
subscale score was 50.96 (S.D. = 19.3) (range 0-90). Scores of 40-60 are associated with severe
disability (Fairbank, Couper, Davies & O’Brien, 1980). Published norm scores for individuals
with spinal metastases is 48.04 (Roland & Fairbank, 2000). Thus, the sample mean scores
reflect worse physical function than scores associated with spinal metastases. This was an
unexpected finding, and may be related to the other factors associated with physical function in
this study, including BMI, smoking, pain VAS and numbness.
For SF-36 subscales, higher scores reflect better physical function. The mean SF-36
physical function subscale score for the sample was 32.57 (S.D. = 11.98, range: 14.94-57.03).
Published healthy population norm score is 54.76 (S.D. = 6.04) (Ware et al., 2007). For
comparison, published mean score for individuals with back pain and sciatica is 46.78 (S.D. =
11.14) and mean scores for individuals in the 25th percentile with back pain and sciatica is 40.87
(S.D. = 11.14). The sample mean SF-36 physical function subscale score was worse than mean
scores for those with diabetes (42.52, S.D. = 11.18) and was only slightly better than scores for
individuals in the 25th percentile with cancer (30.64, S.D. = 11.52) (Ware et al., 2007).
In summary, the physical function scores from both the ODI and the SF-36 physical
function subscale indicate the sample was experiencing significant reduction in physical
function. Moreover, the ODI scores were lower than for those in the 25th percentile for those

128

with similar lumbar diagnoses, and only minimally better than those with cancer, a condition
with more serious health implications. Given the similarity of underlying diagnoses with those
populations from whom the norm scores were obtained, the explanation for worse physical
function scores in the study population is unclear. It is possible that the study population was
overall more obese, had a higher rate of smoking and higher subjective ratings of pain that
affected physical function scores than the populations used to determine the SF-36 population
norm scores for lumbar pathologies. Co-morbidities were not examined in this study, but may
have been a factor in the subjects’ estimations of physical function.
Discussion of Results for Specific Aim 1
The purpose of Specific Aim 1 was to determine the contribution of physiological (BMI,
sex, age, smoking status), situational (employment status, worker’s compensation claim,
insurance type), and psychological (depression) factors in persons with degenerative lumbar
conditions to symptoms and physical function. First, the relationship between patient
characteristics and symptoms will be examined. Next, the associations between patient
characteristics and physical function will be reviewed.
Discussion of Associations between Patient Characteristics and Symptoms.
Multiple regression was used to explore the relationship between BMI, sex, age, smoking
status, employment status and depression and pain VAS. Smoking was the only significant
predictor of pain VAS, explaining 8.6% of the variance. That is, smoking was weakly associated
with worse pain VAS scores, but more than 90% of the variance in pain VAS was not explained
by the variables BMI, sex, age, smoking, employment status and depression. Since BMI and
employment status were the last variables to be eliminated in the first regression, they were kept
in the multiple regression model while insurance types were added to determine the effects on

129

pain VAS. In the final multiple regression model, smoking, having Medicaid insurance, or not
having insurance were associated with worse pain VAS, explaining 13% of the variance in pain
VAS scores. That is, smoking, having Medicaid insurance, or not having insurance were
associated with worse pain VAS scores, leaving 87% of the variance unexplained by BMI,
employment status, smoking, and insurance type. Thus, the variables included in the regression
were not sufficient to explain a large portion of the variance, or, the sample may not have been
large enough. The finding that smoking is related to pain is consistent with previous findings
that associate smoking with higher levels of back pain, but the mechanism behind this is unclear
(Karahan, Kav, Abbasoglu & Dogan, 2009; Shiri, Karppinen, Lein-Arjas, Solovieva, & ViikariJuntura, 2010). And the associations between Medicaid insurance or not having insurance and
pain VAS are consistent with previous findings that lack of insurance or under-funded insurance
are associated with worse health outcomes in general (Greenstein, Moskowitz, Gelijns &
Egorova, 2012; Kruper, et al., 2011; McClelland, Guo & Okuyemi, 2011; Yorio, Yan, Xie &
Gerber, 2012). This finding may be due to a restricted number of physical therapy visits offered
by Medicaid insurance. At the time these data were collected, local county Medicaid plans also
required attendance at a 6 hour class on pain before spine injections were authorized, which
served as a deterrent for some subjects in obtaining this treatment.
Next, the relationship between patient characteristics and the symptoms of numbness and
weakness was tested. While age was the only significant variable in the final logistic regression
model for weakness, it had no discriminatory value. And while employment status was the only
significant variable in the final logistic regression model for numbness, it also had no
discriminatory value. Therefore, it was concluded that BMI, sex, age, smoking, employment
status and depression had no influence on the symptoms of numbness or weakness. It is likely

130

that these patient characteristic variables are not related to the symptoms of numbness and
weakness.
Since pain location was a categorical variable, discriminant analysis was used to
determine if patient characteristics (BMI, sex, age, smoking, employment status and depression)
could predict pain location (back pain only, back and leg pain, leg pain only). Age and smoking
were significant predictors of pain location, but age alone accounted for only 8% of the variance,
and age and smoking together accounted for only 5% of the variance. Smoking is associated
with higher levels of back pain and spinal degeneration increases with age. While there are no
studies exploring smoking with location of pain in individuals with spinal degeneration, it is
possible that higher estimations of pain, including more widespread estimations of pain location
may be related in smokers. Using analysis of variance, BMI was not found to influence pain
location. With chi-square testing, sex, depression and worker’s compensation were not related to
pain location (only 3 subjects had worker’s compensation insurance). There were no patient
characteristic variables that were sufficient to explain pain location (back pain only, back pain
and leg pain, leg pain only). It is likely that these variables also had no influence over pain
location.
Discussion of Associations between Patient Characteristics and Physical Function.
Since two separate measures of physical function were available in the database to
measure physical function, separate statistical testing was conducted with the ODI and the
physical function subscale of the SF-36. These results will be described in the following section.
First, the patient characteristics of BMI, sex, age, smoking, employment status and
depression were examined with ODI scores as the dependent variable. Because the variable
insurance type had several levels, this was added in a subsequent step. Initially, BMI, smoking

131

and employment status were significant, explaining 15.4% of the variance in ODI scores. Higher
BMI and smoking were associated with worse ODI scores, while being employed was associated
with better ODI scores. Next, insurance type was added, with the previously significant
variables of BMI, smoking and employment status. BMI, smoking, having commercial or
Medicare insurance were all significant, explaining 18% of the variance in ODI scores. Higher
BMI and smoking were associated with worse ODI scores, while having commercial or
Medicare insurance were associated with better ODI scores.

These findings are consistent with

previous findings, that higher BMI and smoking are associated with worse physical function
(Prasarn, Horodyski, Behrend, Wright & Rechtine, 2012; Rihn et al, 2013). Also, having
insurance has been associated with better health outcomes in general, consistent with the findings
in this study (Greenstein, Moskowitz, Gelijns & Egorova, 2012; Kruper, et al., 2011;
McClelland, Guo & Okuyemi, 2011; Yorio, Yan, Xie & Gerber, 2012). However, the variables
studied explained only 18% of the variance, leaving 82% of the variance in ODI scores
unexplained. The sample size may have been too small to detect larger effects with these
variables.
The combined effects of BMI, sex, age, smoking, employment status and depression were
examined for their effects on SF-36 physical function subscale scores. BMI, age, employment
status and depression were significant in the final model, explaining 17.7% of the variance in SF36 physical function scores. Specifically, higher BMI, older age and depression were associated
with worse SF-36 physical function subscale scores, while being employed was associated with
better SF-36 physical function subscale scores. Next, insurance type was added, with the
previously significant variables of BMI, older age, depression and employment status. In the
final model, BMI, age, smoking and having Medicaid insurance were significant, predicting

132

18.2% of the variance in SF-36 physical function subscale scores. Specifically, higher BMI,
older age, smoking, and having Medicaid insurance were all associated with worse SF-36
physical function subscale scores. However, the variables studied explained only 18.2% of the
variance, leaving almost 82% of the variance unexplained by the study variables. It is possible
that the sample size was too small to detect a larger effect.
The factors found to be associated with physical function in this study include BMI,
smoking, age, and the insurance types of Medicaid, Medicare or Commercial insurance. The
factors common to both ODI and SF-36 physical function subscale scores are higher BMI and
smoking, having deleterious effects on both measures of physical function. The findings that
BMI and smoking have a negative effect on physical function in this population are consistent
with previous research (Prasarn, Horodyski, Behrend, Wright & Rechtine, 2012; Rihn, et al.,
2013). Medicare and Commercial insurance is positively associated with physical function in
this study. It is likely that Medicare and Commercial insurance plans have better coverage for
interventions such as physical therapy and spinal injections that improve physical function in this
population. The finding that Medicaid insurance is associated with worse physical function is
also likely due to reduced coverage for physical therapy and spinal injections.
The factors associated with ODI and SF-36 physical function subscale scores did differ.
While Medicare and Commercial insurance were positively associated with ODI scores, they
were not significant for SF-36 physical function subscale scores. And, while older age and
Medicaid insurance were negatively associated with SF-36 physical function scores, they were
not significant for ODI scores. This difference may be due to the ODI being a lumbar-specific
physical function measure and the SF-36 being a generic measure of overall well-being.

133

To summarize, the study variables smoking, having Medicaid insurance, and not having
insurance were weakly associated with pain VAS, explaining 13% of the variance. There were
no patient characteristic variables that were sufficient to explain the symptoms of numbness and
weakness. It is likely that these variables had no influence over the symptoms of numbness,
weakness, or pain location. Finally, there were no patient characteristic variables that were
sufficient to explain pain location (back pain only, back pain and leg pain, leg pain only). It is
possible that these variables also had no influence over pain location.
Higher BMI, smoking, older age and having Medicaid insurance were all associated with
worse physical function scores, while having commercial insurance was associated with better
physical function scores. These findings are expected because they are consistent with previous
literature associating higher BMI and smoking with worse physical function (Prasarn,
Horodyski, Behrend, Wright & Rechtine, 2012; Rihn, et al., 2013) and indigent insurance plans
with worse health outcomes in general (Greenstein, Moskowitz, Gelijns & Egorova, 2012;
Kruper, et al., 2011; McClelland, Guo & Okuyemi, 2011; Yorio, Yan, Xie & Gerber, 2012).
Finally, since a very small number of subjects had worker’s compensation insurance, it
was not possible to obtain findings that supported the literature associating worker’s
compensation with worse physical function outcomes for individuals with lumbar degenerative
conditions.
Discussion of Additional Results
Analysis of variance was computed to determine if location of pain was associated with
worse physical function. After multiple comparison testing, the presence of back and leg pain
was associated with worse ODI scores. However, using analysis of variance and multiple
comparison testing, the presence of back and leg pain was not significantly associated with

134

worse SF-36 physical function scores. These findings partially support previous findings that
have associated the presence of back and leg pain with worse physical function (Kongstead,
Kent, Albert, Jensen & Manniche, 2012). Since the ODI is a lumbar-specific measure of
physical function, it is possible this instrument is more sensitive than the physical function
subscale of the SF-36 to the relevant factors that influence physical function in this population,
hence the significant association between the presence of back and leg pain together and worse
physical function for the ODI, but not for the SF-36 physical function subscale.
Discussion of Results for Specific Aim 2
The purpose of Specific Aim 2 was to develop a predictive model for physical function in
persons with lumbar degenerative conditions, using symptoms (back and/or leg pain, numbness,
and weakness) and physiological, situational, and psychological patient factors. Since there were
two measures of physical function, the ODI and the physical function subscale of the SF-36,
separate statistical tests were used. First, the results of testing with ODI scores as the dependent
variable will be discussed.
Using the significant predictors for ODI score from the Aim 1 analysis (smoking and
BMI), symptoms were added into an analysis of variance. With general linear modeling, BMI,
smoking, pain VAS and extremity numbness were significantly associated with ODI scores,
explaining almost 35% of the variance. Specifically, higher BMI, smoking, higher pain VAS
and the presence of extremity numbness was associated with worse ODI scores. The association
between higher BMI, smoking, and higher pain level are consistent with previous research
(Prasarn, Horodyski, Behrend, Wright & Rechtine, 2012; Rihn et al, 2013). The finding of
extremity numbness associated with worse ODI scores is unexpected, and the explanation for
this is unclear. It is possible that numbness of the lower limb and foot impairs proprioception,

135

thereby interfering with mobility. A search of the literature revealed no studies exploring the
relationship of extremity numbness and physical function in individuals with lumbar
degenerative conditions. Since the ODI is a lumbar-specific measure of physical function, it is
possibly more sensitive to the factors that may impact physical function in this population.
Numbness of the lower extremity may be a relevant symptom for its effects on physical function
in individuals with lumbar degenerative conditions.
Next, the significant predictors for SF-36 physical function scores from Aim 1 analysis
(BMI, age and smoking) were added into an analysis of variance with symptoms. BMI, age and
pain VAS were significant and kept in the final model. These variables were entered into a
backwards step-wise multiple regression with the SF-36 physical function subscale the
dependent variable. BMI, age and pain VAS were significantly associated with worse SF-36
physical function subscale scores, explaining 26% of the variance. Specifically, higher BMI,
older age and higher pain level are associated with worse SF-36 physical function subscale
scores. These findings are consistent with previous research associating higher BMI and higher
pain level with worse physical function (Prasarn, Horodyski, Behrend, Wright & Rechtine, 2012;
Rihn et al, 2013). Older age has been associated with greater degenerative changes in the lumbar
spine (Cheung, et al, 2009; Kalichman, Kim, Li, Guermazi & Hunter, 2010).
Higher BMI was a relevant factor in predicting worse physical function measured by both
the ODI and the SF-36 physical function subscale. This finding was expected and consistent
with previous literature (Prasarn, Horodyski, Behrend, Wright & Rechtine, 2012; Rihn et al,
2013). Higher BMI is associated with back pain, also a significant predictor of physical function
in this study, as measured by both the ODI and the SF-36 physical function subscale. Smoking
and extremity numbness were significant predictors for worse ODI scores but not worse SF-36

136

physical function subscale scores. Smoking is associated with higher levels of back pain, also a
significant predictor of worse ODI scores in this study. Smoking also may have an effect on
physical function through its effect on pulmonary function, a relationship not explored in this
study. Lower extremity numbness may be a relevant factor for physical function in individuals
with lumbar degenerative conditions, and detectable only by ODI scores because the ODI is a
lumbar-specific physical function measure. Age was a significant predictor for worse SF-36
physical function subscale scores, but not for ODI scores. Older age is associated with more
degenerative lumbar changes (Cheung, et al., 2009). In summary, the significant predictors for
reduced physical function in this study are expected, and some variables are known to affect the
others. The discrepancy between the significant predictors of ODI scores and SF-36 physical
function scores may be related to the particular sensitivities of each instrument.
The findings of this study add to science by connecting the significant patient
physiological, situational and psychological characteristics to symptoms and to physical function
in a population affected by lumbar degenerative conditions. Much of the medical literature with
this population does not consistently consider the impact of symptoms on physical function. By
identifying the relevant physiological, situational and psychological factors that combine with
symptoms to predict physical function, a risk assessment may allow early identification of
individuals with characteristics that place them at risk for poorer physical function. Moreover,
by studying symptoms with other well-studied patient characteristics for their combined effects
on physical function, this study considers how the patient perception of symptoms affects
physical function. Using patient-reported information and patient identified priorities in the plan
of care with the identified risk factors may transform a standardized model of care for patients
with lumbar degenerative conditions to an individualized approach to care.

137

This study also illustrates the need for multiple instruments to adequately measure
physical function in this population. The multiplicity of patient factors significantly associated
with physical function in this study also highlights the complex nature of spinal degenerative
conditions and their effects on patients. Last, this study illustrates the need for development of
instruments to capture the patient experience of symptoms particular to lumbar degenerative
conditions, including the dimensions of distress, duration, quality and intensity. More robust
data regarding the symptoms experienced by this population will enhance the understanding of
the relationship of symptoms to physical function in this population.
Discussion of Study Results for Exploratory Aim 3
The purpose of Exploratory Aim 3 was to explore the impact of the physiological factor
genotype (disc structural genes and pain genes) on symptoms (back and/or leg pain, numbness,
and weakness) and on physical function (ODI scores and SF-36 physical function subscale
scores). First, the relationship between genotype and symptoms will be examined. Next,
associations between genotype and physical function will be reviewed.
Discussion of Associations between Genotype and Symptoms.
Using analysis of variance, COL9A2, COL9A3, OPRM1, COMT and VDR SNPs were
analyzed for their associations with pain VAS. There were no statistically significant
relationships between COL9A2, COL9A3, OPRM1, COMT and VDR SNPs and pain VAS. There
was a trend toward higher pain VAS in individuals homozygous for A/A OPRM1 genotype,
lower pain VAS for individuals who were A/G, and the lowest pain VAS for individuals
homozygous for G/G. Although this finding was not statistically significant, the trend is
consistent with other findings associating the *G OPRM1 allele with decreased pain sensitivity
in men after lumbar disc herniation (Olsen et al, 2012). However, the literature supporting this is

138

inconsistent. Some studies have identified a sex specific interaction with the *G allele and pain
sensitivity in individuals with lumbar disc herniation, with *G males experiencing reduced pain
intensity and *G females experiencing increased pain intensity after disc herniation (Hasvik,
Schistad, Grovle, Haug, Roe & Gjerstad, 2014; Olsen et al, 2012). The findings from this
exploratory Aim are consistent with some previous findings in the literature, and larger sample
size would allow for further exploration of the sex interaction with genotype in this population.
ACAN VNTR was tested for association with pain VAS using correlation. No
statistically significant relationship was found between ACAN VNTR and pain VAS (r2 = -.047,
p. = .821).
Using chi-square, COL9A2, COL9A3, OPRM1, COMT and VDR were tested for
associations with pain location (back pain only, back pain and leg pain, leg pain only) and
numbness and weakness. There were no statistically significant associations between genotype
and pain location, numbness or weakness.
When the relationship between ACAN VNTR and pain location (back pain only, back and
leg pain, leg pain only) was analyzed with chi-square, no significant association was identified
(x2 = 17.011, p. = .149). It is likely that the genotype sample size was too small to identify
significant associations between SNPs and symptoms, given the seven different ACAN VNTR
alleles identified in the genotyped subjects. Also, there was limited variability in SNPs
represented within the COL9A3 and OPRM1 genes, thus not allowing for adequate comparison
of diverse SNPs with phenotype. Further study with larger sample sizes could reveal more
significant findings. In particular, future questions regarding the representation of different
alleles in larger populations would improve understanding of the relationship between genotype
and phenotype. Future studies should illuminate which candidate genes exert the most

139

significant influences on symptoms and physical function in this population. Other candidate
genes should be included in future studies to detect their contributions as well. Because
substances involved in disc degeneration have also been identified, genes encoding for these
substances should also be included in future studies, to explore their relationship to symptoms
and physical function in this population.
Last, there were no associations in the literature between genotype and numbness and
weakness. Thus, the findings of no significant association between genotype and numbness and
weakness are not unexpected.
Discussion of Associations between Genotype and Physical Function.
Since two separate measures of physical function were used, separate statistical tests
were conducted with the ODI and the physical function subscale of the SF-36. These results will
be described in the following section.
Using analysis of variance, COL9A2, COL9A3, OPRM1, COMT and VDR SNPs were
tested for association with ODI scores. There were no significant associations between COL9A2,
COL9A3, COMT or VDR SNPs and ODI scores. However, there was a significant association
between OPRM1 and ODI scores. Individuals with A/A genotype were found to have
significantly higher (worse) ODI scores than those with one or two copies of the G allele. And,
while there were no significant associations between COL9A2, COL9A3, OPRM1, COMT or
VDR SNPs and SF-36 physical function subscale scores with analysis of variance, when OPRM1
was analyzed as a dichotomous variable (A/A and */G), there was a significant association
between OPRM1 and SF-36 physical function subscale scores, consistent with the findings for
ODI scores. Individuals with A/A genotype (or no copies of the G allele) were found to have
significantly lower (worse) SF-36 physical function subscale scores than those with one or two

140

copies of the G allele (*/G). These findings are consistent with previous research suggesting a
difference between A/A and */G alleles and pain and physical function in individuals with
herniated lumbar discs. Olsen et al., (2012) identified a sex and genotype interaction that
influenced differences in individuals pain VAS and ODI scores over time after treatment for
lumbar disc herniation. While the */G males had greater improvements in pain VAS and ODI
scores after treatment for lumbar disc herniation than A/A males, */G females had the least
improvement in pain VAS and ODI scores compared to */G males, A/A males, and A/A females.
The small sample size (n = 28) limits the ability to examine sex and genotype interaction for
OPRM1, but these findings support the connection between pain and physical function in this
population.
Correlations were computed to test the relationship between ACAN VNTR alleles and
ODI and SF-36 physical function subscale scores. There were no significant correlations
between ACAN VNTR alleles and physical function.
In summary, the only gene found to have significant associations with physical function
was OPRM1, and the findings were partially consistent with what has been identified in the
literature. And, although not statistically significant, OPRM1 did show a trend toward
association with pain VAS. However, the literature is not consistent with regard to the genotype
universally associated with greater pain experience (Olsen, et al., 2012; Walter & Lotsch, 2009).
Study Limitations
The limitations of this study include the descriptive, cross-sectional design and the use of
secondary data. The cross-sectional design limits the ability to establish a temporal relationship
between the predictors and the outcome. Therefore, while associations between patient
physiological, situational and psychological factors and symptoms and physical function can be

141

determined by statistical analysis, the lack of a prospective design limits the conclusions
regarding the nature of temporal relationships between patient characteristics and symptoms and
physical function.
This study, though conducted using a random sample, reflects findings from one tertiary
spine center located in West Michigan. The findings therefore may not be generalizable to other
populations. Indeed, the study population measures of physical function were worse than
expected, and worse than populations with other similar and more severe lumbar conditions,
comparing study population physical function scores to scores of populations with spinal
metastatic disease and disc herniations.
The presence of depression was based on review of the medical record received from the
referring physician in this study and not measured directly. The validity of this variable and the
interpretation of its significance in this study were therefore limited.
Information regarding medical co-morbidities and their effects on symptoms and physical
function was not examined in this study. Medical co-morbidities may have accounted for some
of the unexplained variance in physical function in this study. This study is therefore limited in
its ability to explain all of the possible variables that may have affected physical function.
Although the organizing framework used was the TOUS, limited detail on symptoms was
explored in this study. Pain VAS, location of pain, numbness and weakness were the symptoms
studied. In reality, there may be more pertinent and influential symptoms that affect the outcome
of physical function in these subjects. Also, this study did not distinguish between those
individuals with acute lumbar spinal conditions or chronic spinal conditions. The nature of the
acuity of the condition may have affected the symptom experience and/or the outcome of
physical function.

142

While it was suggested by the author that an explanation for the association between
having Medicaid and not having insurance and worse physical function may have been lack of
coverage for evidence-based treatments such as physical therapy and spinal injections, this
relationship was not studied.
Data on patient characteristics, symptoms and physical function predates genotyping data
by as many as four years but this should not affect interpretation of results. Genotype does not
change over time and the subjects in this study possessed their genotypes at the time that data
were collected on patient characteristics, symptoms and physical function. The difficulty
experienced in contacting potential subjects for genotyping was largely related to persons not
answering phones and some phone numbers being disconnected. There was as many as five
years between some subjects presentation to the spine service for care and attempts to contact
those same subjects for genotyping. It is possible that those potential genotyping subjects with
disconnected or changed phone numbers represented a subset of individuals with lower
socioeconomic status and as such, may have affected the true randomness of the genotyping
sample.
With regard to genotype, the multiple comparisons performed may lead to identification
of significant results that are attributable to random chance. Therefore, the finding of a
significant association of OPRM1 genotype to physical function as measured by the ODI and SF36 physical function subscale scores should be interpreted with caution.
Last, while the list of analgesic, non-steroidal anti-inflammatory, anti-convulsant and
narcotic medications taken by subjects at the time of their initial evaluation at the spine service
was recorded, medication use could not be factored into the statistical analysis. The doses were
not consistently recorded in the medical record, nor were the frequency or last dose taken.

143

Information regarding medications taken for other medical co-morbidities were not recorded in
this study and may have influenced physical function. Therefore, the influence of medication on
symptoms and physical function could not be determined in this study. Medication use could
have influenced subjects’ estimations of their symptoms and their physical function.
Implications for Nursing Practice
The results of this study provided limited support for the usefulness of the TOUS to
organize the approach to study of the phenomena related to lumbar degenerative conditions.
There were significant associations between the physiological characteristic smoking and the
situational characteristic of insurance type (Medicaid and no insurance) and the symptom of pain
VAS. Two physiological characteristics (BMI and smoking) and one situational characteristic
insurance type (commercial and Medicare insurance) were found to be associated with physical
function (ODI scores). Two physiological characteristics (BMI and smoking) and one situational
characteristic insurance type (medicaid) were found to be associated with physical function (SF36 physical function subscale scores).
The patient characteristics of BMI, genotype and smoking and the symptoms of higher
pain VAS and extremity numbness were useful in predicting physical function as expressed by
ODI scores. The patient characteristics of BMI, genotype and age and the symptom of higher
pain VAS were useful in predicting physical function as expressed by SF-36 physical function
subscale scores. There were no associations identified between the psychological characteristic
depression and symptoms or physical function. Thus, while there was some support for the
influence of patient characteristics on symptoms and physical function in this study, the evidence
was not strong.

144

Based on the relationships between the variables in this study, it seems imperative that
nurses and health care providers include interventions targeted at the physiological
characteristics of obesity and smoking in order to reduce symptoms and improve physical
function. The standard care approaches that are focused on identifying the anatomic pain
generator will no longer be sufficient. Incorporating interventions aimed at reducing BMI and
smoking cessation with teaching regarding the effects of obesity and smoking on back pain and
physical function should be an integral part of spine care. Tailored approaches that incorporate
change theory and patient preferences can be developed to target the patient characteristics that
are significantly associated with worse physical function outcomes.
Based on the relationships identified between pain VAS and physical function, nurses
and health care professionals should focus on techniques to reduce pain. Finally, since insurance
type was found to have associations with the symptom of pain VAS and physical function,
healthcare professionals should advocate for consistency in coverage for all insurance plans for
evidence-based interventions such as physical therapy and injections, to reduce pain and improve
physical function.
Even though the propositions of the TOUS were only partially supported, organizing care
for individuals with lumbar degenerative conditions based on the theory may provide more
comprehensive care than the current standard care. Specifically, assessment that includes
physiological, situational and psychological factors could identify risk factors that if addressed
early, could reduce symptoms and maintain physical function. Current approaches that address
only the presumed anatomic pain generator and do not incorporate patient characteristics that
place patients at risk for worse physical function may not be sufficient to produce meaningful
improvements in physical function. An awareness of the associations between patient

145

characteristics and symptoms and their effects on physical function could allow nurses and
healthcare providers to intervene early to preserve physical function through tailored approaches
that incorporate the identified risks and the preferences for that individual. The multiplicity of
factors that are associated with symptoms and physical function in this population requires a
trans-disciplinary approach that incorporates nurses, physicians, pain care providers, behavioral
specialists and physical therapists.
Implications for Research
Because the use of the TOUS in this study was not sufficient to explain a significant
portion in the variability in symptoms and physical function, other models may need to be
considered for organizing the approach to study of the factors that influence the outcome of
physical function in this population. One such model is the Disablement Process, described as a
“socio-medical” model (Verbrugge & Jette, 1994). Disablement is conceptualized as a pathway
on a continuum, moving from pathology to impairments to functional limitations to disability.
This pathway is influenced by factors external to the individual, factors within the individual,
and other attributes considered to be risk factors that elevate the probability of disability. These
factors may speed or slow the disablement process. Acute and chronic conditions are included in
the model, and the authors discuss interventions aimed at slowing the disablement process. The
disablement process model may provide more salient variables and useful propositions for
organizing the approach to study in this population.
There are no studies evaluating the combined effects of numbness and weakness on
physical function in this population. Future studies should consider analyzing numbness and
weakness together, in order to determine whether these symptoms catalyze each other in their
effect on physical function. Numbness may affect physical function because of a loss of

146

sensation involving the foot, affecting proprioception. Weakness may affect physical function
through interference with normal gait mechanics and trips. Together, these symptoms may have
a greater influence on physical function than either symptom alone.
Further studies examining the frequency of back pain alone, back pain and leg pain
together and leg pain alone for their effects on physical function could help health professionals
identify and stratify those at risk for worse physical function. Though different anatomic pain
generators in the spine share similar pain patterns (Taylor, Coxon & Watson, 2013; Cohen &
Raja, 2007; van der Werff, Buijs & Groen, 2006), there is limited knowledge regarding the
effects of pain patterns on physical function.
Measuring the influence of patient characteristics and genotype on symptoms and
physical function over time would allow nurse scientists to determine temporal relationships
between these variables. A longitudinal design could aid in determining the effects of patient
characteristics on patient’s response to treatment for lumbar degenerative conditions, thereby
indentifying those at risk for poorer responses to treatment.
This study highlights the need for further research that includes other important patient
characteristics that may influence symptoms and physical function in persons with lumbar
degenerative conditions. There was unexplained variance accounting for the effects of patient
characteristics on symptoms and physical function, suggesting that other as yet unidentified
variables may play a role. Important variables for future study may include race, ethnicity and
socioeconomic status (SES) in order to discover other relevant factors to include in a risk
assessment for individuals with lumbar degenerative conditions. Associations between ethnicity
and representation of SNPs would provide further insight into how genotypes are represented in

147

different populations and their effects on symptoms and physical function specific to those
populations.
Since Aim 3 was exploratory, the genes selected for study were based on a review of
those most commonly studied in relationship to the structure of the intervertebral disc and those
related to the experience of pain. There are many other genes that have been studied relative to
disc structure and to compounds that have been shown to affect the rate and severity of disc
degeneration. None of the genes encoding for substances that affect the rate and severity of disc
degeneration were included in this study.
Larger sample sizes for genotyping could provide more evidence for the connections
between genotype and phenotype in this population, as well as more information regarding the
representation of genotype in different ethnic populations. Caution should be used, however,
when interpreting the results of these multiple comparisons because of the likelihood of finding
significant results that are the result of chance alone.
Since the psychological characteristic depression was not measured directly in this study,
future research should incorporate a method to measure this variable directly. This would
strengthen the validity of this variable and the conclusions made regarding the associations
between depression and symptoms and physical function in this population.
This study also highlights the issue of a lack of instruments to measure symptoms in
individuals with lumbar degenerative conditions. While there are existing instruments to
measure pain, the unique features of the symptoms that accompany lumbar degenerative
conditions (the nature, location, characteristics of back and limb pain, with numbness and
weakness) may require measurement techniques that are sensitive to these features. A spinal
stenosis symptom measure has been developed and tested, (Stucki et al., 1996). More work must

148

be done to develop and refine instruments for measuring symptoms and their impact in
individuals with lumbar degenerative conditions.
There were different results for the statistical tests involving physical function as
measured by the ODI and the SF-36 physical function subscale. This finding reflects that the
ODI is clearly a disease-specific instrument designed to measure physical function in the
population of individuals with lumbar degenerative conditions, and is superior to the physical
function subscale of the SF-36 for this purpose.
Finally, this study provides limited support for the use of the TOUS as an organizing
framework for future studies on the effects of patient characteristics in persons with lumbar
degenerative conditions. This study partially supports the notion that different categories of
patient characteristics have an influence on symptoms and physical function in this population.
The concept of how symptoms interact with patient characteristics and their combined effects on
physical function has not been sufficiently explored. This may represent an important
opportunity for nurses to add to the body of knowledge by incorporating the study of symptoms
with other patient characteristics for their effects on physical function in persons with lumbar
degenerative conditions. Revisions to the TOUS may improve its use in the future. Better
definitions of the specific patient physiological, situational and psychological characteristics may
improve the testability of the model and its use in clinical practice. As the science of genotype
and phenotype progresses, it would be helpful to define how this is incorporated into the
TOUS—does it fit as a physiological variable?
Implications for Policy
Obesity and smoking have been identified as an important health concerns in the U.S.
This study suggests that obesity and smoking are also specific concerns in persons with lumbar

149

degenerative conditions for their effects on symptoms and physical function. Spine care
programs should incorporate interventions designed to address all of the risk factors that affect
symptoms and physical function in this population and should include specific interventions that
target weight loss and smoking cessation.
This study also identified associations between insurance plans and pain and physical
function in this population. Specifically, not having insurance or having Medicaid insurance was
associated with higher pain scores and worse physical function. Conversely, having Medicare or
Commercial insurance was associated with better physical function scores on the ODI. This
difference may be due to better coverage with Medicare and Commercial insurance for evidencebased interventions such as spinal injections and physical therapy, designed to reduce pain and
improve physical function in this population. Steps should be taken to provide for consistency in
coverage for evidence-based interventions like spinal injections and physical therapy across
insurance plans. Consideration should be given for providing coverage for weight loss
treatments in this population. Because these data were collected before implementation of the
Affordable Care Act, the scope and effects of coverage provided under these insurance plans is
not yet known.
Last, the body of knowledge related to the use of genotyping for personalized medicine is
a growing field of study. Controversy exists surrounding the implications of the use of
genotyping and confidentiality issues. While the science of genotype and phenotype in
populations with spinal degeneration is still developing, this information could provide helpful
knowledge in the future to reduce pain and maintain physical function. This study raises
questions regarding which genes contribute the most to symptoms and physical function in this
population. Genes involved in the breakdown of the intervertebral disc were not included in this

150

study, and should be considered in combination with disc structural genes and genes associated
with the experience of pain, in order to further explore the relationship of genotype to symptoms
and physical function in individuals with lumbar degenerative conditions.
Conclusion/Summary
The primary purpose of this study was to examine the patient characteristics and
symptoms that contribute to the outcome of physical function in a population experiencing
lumbar degenerative conditions. Additionally, the novel physiological patient characteristic
genotype was explored for its association with symptoms and physical function.
The physiological characteristic (smoking) and the situational characteristics (Medicaid
insurance and no insurance) had significant negative influences on pain VAS. Higher BMI,
smoking, older age and Medicaid insurance were significantly associated with worse physical
function, while having Commercial insurance or Medicare were significantly associated with
better physical function.
The variables of patient physiological, situational and psychological characteristics and
symptoms were analyzed in order to develop a predictive model for the outcome physical
function. Higher BMI and higher pain VAS were significant predictors for worse physical
function for both the ODI and the SF-36 physical function subscale scores, while smoking and
the presence of numbness were significant predictors for worse ODI scores and older age was a
significant predictor of worse SF-36 physical function subscale scores. Both measures of
physical function were available in the database that was used for the variables tested in this
study, so both were used in the analysis and analyzed separately. The differences in the variables
that predicted physical function scores between the two instruments were likely related to the

151

ODI being a lumbar-specific instrument and the SF-36 being a multi-purpose measure of
functional health and well-being.
Last, a small sample (n = 28) of the study population provided saliva samples for
genotyping. Genotype data for 4 genes implicated in maintaining disc structure and 2 genes
implicated in the experience of pain were collected and analyzed for associations with symptoms
and physical function. There was limited diversity of SNPs for the COL9A3 and the OPRM1
genotypes. While there were no statistically significant associations between genotype and
symptoms, OPRM1 genotype was significantly associated with physical function scores.
The findings from this study are an important first step that connects patient
characteristics (including genotype) and symptoms to show their influence on physical function
in persons with lumbar degenerative conditions. This study also raises important questions
regarding which genes have the greatest impact with other patient characteristics on symptoms
and physical function for individuals with lumbar degenerative conditions. Genes known to
influence the breakdown of the intervertebral disc were not studied, and could be included in
future studies. Other candidate genes known to influence intervertebral disc structural integrity
should also be included in future studies.

152

APPENDICES

153

Appendix A: Figures

Figure 1

The Theory of Unpleasant Symptoms with Study Variables

154

Figure 2

Neurosurgery/Spine Health History

155

Figure 2 (cont’d)

156

Figure 2 (cont’d)

157

Figure 2 (cont’d)

158

Figure 3

SF-36

159

Figure 3 (cont’d)

160

Figure 3 (cont’d)

161

Figure 3 (cont’d)

162

Figure 3 (cont’d)

163

Figure 3 (cont’d)

164

Figure 4

IRB Approval

165

Figure 4 (cont’d)

166

Figure 5

Pain Diagram Overlay

Management of Non-radicular Low Back Pain: A Pilot Clinical Trial. Manual Therapy, 11, 279286.

167

Appendix B

Oswestry Disability Index

168

Appendix C

Data Collection Tool

INDIVIDUAL CHARACTERISTICS, SYMPTOMS AND PHYSICAL FUNCTION IN
LUMBAR DEGENERATION
DATA COLLECTION TOOL
ID Number_____
Physiological Factors
Height___________________________
Weight___________________________
BMI calculation___________________
BMI Category____________________
Sex Female_______Male___________
Age__________
Smoking Y_____N_____
Genotype
OPRM-1 SNP_____
COMT SNP_____
COL9A2 SNP_____
COL9A3 SNP_____
ACAN VNTR______

169

Appendix C (cont’d)
VDR SNP_____

Situational Factors
Employment Status
Currently working? Y_____N_____
Worker’s Compensation claim? Y_____N_____
Insurance Type:
Commercial_____
Medicare______
Champus_____
Medicaid_____
None_____
Psychological Factors
Depression Y_____N_____
Symptoms
Pain VAS score_____ (Measured in Cm)
Pain location
1_____
2_____
3_____
4_____

170

Appendix C (cont’d)
5_____
6_____
Limb Numbness Y_____N_____
Weakness Y____N____
Outcome Measures:
SF-36 Physical Function subscale score_____
ODI score_____
Medical Co-morbidities
HTN Y_____N_____
Diabetes Y_____N_____
CHD Y _____N_____
Fibromyalgia Y_____N_____
Other (list)______________________________________________________________
Total_____
Medications: (Record all oral medications the individual is currently taking, both scheduled and
prn, in the following categories: non-steroidal anti-inflammatories(NSAIDS), steroids,
analgesics and narcotics. Code 1 for NSAIDS, 2 for steroids, 3 for analgesics and 4 for
narcotics.)
Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

171

Appendix C (cont’d)
Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

Medication Name:__________________________

Code:___________

172

Appendix D

Permission to Use TOUS
WOLTERS KLUWER HEALTH LICENSE
TERMS AND CONDITIONS
Nov 18, 2013

This is a License Agreement between Teri L Holwerda ("You") and Wolters Kluwer Health
("Wolters Kluwer Health") provided by Copyright Clearance Center ("CCC"). The license
consists of your order details, the terms and conditions provided by Wolters Kluwer Health,
and the payment terms and conditions.
All payments must be made in full to CCC. For payment instructions, please see
information listed at the bottom of this form.
License Number

3272201448457

License date

Nov 18, 2013

Licensed content publisher

Wolters Kluwer Health

Licensed content publication Advances in Nursing Science
Licensed content title

The Middle-Range Theory of Unpleasant Symptoms: An Update

Licensed content author

Elizabeth R. Lenz, Linda C. Pugh, Renee A. Milligan, et al

Licensed content date

Jan 1, 1997

Volume Number

19

Issue Number

3

Type of Use

Dissertation/Thesis

Requestor type

Individual

Author of this Wolters
Kluwer article

No

Title of your thesis /
dissertation

The Effects of Individual Characteristics and Symptoms on Physical
Function in Persons with Lumbar Degenerative Conditions

Expected completion date

Feb 2014

Estimated size(pages)

130

Billing Type

Invoice

173

Appendix D (cont’d)

Billing address

650 Griswold St SE

GRAND RAPIDS, MI 49507

United States
Total

0.00 USD

Terms and Conditions

Terms and Conditions
1. A credit line will be prominently placed and include: for books - the author(s), title of
book, editor, copyright holder, year of publication; For journals - the author(s), title of
article, title of journal, volume number, issue number and inclusive pages.
2. The requestor warrants that the material shall not be used in any manner which may be
considered derogatory to the title, content, or authors of the material, or to Wolters
Kluwer.
3. Permission is granted for a one time use only within 12 months from the date of this
invoice. Rights herein do not apply to future reproductions, editions, revisions, or other
derivative works. Once the 12-month term has expired, permission to renew must be
submitted in writing.
4. Permission granted is non-exclusive, and is valid throughout the world in the English
language and the languages specified in your original request.
5. Wolters Kluwer cannot supply the requestor with the original artwork or a "clean copy."
6. The requestor agrees to secure written permission from the author (for book material
only).
7. Permission is valid if the borrowed material is original to a Wolters Kluwer imprint
(Lippincott-Raven Publishers, Williams & Wilkins, Lea & Febiger, Harwal, Igaku-Shoin,
Rapid Science, Little Brown & Company, Harper & Row Medical, American Journal of
Nursing Co, and Urban & Schwarzenberg - English Language).
8. If you opt not to use the material requested above, please notify Rightslink within 90
days of the original invoice date.
9. Please note that articles in the ahead-of-print stage of publication can be cited and the
content may be re-used by including the date of access and the unique DOI number. Any
final changes in manuscripts will be made at the time of print publication and will be
reflected in the final electronic version of the issue.?Disclaimer: Articles appearing in the
Published Ahead-of-Print section have been peer-reviewed and accepted for publication
in the relevant journal and posted online before print publication. Articles appearing as
publish ahead-of-print may contain statements, opinions, and information that have
errors in facts, figures, or interpretation. Accordingly, Lippincott Williams & Wilkins, the
editors and authors and their respective employees are not responsible or liable for the
use of any such inaccurate or misleading data, opinion or information contained in the
articles in this section.

174

Appendix D (cont’d)
10. 1This permission does not apply to images that are credited to publications other than
Wolters Kluwer journals. For images credited to non-Wolters Kluwer journal publications,
you will need to obtain permission from the journal referenced in the figure or table
legend or credit line before making any use of the image(s) or table(s).
11. In case of Disease Colon Rectum, Plastic Reconstructive Surgery, The Green
Journal, Critical Care Medicine, Pediatric Critical Care Medicine, the American
Heart Publications, the American Academy of Neurology the following guideline
applies: no drug brand/trade name or logo can be included in the same page as the
material re-used
12. When requesting a permission to translate a full text article, Wolters Kluwer/Lippincott
Williams & Wilkins requests to receive the pdf of the translated document
13. “Adaptations of single figures do not require Wolters Kluwer further approval if the
permission has been granted previously. However, the adaptation should be credited as
follows:?Adapted with permission from Lippincott Williams and Wilkins/Wolters Kluwer
Health: [JOURNAL NAME] (reference citation), copyright (year of publication)”
Please note that modification of text within figures or full-text articles is strictly
forbidden.
14. The following statement needs to be added when reprinting the material in Open Access
journals only: ‘promotional and commercial use of the material in print, digital or mobile
device format is prohibited without the permission from the publisher Lippincott Williams
& Wilkins. Please contact journalpermissions@lww.com for further information”.
15. Other Terms and Conditions:

v1.8
If you would like to pay for this license now, please remit this license along with your
payment made payable to “COPYRIGHT CLEARANCE CENTER” otherwise you will be
invoiced within 48 hours of the license date. Payment should be in the form of a check
or money order referencing your account number and this invoice number
RLNK501162191.
Once you receive your invoice for this order, you may pay your invoice by credit card.
Please follow instructions provided at that time.
Make Payment To:
Copyright Clearance Center
Dept 001
P.O. Box 843006
Boston, MA 02284-3006
For suggestions or comments regarding this order, contact RightsLink Customer
Support: customercare@copyright.com or +1-877-622-5543 (toll free in the US) or +1978-646-2777.
Gratis licenses (referencing $0 in the Total field) are free. Please retain this printable
license for your reference. No payment is required.

175

Appendix E

Informed Consent

176

Appendix E (cont’d)

177

Appendix E (cont’d)

178

Appendix E (cont’d)

179

Appendix E (cont’d)

180

Appendix E (cont’d)

181

Appendix F

Data Use Agreement

182

Appendix F (cont’d)

183

REFERENCES

184

REFERENCES

Abraham, J., Maranian, M., Spiteri, I., Russell, R., Ingle, S., Luccarini, C., Earl, H., Pharoah, P.,
Dunning, A. & Caldas, C. (2012). Saliva samples are a viable alternative to blood
samples as a source of DNA for high throughput genotyping. BMC Medical Genomics,
5, retrieved from http://www.biomedcentral.com/1755-8794/5/1/19
Ahles, T., Ruckdeschel, J. & Blanchard, E. (1984). Cancer-related pain—II. Assessment with
visual analogue scales. Journal of Psychosomatic Research, 28, 121-124.
Aladin, D., Cheung, K., Chan, D., Jim, J., Luk, K. & Lu, W. (2007). Expression of the Trp2
allele of COL9A2 is associated with alterations in the mechanical properties of human
intervertebral discs. Spine, 32, 2820-2826.
Altinkaya, N., Yildirim, T., Demir, S., Alkan, O. & Sarica, F. (2011). Factors associated with
thickening of the ligamentum flavum is thickening of the ligamentum flavum thickening
due to hypertrophy or buckling? Spine, 36, E1093-E1097.
Anderson, P., Schwaegler, P., Cizek, D. & Leverson, G. (2006). Work status as a predictor of
surgical outcome of discogenic low back pain. Spine, 31, 2510-2515.
Anderson, P., Subach, B. & Riew, D. (2009). Predictors of outcome after anterior cervical
discectomy and fusion a multivariate analysis. Spine, 34, 161-166.
Annunen, S., Paassilta, P., Lohiniva, J., Perala, M., Pihlajamaa, T., Karppinen., J., Tervonen, O.,
Kroger, H., Lahde, S., Vanharanta, H., Ryhanen, L., Goring, H., Ott, J., Prockop, D. &
Ala-Kokko, L. (1999). An allele of COL9A2 associated with intervertebral disc disease.
Science, 285, 409-412.
Argoff, C. (2010). Clinical implications for opioid pharmacogenetics. Clinical Journal of Pain,
26, S16-20.
Armstrong, T. (2003). Symptoms experience: A concept analysis. Oncology Nursing Forum,
30, 601-606.
Atlas, S., Chang, Y., Kammann, E., Keller, R., Deyo, R. & Singer, D. Long-term disability and
return to work among patients who have a herniated lumbar disc: The effect of disability
compensation. Journal of Bone & Joint Surgery, 82A, 4-15.
Atlas, S., Deyo, R., Patrick, D., Convery, K., Keller, R. & Singer, D. (1996). The Quebec Task
Force classification for spinal disorders and the severity, treatment, and outcomes of
sciatica and lumbar spinal stenosis. Spine, 21, 2885-2892.

185

Balmain, N., Hauchecorne, M., Pike, J., Cuisinier-Gleizes, P. & Mathieu, H. (1993).
Distribution and subcellular immunolocalization of 1,25-dihydroxyvitamin D3 receptors
in rat epiphyseal cartilage. Cellular and Molecular Biology (Noisy-le-grand). 39, 339-50.
Battie, M., Videman, T., Gill, K., Moneta, G., Nyman, R., Kaprio, J. & Koskenvuo, M. (1991).
1991 Volvo Award in clinical sciences. Smoking and lumbar intervertebral disc
degeneration: An MRI study of identical twins. Spine, 16, 1015-1021.
Battie, M., Videman, T., Levalahti, E., Gill, K. & Kaprio, J. (2007). Heritability of low back
pain and the role of disc degeneration. Pain, 131, 272-280.
Battie, M. Videman, T., Levalahti, E., Gill, K. & Kaprio, J. (2008). Genetic and environmental
effects on disc degeneration by phenotype and spinal level a multivariate twin study.
Spine, 33, 2801-2808.
Beaton, D. & Schemitsch, E. (2003). Measures of health-related quality of life and physical
function. Clinical Orthopaedics and Related Research, 413, 9-105.
Behrend, C., Prasarn, M., Coyne, E., Horodyski, M., Wright, J. & Rechtine, G. (2012). Smoking
cessation related to improved patient-reported pain scores following spinal care. Journal
of Bone & Joint Surgery, 94, 2161-2166.
Bener, A., Verjee, M., Dafeeah, E., Falah, O., Al-Juhaishi, T., Schlog, J., Sedeeq, A. & Khan, S.
(2013). Psychological factors: Anxiety, depression and somatization in low back pain
patients. Journal of Pain Research, 6, 95-101.
Benoist, M., Boulo, P. & Hayem, G. (2012). Epidural steroid injections in the management of
low-back pain with radiculopathy: An update on their efficacy and safety. European
Spine Journal, 21, 204-213.
Bogduk, N. (2012). Degenerative joint disease of the spine. Radiologic Clinics of North
America, 50, 613-628.
Brant, J., Beck, S. & Miaskowski, C. (2010). Building dynamic models and theories to advance
the science of symptom management research. Journal of Advanced Nursing, 66, 228240.
Burnham, R., Warren, S., Saboe, L., Davis, L., Russell, G. & Reid, D. (1996). Factors predicting
employment 1 year after traumatic spine fracture. Spine, 21, 1066-1071.
Busija, L., Osborne, R., Nilsdotter, A., Buchbinder, R. & Roos, (2008). Magnitude and
meaningfulness of change in SF-36 scores in four types of orthopedic surgery. Health
and Quality of Life Outcomes, 6, 1-12. doi: 10.1186/1477-7525-6-55

186

Cai, C., Pua, Y. & Lim, K. (2007). Correlates of self-reported disability in patients with low
back pain: The role of fear-avoidance beliefs. Annals of the Academy of Medicine,
Singapore, 36, 1013-1020.
Carragee, E., Alamin, T., Miller, J. & Carragee, J. (2005). Discographic, MRI and psychosocial
determinants of low back pain disability and remission: A prospective study in subjects
with benign persistent back pain. The Spine Journal, 5, 24-35.
Cawthon, P., Fox, K., Gandra, S., Delmonico, M., Chiun-Fang, C., Anthony, M., Caserotti, P.,
Kritschevsky, S., Newman, A., Goodpaster, B., Satterfield, S., Cummings, S. & Harris, T.
(2011). Clustering of strength, physical function, muscle and adiposity characteristics
and risk of disability in older adults. Journal of the American Geriatric Society, 59, 781787.
Centers for Disease Control and Prevention, 2014. Prevalence of Self-Reported Obesity Among
U.S. Adults. Retrieved from: http://www.cdc.gov/obesity/data/adult.html
Centers for Disease Control and Prevention, 2014. Adult Cigarette Smoking in the United States:
Current Estimates. Retrieved from:
http://www.cdc.gov/tobacco/data_statistics/fact_sheets/adult_data/cig_smoking/
Chapman, J., Norvell, D., Hermsmeyer, J., Bransford, R., De Vine, J., Mc Girt, M. & Lee, M.
(2011). Evaluating common outcomes for measuring treatment success for chronic low
back pain. Spine, 36, S54-S68.
Cheng, J., Lee, M., Massicotte, E., Ashman, B., Gruenberg, M., Pilcher, L. & Skelly, A. (2011).
Clinical guidelines and payer policies on fusion for the treatment of chronic low back
pain. Spine, 36, S144-S163.
Cheung, K., Karppinen, J., Chan, D., Song, Y., Sham, P., Cheah, K., Leong, J. & Luk, K. (2009).
The prevalence and pattern of lumbar magnetic resonance imaging changes in a
population study of one thousand forty-three individuals. Spine, 34, 934-940.
Cheung, K., Samartzis, D., Karppinen, J. & Luk, K. (2012). Are “patterns” of lumbar disc
degeneration associated with low back pain? New insights based on skipped level disc
pathology? Spine, 37, E430-E438.
Chokshi, F., Quencer, R. & Smoker, W. (2010). The “thickened” ligamentum flavum: Is it
buckling or enlargement? American Journal of Neuroradiology, 31, 1813-1816.
Choma, T., Schuster, J., Norvell, D., Dettori, J. & Chutkan, N. (2011). Fusion versus nonoperative treatment for chronic low back pain do comorbid diseased or general health
factors affect outcome? Spine, 36, S87-S95.
Chou, W., Yang, L., Lu, H., Ko, J., Wang, C., Lin, S., Lee, T., Concejero, A. & Hsu, C. (2006).
Association of µ-opioid receptor gene polymorphism (A118G) with variations in

187

morphine consumption for analgesia after total knee arthorplasty. Acta
Anaesthesiologica Scandinavica, 50, 787-792.
Chung-Wei, C., McAuley, J., Macedo, L., Barnett, D., Smeets, R., & Verbunt, J. (2011).
Relationship between physical activity and disability in low back pain: A systematic
review and meta-analysis. Pain, 152, 607-613.
Cleland, J., Childs, J., Palmer, J. & Eberhart, S. (2006). Slump stretching in the management of
non-radicular pain: A pilot clinical trial. Manual Therapy, 11, 279-286.
Cleland, J., Gillani, R., Bienen, J. & Sadosky, A. (2010). Assessing dimensionality and
responsiveness of outcomes measures for patients with low back pain. Pain Practice, 11,
57-69.
Cohen, S. & Raja, S. (2007). Pathogenesis, diagnosis and treatment of lumbar zygapophysial
(facet) joint pain. Anesthesiology, 106, 591-614.
Cork, R., Isaac, I., Elsharydah, A., Saleemi, S., Zavisca, F. & Alexander, L. (2004). A
comparison of the Verbal Rating Scale and the Visual Analog Scale for pain assessment.
Internet Journal of Anesthesiology, 8, 1-4.
Corwin, E., Brownstead, J., Barton, N., Heckard, S & Morin, K. (2005). The impact of fatigue
on the development of postpartum depression. Journal of Obstetric, Gynecologic and
Neonatal Nursing, 34, 577-586.
Corwin, E., Klein, L. & Rickelman, K. (2002). Predictors of fatigue in healthy young adults:
Moderating effects of cigarette smoking and cancer. Biological Research for Nursing, 2,
222-233.
Crook, J., Milner, R., Schultz, I. & Stringer, B. (2002). Determinants of an occupational
disability following a low back injury: A critical review of the literature. Journal of
Occupational Rehabilitation, 12, 277-295.
Dai, F., Belfer, I., Schwartz, C., Banco, R., Martha, J., Tigioughart, H., Tromanhouser, S., Jenis,
L. & Kim, D. (2010). Association of catechol-O-methyltransferase genetic variants with
outcome in patients undergoing surgical treatment for lumbar degenerative disc disease.
The Spine Journal, 10, 949-957.
Daffner, S., Hymanson, H., & Wang, J. (2010). Cost and use of conservative management of
lumbar disc herniation before surgical discectomy. Spine, 10, 463-468.
Dagenais, S., Caro, J. & Haldeman, S. (2008). A systematic review of low back pain cost of
illness studies in the United States and internationally. The Spine Journal, 8, 8-20.
Dasenbrock, H., Wolinsky, J., Sciubba, D., Witham, T., Gokaslan, Z. & Bydon, A. (2012). The
impact of insurance status on outcomes after surgery for spinal metastases. Cancer, 118,
4833-4841.
188

Davidson, M. & Keating, J. (2002). A comparison of five low back disability questionnaires:
Reliability and responsiveness. Physical Therapy, 82, 8-24.
Downie, W., Leatham, P., Rhind, V., Wright, V., Branco, J. & Anderson, J. (1978). Studies with
pain rating scales. Annals of the Rheumatic Diseases, 37, 378-381.
Deyo, R., Mirza, S., Martin, B., Kreuter, W., Goodman, D. & Jarvik, J. (2010). Trends, major
medical complications, and charges associated with surgery for lumbar spinal stenosis in
older adults. Journal of the American Medical Association, 303, 1259-1265.
Dodd, M., Janson, S., Facione, N., Faucett, J., Froelicher, E., Humphreys, J., Lee, K.,
Miaskowski, C., Puntillo, K., Rankin, S., & Taylor, D. (2001). Advancing the science of
symptom management. Journal of Advanced Nursing, 33, 668-676.
Edward, R., Smith, M., Klick, B., Magyar-Russell, G., Haythornthwaite, J., Holavanahalli, R.,
Patterson, D., Blakeney, P., Lezotte, D., McKibben, J., & Fauerbach, J. (2007).
Symptoms of depression and anxiety as unique predictors of pain-related outcomes
following burn injury. Annals of Behavioral Medicine, 34, 313-322.
Eggermont, L., Milberg, W., Lipsitz, L., Scherder, E. & Leveille, S. (2009). Physical activity
and executive function in aging: The MOBILIZE Boston study. Journal of the American
Geriatric Society, 57, 1750-1756.
El-sayed, A., Ziewacz, J., Davis, M., Lau, D., Siddiqi, H., Zamora-Berridi, G. & Sullivan, S.
(2012). Insurance status and inequalities in outcomes after neurosurgery. World
Neurosurgery, 76, 459-466.
Eser, B., Cora, T., Eser, O., Kalkan, E., Haktanir, A., Erdogan, M. & Solak, M. (2010).
Association of the polymorphisms of vitamin D receptor and aggrecan genes with
degenerative disc disease. Genetic Testing and Molecular Biomarkers, 14, 313-317.
Fairbank, J., Couper, J., Davies, J. & O’Brien, J. (1980). The Oswestry low back pain
questionnaire. Physiotherapy, 66, 271-273.
Fairbank, J., Gwilym, S., France, J., Daffner, S., Dettori, J., Hermsmeyer, J. & Andersson, G.,
(2011). The role of classification of chronic low back pain. Spine, 36, S19-S42.
Fairbank, J., & Pynsent, P. (2007). The Oswestry disability index. Spine, 25, 2940-2953.
Falco, F., Manchikanti, L., Datta, S., Sehgal, N., Geffert, S., Onyewu, O., Zhu, J., Coubarous, S.,
Hameed, M., Ward, S., Sharma, M., Hameed, H., Singh, V & Boswell, M. (2012). An
update of the effectiveness of therapeutic lumbar facet joint interventions. Pain
Physician, 15, E909-E953.
Farrar, J., Berlin, J., & Strom, B. (2003). Clinically important changes in acute pain outcome
measures: A validation study. Journal of Pain and Symptom Management, 25, 406-411.
189

Farrell, D. & Savage, E. (2010). Symptom burden in inflammatory bowel disease: Rethinking
conceptual and theoretical underpinnings. International Journal of Nursing Practice, 16,
437-442.
Fawcett, F. (2005). Framework for analysis and evaluation of nursing theories. In Fawcett, J.
(2005). Contemporary nursing knowledge analysis and evaluation of nursing models and
theories (2nd Ed.). (pp. 441-449). Philadelphia: F.A. Davis Company.
Ferrans, C., Johnson Zerwic, J., Wilbur, J., & Larson, J. (1999). Conceptual model of healthrelated quality of life. Journal of Nursing Scholarship, 37, 336-342.
Ferreira, M. & Pereira, M. (2013). The mediator role of psychological morbidity in patients with
chronic low back pain in differentiated treatments. Journal of Health Psychology, doi:
10.1177/1359105313488970
Fillingim, R., Kaplan, L., Staud, R., Ness, T., Glover, Tl, Campbell, C., Mogil, J. & Wallace, M.
(2005). The A118G single nucleotide polymorphism of the µ-opioid receptor gene
(OPRM1) is associated with pressure pain sensitivity in humans. The Journal of Pain, 6,
159-167.
Freemont, A. (2009). The cellular pathobiology of the degenerate intervertebral disc and
discogenic back pain. Rheumatology, 48, 5-10.
Fritz, J. & Irrgang, J. (2001). A comparison of a modified Oswestry Low Back Pain Disability
Questionnaire and the Quebec Back Pain Disability Scale. Physical Therapy, 81, 776788.
Fu, M., Lemone, P. & McDaniel, R. (2004). An integrated approach to an analysis of symptom
management in patients with cancer. Oncology Nursing Forum, 31, 65-70.
Fu, M., McDaniel, R., & Rhodes, V. (2007). Measuring symptom occurrence and symptom
distress: Development of the symptom experience index. Journal of Advanced Nursing,
59, 623-634.
Genevay, S. & Atlas, S. (2010). Lumbar spinal stenosis. Best Practice & Research Clinical
Rheumatology, 24, 253-265.
George, S., Coronado, R., Beneciuk, J., Valencia, C., Werneke, M. & Hart, D. (2011).
Depressive symptoms, anatomical region, and clinical outcomes for patients seeking
outpatient physical therapy for musculoskeletal pain. Physical Therapy, 91, 358-372.
Gift, A. (2009). Unpleasant Symptoms. In Peterson, S. & Bredow, T. Middle Range Theories
Application to Nursing Research (2nd ed.) (pp. 82-98). Philadelphia: Wolters-Kluwer
Health/Lippincott Williams & Wilkins.

190

Gift, A., Jablonski, A., Stommel, M. & Given, W. (2004). Symptom clusters in elderly patients
with lung cancer. Oncology Nursing Forum, 31, 203-210.
Gillum, R. & Obesisan, T. (2010). Physical activity, cognitive function, and mortality in a U.S.
cohort. Annals of Epidemiology, 20, 251-257.
Glassman, S., Copay, A., Berven, S., Polly, D., Subach, B. & Carreon, L. (2008). Defining
substantial clinical benefit following lumbar spine arthrodesis. The Journal of Bone &
Joint Surgery, 90, 1839-1847. doi: 10.2106/JBJS.G.01095
Golatowski, C., Salazar, M., Dhople, V., Hammer, E., Kocher, T., Jehmlich, N. & Volker. U.
(2013). Comparative evaluation of saliva collection methods for proteome analysis.
Clinica Chimica Acta, 419, 42-46.
Gong, X., Wang, J., Liu, F., Yuan, H., Zhang, W., Guo, Y. & Jiang, B. (2013). Gene
polymorphisms of OPRM1 A118G and ABCB1 C3435T may influence opioid
requirements in Chinese patients with cancer pain. Asian Pacific Journal of Cancer
Prevention, 14, 2937-2943.
Greenstein, A., Moskowitz, A., Gelijns, A. & Egorova, N. (2012). Payer status and treatment
paradigm for acute cholecystitis. Archives of Surgery, 147, 453-458.
Grotle, M., Brox, J. & Vollestad, N. (2004). Concurrent comparison of responsiveness in pain
and functional status measurements used for patients with low back pain. Spine, 29,
E492-E501.
Gun, R., Osti. O., O’Riordan, A., Mpelasoka, F., Eckerwall, C. & Smyth, J. (2005). Risk factors
for prolonged disability after whiplash injury: A prospective study. Spine, 30, 386-391.
Guyer, R., Siddiqui, S., Zigler, J., Ohnmeiss, D., Blumenthal, S., Sachs, B., Hochschuler, S. &
Rashbaum, R. (2008). Lumbar spinal arthroplasty analysis of one center’s twenty best
and twenty worst clinical outcomes. Spine, 33, 2566-2569.
Hadjipavlou, A., Tzermiadianos, M., Bogduk, N. & Zindrick, M. (2008). The pathophysiology
of disc degeneration. The Journal of Bone and Joint Surgery, 90-B, 1261-1270.
Hammer, J., Howell, S., Bytzer, P., Horowitz, M., & Talley, N. (2003). Symptom clustering in
subjects with and without diabetes mellitus: A population-based study of 15,000
Australian adults. The American Journal of Gastroenterology, 98, 391-398.
Hangai, M., Kaneoka, K., Kuno, S., Hinotsu, S., Sakane, M., Mamizuka, N., Sakai, S. & Ochiai
N. (2008). Factors associated with lumbar intervertebral disc degeneration in the elderly.
The Spine Journal, 8, 732-740.
Harris, I., Mulford, J., Solomon, M., van Gelder, J. & Young, J. (2005). Association between
compensation status and outcome after surgery: A meta-analysis. Journal of the
American Medical Association, 293, 1644-1652.
191

Hartvigsen, J., Nielsen, J., Ohm Kyvik, K., Fejer, R., Vach, W., Iachine, I. & Leboef-Yde, C.
(2009). Heritability of spinal pain and consequences of spinal pain: A comprehensive
genetic epidemiologic analysis using a population-based sample of 15,328 twins ages 2071 years. Arthritis & Rheumatism, 61, 1343-1351.
Hastie, B., Riley, J., Kaplan, L., Herrera, D., Campbell, C., Virtusio, K., Mogil, J., Wallace, M.
& Fillingim, R. (2012). Ethnicity interacts with the OPRM1 gene in experimental pain
sensitivity. Pain, 153, 1610-1619.
Hasvik, E., Schistad, E., Grovle, L., Haugen, A., Roe, C. & Gjerstad, J. (2014). Subjective
health complaints in patients with lumbar radicular pain and disc hernation are associated
with a sex-OPRM1 A11G polymorphism interaction: A prospective 1-year observational
study. BMC Musculoskeletal Disorders, 15, http://www.biomedcentral.com/14712474/15/161
Healthy People 2020 (2012). Arthritis, Osteoporosis, and Chronic Back Conditons. Retrieved
from: http://healthypeople.gov/2020/topicsobjectives2020/overview.aspx?topicid=3
Heidrich, S., Egan, J., Hengudomsub, P., & Randolph, S. (2006). Symptoms, symptom beliefs,
and quality of life of older breast cancer survivors: A comparative study. Oncology
Nursing Forum, 33, 315-322.
Heuch, I., Hagen, K., Heuch, I., Nygaard, O. & Zwart, J. (2010). The impact of body mass index
on the prevalence of low back pain: The HUNT study. Spine, 35, 764-768.
Heuch, I. Heuch, I., Hagen, K. & Zwart, J. (2013). Body mass index as a risk factor for
developing chronic low back pain a follow up in the Nord-Trondelag Health Study.
Spine, 38, 133-139.
Higashino, K., Matsui, Y., Yagi, S., Takata, Y., Goto, T., Sakai, T., Katoh, S. & Yasui, N.
(2007). The alpha2 type IX collagen tryptophan polymorphism is associated with the
severity of disc degeneration in younger patients with herniated nucleus pulposus of the
lumbar spine. International Orthopaedics (SICOT), 31, 107-111.
Hirano, K., Imagama, S., Hasegawa, Y., Ito, Z., Muramotot, A. & Ishiguro, N. (2014). Impact of
low back pain, knee pain, and timed up-and-go test on quality of life in community-living
people. Journal of Orthopaedic Science, 19, 164-171.
Hobart, J., Williams, L., Moran, K. & Thompson, A. (2002). Quality of life measurement after
stroke: Uses and abuses of the SF-36. Stroke Journal of the American Heart
Association, 33, 1348-1356. doi: 10.1161/01.STR.0000015030.59594.B3
Hoffman, A.J., von Eye, A., Gift, A.G., Given, B.A., Given, C.W., & Rothert, M. (2009).
Testing a theoretical model of perceived self-efficacy for cancer-related fatigue selfmanagement and optimal physical functional status. Nursing Research, 58(1), 32-41.
192

Holliday, K. & McBeth, J. (2011). Recent advances in the understanding of genetic
susceptibility to chronic pain and somatic symptoms. Current Rheumatology Reports,
13, 521-527. doi: 10.1007/s11926-0110208-4
Huang, C., Liu, H., Su, N., Hsu, Y., Yang, C., Chen, C & Tsai, P. (2008). Association between
human opioid receptor genes polymorphisms and pressure pain sensitivity in females.
Journal of the Association of Anaesthetists of Great Britain and Ireland, 63, 1288-1295.
Hutchinson, S. & Wilson, H. (1998). The Theory of Unpleasant Symptoms and Alzheimer’s
Disease. Scholarly Inquiry for Nursing Practice, 12, 143-158.
Jarvik, J., Hollingworth, W., Heagerty, P., Haynor, D., Boyko, D., & Deyo, R. (2005). Threeyear incidence of low back pain in an initially asymptomatic cohort clinical and imaging
risk factors. Spine, 30, 1541-1548.
Jensen, M., Karoly, P. & Braver, S. (1986). The measurement of clinical pain intensity: A
comparison of six methods. Pain, 27, 117-126.
Jensen, M., Smith, D., Ehde, D. & Robinsin, L. (2001). Pain site and the effects of amputation
pain: Further clarification of the meaning of mild, moderate, and severe pain. Pain, 91,
317-322.
Jhawar, B., Fuchs, C., Colditz, G. & Stampfer, M. (2006). Cardiovascular risk factors for
physician-diagnosed lumbar disc herniation. The Spine Journal, 6, 684-691.
Jim, J., Noponen-Hietala, N., Cheung, K., Ott, J., Karppinen, J., Sahraravand, A., Luk, K., Yip,
S., Sham, P., Song, Y., Leong, J., Cheah, K., Ala-Kokko, L. & Chan, D. (2005). The
TRP2 allele of COL9A2 in an age-dependent risk factor for the development and severity
of intervertebral disc degeneration. Spine, 30, 2735-2742.
Jurgens, C., Moser, D., Armola, R., Carlson, B., Sethares, K. & Riegel, B. (2009). Symptom
clusters of heart failure. Research in Nursing & Health, 32, 551-560.
Kales, S., Linos, A., Chatzis, C., Sai, Y., Halla, M., Nasioulas, G. & Christiani, D. (2004). The
role of collagen IX tryptophan polymorphisms in symptomatic intervertebral disc disease
in southern European patients. Spine, 29, 1266-1270.
Kalichman, L. & Hunter, D. (2007). Lumbar facet joint osteoarthritis: A review. Seminars in
Arthritis and Rheumatism, 37, 69-80.
Kalichman, L. & Hunter, D. (2008). The genetics of intervertebral disc degeneration.
Associated genes. Joint Bone Spine 75, 388-396.

193

Kalichman, L., Kim, D., Li, L., Guermazi, A. & Hunter, D. (2010). Computed tomographyevaluated features of spinal degeneration: Prevalence, intercorrelation, and association
with self-reported low back pain. The Spine Journal, 10, 200-208.
Kaptan, H., Yelcin, H. & Kasimcan, O. (2012). Correlation of low back pain caused by lumbar
spinal stenosis and depression in women: A clinical study. Archives of Orthopaedic and
Trauma Surgery, 132, 963-967.
Karahan, A, Kav, S. Abbasoglu, A. & Dogan, N. (2009). Low back pain: Prevalence and risk
factors among hospital staff. Journal of Advanced Nursing, 65, 516-524.
Kathy, K., Harris, K., Hadi, S. & Chow, E. (2007). What should be the optimal cut points for
mild, moderate, and severe pain? Journal of Palliative Medicine, 10, 1338-1346.
Kauppila, L. (2009). Atherosclerosis and disc degeneration/Low back pain: A systematic
review. European Journal of Vascular and Endovascular Surgery, 37, 661-670.
Kawaguchi, Y., Kanamori, M., Ishihara, H., Ohmori, K., Matsui, H. & Kimura, T. (2002). The
association of lumbar disc disease with vitamin-D receptor gene polymorphism. The
Journal of Bone and Joint Surgery, 84-A, 2022-2028.
Kim, H. & Schwartz, C. (2010). The genetics of pain: Implications for evaluation and treatment
of spinal disease. The Spine Journal, 10, 827-840.
Kim, H., Suh, B., Lee, D., Park, J., Kang, K., Chang, B., Lee, C. & Yeom, J. (2013). Gender
difference of symptom severity in lumbar spinal stenosis: Role of pain sensitivity. Pain
Physician, 16, E715-E723.
Kleiber, C., Schutte, C., McCarthy, A., Floria-Santos, M., Mirray, J. & Hanrahan, K. (2007).
Predictors of topical analgesic effectiveness in children. Journal of Pain, 8, 168-174.
Kohlboeck, G., Greimel, K., Piotrowski, P., Leibetseder, M., Krombholz-Reindl, M., Neuhofer,
R., Schmid, A. & Klinger, R. (2004). Prognosis of multifactorial outcome in lumbar
discectomy: A prospective longitudinal study investigating patients with disc prolapse.
Clinical Journal of Pain, 20, 455-461.
Kongsted, A., Kent, P., Albert, H., Jensen, T. & Manniche, C. (2012). Patients with low back
pain differ from those who also have leg pain or signs of nerve root involvement—a
cross-sectional study. BMC Musculoskeletal Disorders, 13. Retrieved from
http://www.biomedcentral.com/1471-2474/13/236
Konstantinou, K., Hider, S., Jordan, J., Lewis, M., Dunn, K. & Hay, E. (2013). The impact of
low back-related leg pain on outcomes as compared with low back pain alone: A
systematic review of the literature. Clinical Journal of Pain, doi:
10.1097/AJP.obo13e31826f9a52

194

Kosinski, M., Keller, S., Hatoum, H., Kong, S., & Ware, J. (1999). The SF-36 health survey as a
generic outcome measure in clinical trials of patients with osteoarthritis and rheumatoid
arthritis. Medical Care, 37, MS10-MS22.
Kruper, L., Holt, A., Xu, S., Duan, L., Henderson, K., Bernstein, L. & Ellenhorn, J. (2011).
Disparities in reconstruction rates after mastectomy: Patterns of care and factors
associated with the use of breast reconstruction in southern California. Annals of
Surgical Oncology, 18, 2158-2165.
LaCaille, R., DeBerard, S., LaCaille, L., Masters, K. & Colledge, A. (2007). Obesity and
litigation predict workers’ compensation costs associated with interbody cage lumbar
fusion. The Spine Journal, 7 266-272.
Lachman, H., Morrow, B., Shprintzen, R., Veit, S., Parsia, S., Faedda, G., Goldberg, R.,
Kucherlapati, R. & Papolos, D. (1996). Association of codon 108/158 catechol-Omethyltransferase gene polymorphism with the psychiatric manifestations of velo-cardiofacial syndrome. American Journal of Medical Genetics, 67, 468-472.
Lee, E. (2005). Relationships of mood disturbance, symptom experience and attentional function
in women with breast cancer based upon the Theory of Unpleasant Symptoms. Journal of
Korean Academy of Nursing, 35, 728-736.
Lee, M., Dettori, J., Standaert, C., Brodt, E. & Chapman, J. (2012). The natural history of
degeneration of the lumbar and cervical spines. Spine, 37, S18-S30.
Leidy, N. (1994). Functional status and the forward progress of merry-go-rounds: Toward a
coherent analytical framework. Nursing Research, 43, 196-202
Lenz, E.R., Pugh, L.C., Milligan, R.A., Gift, A., & Suppe, F. (1997). The middle-range theory
of unpleasant symptoms: An update. Advances in Nursing Science, 19(3), 14-27.
Lenz, E.R., Suppe, F., Gift, A.G., Pugh, L.C., & Milligan, R.A., (1995). Collaborative
development of middle-range nursing theories: Toward a theory of unpleasant
symptoms. Advances in Nursing Science, 17(3), 1-13.
Lin, M., Hwang, H., Chen, C. & Chiu, W. (2007). Comparisons of the Brief Form of the World
Health Organization Quality of Life and Short Form-36 for persons with spinal cord
injuries. American Journal of Physical Medicine and Rehabilitation, 86, 104-113.
Doi10.1097/01.phm.0000247780.64373.Oe
Lin, S., Lin, R., & Huang, L. (2006). Disability in patients with degenerative lumbar spinal
stenosis. Archives of Physical Medicine and Rehabilitation, 87, 1250-1256.
Liu, H. (2006). Fatigue and associated factors in hemodialysis patients in Taiwan. Research in
Nursing & Health, 29, 40-50.

195

Liuke, M., Solovieva, S., Lamminen, A., Luoma, K., Leino-Arjas, P., Luukkonen, R. &
Riihimaki, H. (2005). Disc degeneration of the lumbar spine in relation to overweight.
International Journal of Obesity, 29, 903-908.
Livshits, G., Popham, M., Malkin, I., Sambrook, P., MacGregor, A., Spector, T. & Williams, F.
(2011). Lumbar disc degeneration and genetic factors are the main risk factors for low
back pain in women: The UK Twin Spine Study. Annals of Rheumatic Diseases, 70,
1740-1745.
Lotta, T., Vidgren, J., Tilgmann, C., Ulmanen, I., Melen, K., Julkunen, I. & Taskinen, J. (1995).
Kinetics of human soluble and membrane-bound catechol-O-methyltransferase: A
revised mechanism and description of thermolabile variant of the enzyme. Biochemistry,
34, 4202-4210.
MacGregor, A., Andrew, T., Sambrook, P. & Spector, T. (2004). Structural, psychological, and
genetic influences on low back and neck pain: A study of adult female twins. Arthritis &
Rheumatism, 51, 160-167.
Maughan, E. & Lewis, J. (2010). Outcome measures in chronic low back pain. European Spine
Journal, 19, 1484-1494.
McCann, K. & Boore, J. (2000). Fatigue in persons with renal failure who require maintenance
haemodialysis. Journal of Advanced Nursing, 32, 1132-1142.
McClelland, S., Guo, H. & Okuyemi, K. (2011). Population-based analysis of morbidity and
mortality following surgery for intractable temporal lobe epilepsy in the United States.
Archives of Neurology, 68, 725-729.
McGuire, D. (1997). Measuring Pain. In M. Frank-Stromborg & S Olsen, (Eds.), Instruments
for clinical health-care research (2nd ed.) (pp. 528-554). Sudbury, Massachusetts: Jones
and Bartlett Publishers.
McHorney, C. (1996). Measuring and monitoring general health status in elderly persons:
Practical and methodological issues in using the SF-36 health survey. The Gerontologist,
36, 571-583.
Menon, S., Lea, R., Roy, B., Hanna, M., Wee, S., Haupt, M. & Griffiths, L. (2012). The human
µ-opioid receptor gene polymorphism (A118G) is associated with head pain severity in a
clinical cohort of female migraine with aura patients. The Journal of Headache and
Pain, 13, 513-519.
Miaskowski, C. (2009). Understanding the genetic determinants of pain and pain management.
Seminars in Oncology Nursing, 25, S1-S7.
Mohamed, R., Campbell, J., Cooper-White, J., Dimeski, G. & Punyadeera, C. (2012). Clinical
and Translational Medicine, 1, retrieved from
http://www.clintransmed.com/conent/1/1/19
196

Mok, L. & Lee, I (2008). Anxiety, depression and pain intensity in patients with low back pain
who are admitted to acute admitted to acute care hospitals. Journal of Clinical Nursing,
17, 1471-1480.
Monticone, M., Baiardi, P., Ferrari, S., Foti, C., Mugnai, R., Pillastrini, P., Vanti, C. & Zanoli, G.
(2009). Development of the Italian version of the Oswestry Disabiltiy Index (ODI) A
cross-cultural adaptation, reliability, and validity study. Spine, 34, 2090-2095.
Moon, H., Kim, J., Lee, H., Chotai, S., Kang, J., Suh, J. & Park, Y. (2012). Annulus fibrosus
cells interact with neuron-like cells to modulate production of growth factors and
cytokines in symptomatic disc degeneration. Spine, 37, 2-9.
Motl, R. & McAuley, E. (2009). Symptom cluster as a predictor of physical activity in Multiple
Sclerosis: Preliminary evidence. Journal of Pain and Symptom Management, 38, 270279.
Mura, E., Govoni, S., Racchi, M., Carossa, V., Ranzani, G., Allegri, M. & van Schaik, R. (2013).
Consequences of the 118A>G polymorphism in the OPRM1 gene: Translation from
bench to bedside? Journal of Pain Research, 6, 331-353.
Myers, J. (2009). Comparison of the Theory of Unpleasant Symptoms and the Conceptual
Model of Chemotherapy-related Changes in Cognitive Function. Oncology Nursing
Forum, 36, E1-E10.
Nakki, A., Videman, T., Kujala, U., Suhonen, M., Mannikko, M., Peltonen, L., Battie, M.,
Kaprio, J. & Saarela, J. (2011). Candidate gene association study of magnetic resonance
imaging-based hip osteoarthritis (OA): Evidence for COL9A2 gene as a common
predisposing factor for hip OA and lumbar disc degeneration. Journal of Rheumatology,
38, 747-752.
National Center for Health Statistics, (2009). Health, United States, 2009. Retrieved from:
http://www.cdc.gov/nchs/data/hus/hus09.pdf#053
National Institutes for Health (2014). Health risks of obesity. Retrieved from:
http://www.nlm.nih.gov/medlineplus/ency/patientinstructions/000348.htm
National Institute for Occupational Safety and Health, (2004). Worker Health Chartbook.
Retrieved from: http://www.cdc.gov/niosh/docs/2004146/detail/imagedetail.asp@imgid38.htm
Nguyen, T., Randolph, D., Talmage, J., Succop, P. & Travis, R. (2011). Long-term outcomes of
lumbar fusion among Workers’ Compensation subjects. Spine, 36, 320-331.
Ohnhaus, E. & Adler, R. (1975). Methodological problems in the measurement of pain: A
comparison between the verbal rating scale and the visual analogue scale. Pain, 1, 379384.
197

Ostelo, R., Deyo, R., Stratford, P., Waddell, G., Croft, P., Von Korff, M., Bouter, L. & de Vet,
H. (2008). Interpreting change scores for pain and functional status in low back pain.
Toward international consensus regarding minimal important change. Spine, 33, 90-94.
Parks, P., Lenz, E., Milligan, R. & Han, H. (1999). What happens when fatigue lingers for 18
months after delivery? Journal of Obstetric, Gynecologic, and Neonatal Nursing, 28, 8793.
Patrick, D., Deyo, R., Atlas, S., Singer, D., Chapin, A. & Keller, R. (1995). Assisting healthrelated quality of life in patients with sciatica. Spine, 20, 1899-1908.
Pereira, L., Obara, K., Dias, J., Menacho, M., Guariglia, D., Schiavoni, D., Pereira, H. &
Cardoso, J. (2012). Comparing the Pilates method with no exercise or lumbar
stabilization for pain and functionality in patients with chronic low back pain: A
systematic review and analysis. Clinical Rehabilitation, 26, 10-20.
Peterson, S. & Bredow, T. (2009). Introduction to the nature of nursing knowledge. In Peterson,
S. & Bredow, T. Middle range theories application to nursing research (2nd Ed.). (p. 31).
Philadelphia: Wolters Kluwer Lippincott Williams & Wilkins.
Pincus, T., Burton, A., Vogel, S., Field, A. (2002). A systematic review of psychological factors
as predictors of chronicity/disability in prospective cohorts of low back pain. Spine, 27,
E109-E20.
Prasarn, M., Horodyski, M., Behrend, C., Wright, J. & Rechtine, G. (2012). Negative effects of
smoking, worker’s compensation and litigation on pain/disability scores for spine
patients. Surgical Neurology International, 3, S388-S369. doi: 10.4103/21527806.103870
Prins, M., van der Wurff, M. & Groen, G. (2013). Chronic low back pain patients with
accompanying leg pain: The relationship between pain extent and pain intensity,
disability and health status. Journal of Back and Musculoskeletal Rehabilitation, 26, 5561.
Pugh, L., Milligan, R. & Lenz, E. (2000). Response to “Insomnia, fatigue, anxiety, depression
and quality of life of cancer patients undergoing chemotherapy”. Scholarly Inquiry for
Nursing Practice, 14, 291-294.
Rajaee, S., Bae, H., Kanim, L. & Delamarter, R. (2012). Spinal fusion in the United States:
Analysis of trends from 1998-2008. Spine, 37, 67-76.
Rathod, T., Chandanwale, A., Gujrathi, S., Patil, V., Chavan, S. & Shah, M. (2012). Association
between single nucleotide polymorphism in collagen IX and intervertebral disc disease in
the Indian population. Indian Journal of Orthopaedics, 46, 420-426.

198

Redeker, N., Lev, E. & Ruggiero, J. (2000). Insomnia, fatigue, anxiety, depression and quality
of life of cancer patients undergoing chemotherapy. Scholarly Inquiry for Nursing
Practice, 14, 275-290.
Reishtein, J. (2005). Relationship between symptoms and functional performance in COPD.
Research in Nursing & Health, 28, 39-47.
Rejeski, W.J. & Mihalko, S.L. (2001). Physical activity and quality of life in older adults.
Journals of Gerontology 56A, 23-35.
Rihn, J., Kurd, M., Hilibrand, A., Lurie, J., Zhao, W., Albert, T. & Weinstein, J. (2013). The
influence of obesity on the outcome of treatment of lumbar disc herniation. Analysis of
the Spine Patient Outcomes Research Trial (SPORT). The Journal of Bone and Joint
Surgery, 95, 1-8.
Rohan, M., Ohnmeiss, D., Guyer, R., Zigler, J., Blumenthal, S., Hochschuler, S., Sach, B.,
Rashbaum, R. (2009). Relationship between length of time off work preoperatively and
clinical outcome at 24-month follow-up in patients undergoing total disc replacement or
fusion. The Spine Journal, 9, 360-365.
Roland, M. & Fairbank, J. (2000). The Roland-Morris Disability Questionnaire and the
Oswestry Disability Questionnaire. Spine, 25, 3115-3124.
Rychnovsky, J. (2007). Postpartum fatigue in the active duty military woman. Journal of
Obstetric, Gynecologic and Neonatal Nursing, 36, 38-46.
Samartzis, D., Karppinen, J., Chan, D., Luk, K. & Cheung, K. (2012). The association of lumbar
intervertebral disc degeneration on Magnetic Resonance Imaging with body mass index
in overweight and obese adults. Arthritis & Rheumatism, 64, 1488-1496.
Samartzis, D., Karppinen, J., Mok, R., Fong, D., Luk, K. & Cheung, K. (2011). A populationbased study of juvenile disc degeneration and its association with overweight and obesity,
low back pain, and diminished functional status. The Journal of Bone and Joint Surgery,
93, 662-670.
Scott, J. & Huskisson, E. (1974). Graphic representation of pain. Pain, 2, 175-184.
Seki, S., Kawaguchi, Y., Mori, M., Mio, F., Chiba, K., Mikami, Y., Tsunoda, T., Kubo, T.,
Toyama, Y., Kimura, T. & Ikegawa, S. (2006). Association study of COL9A2 with
lumbar disc disease in the Japanese population. Journal of Human Genetics, 51, 10631067.
Serlin, R., Mendoza, T., Nakamura, Y., Edwards, K. & Cleeland, C. (1995). When is cancer
pain mild, moderate or severe? Grading pain severity by its interference with function.
Pain, 61, 277-284.

199

Shiri, R., Karppinen, J., Leino-Arjas, P., Solovieva, S., & Viikari-Juntura, E. (2010). The
association between obesity and low back pain: A meta-analysis. American Journal of
Epidemiology, 171, 135-154.
Shiri, R., Karppinen, J., Leino-Arjas, P., Solovieva, S., & Viikari-Juntura, E. (2010). The
association between smoking and low back pain: A meta-analysis. The American
Journal of Medicine, 123, 87 .e7-87 .35.
Shiri, R., Solovieva, S., Husgafvel-Pursiainen, K., Taimela, S., Saarikoski, L., Huupponen, R.,
Viikari, J., Raitakari, L. & Viikari-Juntura, E. (2008). The association between obesity
and the prevalence of low back pain in young adults: The Cardiovascular Risk in Young
Finns study. American Journal of Epidemiology, 167, 1110-1119.
Sia, A., Lim, Y., Lim, E., Ocampo, C., Lim, W., Cheong, P. & Tan, E. (in press). Influence of
mu-opioid receptor variant on morphine use and self-rated pain following abdominal
hysterectomy. The Journal of Pain. Retrieved from
http://www.sciencedirect.com.proxy1.cl.msu.edu/science/article/pii/S1526590013009395
#
Siemionow, K., An, H., Masuda, K., Andersson, G. & Cs-Szabo, G. (2011). The effects of age,
sex, ethnicity and spinal level on the rate of intervertebral disc degeneration a review of
1712 intervertebral discs. Spine, 36, 1333-1339.
Sigmundsson, F., Jonsson, B. & Stromqvist, B. (2013). Impact of pain on function and health
related quality of life in lumbar spinal stenosis: A register study of 14.821 patients.
Spine, 38, E937-E946.
Silverplats, K., Lind, B., Zoega, B., Halldin, K., Gellerstedt, M., Brisby, H. & Rutberg, L.
(2010). Clinical factors of importance for outcome after lumbar disc herniation surgery:
Long-term follow-up. European Spine Journal, 19, 1459-1467.
Slover, J., Abdu, W., Hanscom, B., Lurie, J. & Weinstein, J. (2006). Can condition-specific
health surveys be specific to spine disease? An analysis of the effect of comorbidities on
baseline condition-specific and general health survey scores. Spine, 31, 1265-1271.
Slover, J., Abdu, W., Hanscom, B. & Weinstein, J. (2006). The impact of comorbidities on the
change in Short-Form-36 and Oswestry scores following lumbar spine surgery. Spine,
31, 1974-1980.
Smith, M., & Liehr, P. (2008). Understanding middle range theory by moving up and down the
ladder of abstraction. In Smith, M. & Liehr, P. (Eds.), Middle range theory for nursing.
(2nd Ed.). (pp. 13-31.). New York: Springer Publishing Company.
So, W., Leung, D., Ho, S., Lai, E., Sit, J. & Chan, C. (in press). Associations between social
support, prevalent symptoms and health-related quality of life in Chinese women
undergoing treatment for breast cancer: A cross-sectional study using structural equation
200

modeling. European Journal of Oncology Nursing. Retrieved from
http//:dx.doi.org/10.1016/j.ejon.2012.11.001
Solovieva, S., Lohinvia, J., Leino-Arjas, P., Raininko, R., Luoma, K., Ala-Kokko, L. &
Riihimaki, H. (2006). Intervertebral disc degeneration in relation to the COL9A3 and IL1β gene polymorphisms. European Spine Journal, 15, 613-619.
Solovieva, S., Noponen, N., Mannikko, M., Lein-Arjas, P., Luoma, K., Raininko, R., Ala-Kokko,
L. & Riihimaki, H. (2007). Association between the aggrecan gene variable number of
tandem repeats polymorphism and intervertebral disc degeneration. Spine, 32, 17001705.
Soysal, M., Kara, B. & Arda, N. (2012). Assessment of physical activity in patients with chronic
low back or neck pain. Turkish Neurosurgery, 23, 75-80.
Tabachnick, B. & Fidell, L. (2007). Using multivariate statistics (5th Ed.). Boston: Pearson
Education, Inc.
Takatalo, J., Karppinen, J., Taimela, S., Niinimaki, J., Laitinen, J., Sequeiros, B., Smartizis, D.,
Korpelainen, R., Nayha, S., Remes, J. & Tervonen, O. (2013). Association of abdominal
obesity with lumbar disc degeneration—a magnetic resonance imaging study. PLoS
ONE, 8, doi: 10.1371/journal.pone.0056244
Tarlov, A., Ware, J., Greenfield, S., Nelson, E., Perrin, E & Zubkoff, M. (1989). The Medical
Outcomes Study: An application of methods for monitoring the results of medical care.
Journal of the American Medical Association, 262, 925-930.
Taylor, C., Coxon, A., Watson, P. & Greenough, C. (in press). Do L5 and S1 nerve root
compression produce radicular pain in a dermatomal pattern? Retrieved from
http://ovidsp.tx.ovid.com.proxy2.cl.msu.edu/sp3.12.0b/ovidweb.cgi?&S=BIMJFPNCDLDDNEAKNCMKFEDCJMIPAA00&Link+Set
=jb.search.49%7c1%7csl_10
Taylor, S., Taylor, A., Foy, M. & Fogg, A. (1999). Responsiveness of common outcome
measures for patients with low back pain. Spine, 24, 1805-1812.
Tetsunaga, T., Misawa, H., Tanaka, M., Sugimoto, Y., Tetsunaga, T., Takigawa, T. & Ozaki, T.
(2013). The clinical manifestations of lumbar disease are correlated with self-rating
depression scale scores. Journal of Orthopaedic Science, 18, 374-379.
Thomas, E., Silman, A., Croft, P., Papageorgiou, A., Jayson, M. & Macfarlane, G. (1999).
Predicting who develops chronic low back pain in primary care: A prospective study.
BMJ, 318, 1662-1667.
Trief, P., Grant, W. & Fredrickson, B. (2000). A prospective study of psychological predictors
of lumbar surgery outcome. Spine, 25, 2616-2621.
201

Tsao, D., Shabalina, S., Gauthier, J,. Dokholyan, N. & Diatchenko, L. (2011). Disruptive mRNA
folding increases translational efficiency catechol-O-methyltransferase variant. Nucleic
Acids Research, 39, 6201-6212.
Tyler, R. & Pugh, L. (2009). Application of the Theory of Unpleasant Symptoms in bariatric
surgery. Bariatric Nursing and Surgical Patient Care, 4, 271-276
U.S. Census Bureau, Statistical Abstract of the United States: 2012. (2012). Resident Population
Projections by Race, Hispanic-Origin Status, and Age: 2010 and 2015. Retrieved from:
http://www.census.gov/compendia/statab/2012/tables/12s0012.pdf
van Hooff, M., Spruit, M., O’Dowd, J., van Lankveld, W., Fairbank, J. & van Limbeek, J.
(2013). Predictive factors for successful clinical outcome 1 year after an intensive
combined physical and psychological programme for chronic low back pain. European
Spine Journal, doi: 10.1007/s00586-013-2844-z
van Kleef, M., Vanelderen, P., Cohen, S., Lataster, A., Zundert, J. & Mekhail, N. (2010). Pain
originating from the lumbar facet joints. Pain Practice, 10, 459-469.
Van Zundert, J., Vanelderen, P., Kessels, A. & van Kleef, M. (2012). Radiofrequency treatment
of facet-related pain: Evidence and controversies. Current Pain and Headache Reports,
16, 19-25.
Varlotta, G., Lefkowitz, T., Schweitzer, M., Errico, T., Spivak, J., Bendo, J. & Rybak, L. (2011).
The lumbar facet joint: A review of current knowledge: Part 1: Anatomy,
biomechanics, and grading. Skeletal Radiology, 40, 13-23.
Verbrugge, L. & Jette, A. (1994). The disablement process. Social Science & Medicine, 38, 114.
Videman, T., Leppavuori, J., Kaprio, J., Battie, M., Gibbons, L., Peltonen, L. & Koskenvuo, M.
(1998). Polymorphisms of the vitamin D receptor gene associated with intervertebral disc
degeneration. Spine, 23, 2477-2485.
Von Korff, M. & Saunders, K. (1996). The course of low back pain in primary care. Spine, 21,
2833-2839.
Voorhies, R., Jiang, X. & Thomas, N. (2007). Predicting outcome in the surgical treatment of
lumbar radiculopathy using the Pain Drawing Score, McGill Short Form Pain
Questionnaire, and risk factors including psychosocial issues and axial joint pain. The
Spine Journal, 7, 516-524.
Vossen, H., Genis, G., Rutten, B., van Os., J., Hermanes, H. & Lusberg, R. (2010). The genetic
influence on the cortical processing of experimental pain and the moderating effect of
pain status. PloS ONE, 5, doi:10.1371/journal.pone.0013641

202

Wahlstrom, J., Burstrom, L., Nilsson, T. & Jarvholm, B. (2012). Risk factors for hospitalization
due to lumbar disc disease. Spine, 37, 1334-1339.
Walsh, T., Hanscom, B., Lurie, J. & Weinstein, J. (2003). Is a condition-specific instrument for
patients with low back pain/leg symptoms really necessary? The responsiveness of the
Oswestry Disability Index, MODEMS, and the SF-36. Spine, 28, 607-615.
Ware, J., Kosinski, M., Bjorner, J., Turner-Bowker, D., Gandek, B. & Maruish, M. (2007).
User’s Manual for the SF-36v2 Survey (2nd Ed.). Quality Metric Incorporated.
Ware, J.E., Snow, K.K., Kosinski, M., & Gandek, B. (1993). SF-36 Health Survey: Manual and
Interpretation Guide. Boston: Nimrod Press.
Ware, J. (2003). SF-36 Health Survey Update. Retrieved 11/22/2009 online from
http://www.sf-36.org/tools/sf36shtml
Weinstein, J., Lurie, J., Olson, P., Bronner, K., & Fisher, E. (2006). Trends and regional
variations in lumbar spine surgery: 1992-2003. Spine, 31, 2707-2714.
Werneke, M. & Hart, D. (2004). Categorizing patients with occupational low back pain by use
of the Quebec Task Force classification system versus pain pattern classification
procedures: Discriminant and predictive validity. Physical Therapy, 84, 243-254.
Werneke, M., Hart, D. & Cook, D. (2006). A descriptive study of the centralization
phenomenon: A prospective analysis. Spine, 24, 676-683.
Wilson, H., Robinson, J., & Turk, D. (2009). Toward the identification of symptom patterns in
people with fibromyalgia. Arthritis & Rheumatism, 61, 527-534.
Woods, S., Kozachik, S. & Hall, R. (2010). Subjective sleep quality in women experiencing
intimate partner violence: Contributions of situational, psychological and physiological
factors. Journal of Traumatic Stress, 23, 141-150.
World Health Organization (2002). International Classification of Functioning, Disability and
Health (ICF). Retrieved from:
http://www.who.int/classifications/icf/icfbeginnersguide.pdf?ua=1
Yang, Z., Lowe, A., de la Harpe, D. & Richardson, M. (2010). Pactors that predict poor
outcomes in patients with traumatic vertebral body fractures. Injury, 41, 226-230.
Yorio, J., Yan, J., Xie, Y. & Gerber, D. (2012). Socioeconomic disparities in lung cancer
treatment and outcomes persist within a single academic medical center. Clinical Lung
Cancer, 13, 448-457.
Yuan, H., Tang, Y., Liang, Y., Lei, L., Xiao, B., Wang, S. & Xia, Z. (2010). Matrix
metalloproteinase-3 and vitamin D receptor genetic polymorphisms and their interactions

203

with occupational exposure in lumbar disc degeneration. Journal of Occupational Health,
52, 23-30.
Zieger, M., Luppa, M., Meisel, H., Gunther, L., Winkler, D., Toussaint, R., Stengler, K.,
Angermeyer, M., Konig, H. & Riedel-Heller, S. (2011). The impact of psychiatric
comorbidity on the return to work in patients undergoing herniated disc surgery. Journal
of Occupational Rehabilitation, 21, 54-65.
Zhang, Y., Sun, Z., Liu, J. & Guo, X. (2008). Advances in susceptibility genetics of
intervertebral degenerative disc disease. International Journal of Biological Sciences, 4,
283-290.
Zubieta, K., Heitzeg, M., Smith, Y., Bueller, J., Xu, K., Koeppe, R., Stohler, C. & Goldman, D.
(2003). COMT val158met genotype affects µ-opioid neurotransmitter responses to a
pain stressor. Science, 299, 1240-1243.

204