 

THRIS

/‘ ill,

This is to certify that the
dissertation entitled

THE OCCUPATIONALLY INJURED WORKER:
INVESTIGATING THE DECISION TO SETTLE A WORKERS’
COMPENSATION CLAIM

presented by
ANDREW PAUL NAY

has been accepted towards fulﬁllment
of the requirements for the

Doctoral degree in Rehabilitation Counselor
Education

(7'

yMajor Professor's Signature
éwﬁ/g, w/o

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

LIBRARY
Michigan State
University

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATEDUE

DAIEDUE

DAIEDUE

 

010612

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIProj/AccauPresIClRCIDateDue.indd

 

THE OCCUPATIONALLY INJURED WORKER:
INVESTIGATING THE DECISION TO SETTLE A WORKERS’
COMPENSATION CLAIM
By

Andrew Paul Nay

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Rehabilitation Counselor Education

2010

ABSTRACT
THE OCCUPATIONALLY INJURED WORKER:
INVESTIGATING THE DECISION TO SETTLE A WORKERS’
COMPENSATION CLAIM
By
Andrew Paul Nay

According to the Bureau of Labor Statistics (2007), a total of 4.2 million work-
place injuries were reported within the United States in 2005, nearly half of which
resulted in days away from work, job transfer, or restriction. In the majority of workers’
compensation claims, wage-replacement beneﬁts are paid voluntarily by the employer or
insurance provider, and upon the individual’s full recovery or return-to-work, beneﬁts are
stopped (Wolkinson and Block, 1996). And while the majority of workers who sustain
an occupational injury or illness eventually return to their previous job, there are those
non-routine cases where individuals are unable to return to their previous job, or may
never return to work due to the nature of their disability and resulting functional
limitations.

Previous studies have shown that a signiﬁcant portion of workers who require
recuperation away from work never return to the workplace (Hester, Decelles & Gaddis,
1986; National Institute of Handicapped Research [NIDR], n.d.; Tate, 1992; Blackwell et
al., 2003) and tend to experience continued inj ury-related pain. Those workers unable to
return to the workplace in a timely manner typically rely on statutory workers’

compensation systems for medical and ﬁnancial support while they recover from various

illnesses or injuries and cope with existing pain.

The purpose of this study was to investigate and examine a variety of individual
and contextual variables that may affect settlement outcome in workers’ compensation
systems using data maintained by the State of Michigan Workers’ Compensation Agency
(WCA). The focus of this study centers on more complex or non-routine workers’
compensation claims in which the worker has or continues to report a signiﬁcant
impairment or disability and is unlikely to return to their pre-injury occupation. The
social context and relationship model proposed by Dembe (2001) was utilized to
conceptualize both individual and contextual factors associated between various
individuals, groups, and social institutions that contribute to the overall experiences of the
injured worker.

This study empirically examines a sample (N=3,111) of occupationally injured
workers who ﬁled a workers’ compensation claim in the year 2007. This sample
identiﬁes both those who have chosen to settle their workers’ compensation claim
(N=1,354) and those who have not (N=1,843) as of January 2010. A variety of
demographic, socioeconomic, and environmental variables were compared dependent
upon settlement outcome to examine if signiﬁcant differences exist between the two
groups. A binary logistic regression identiﬁed pre-injury wage, time between date of
injury and initial wage-loss payment, and female gender as negative predictors of
settlement, whereas attorney involvement was positive predictor of settlement outcome.

Various implications regarding rehabilitation counseling education and research

are also discussed.

ACKNOWLEDGMENTS

Sincere thanks to my academic advisor, Dr. John Kosciulek, Professor of
Rehabilitation Counseling, and to my dissertation committee members, Dr. Jeff Biddle,
Professor of Economics at Michigan State University (MSU); Dr. Michael Leahy,
Professor and Director of the Ofﬁce of Rehabilitation and Disability Studies at MSU;
Dr. Kenneth Rosenman, Professor of Medicine and Chief of the Division of
Occupational and Environmental Medicine at MSU; and Mr. David Campbell,
Vocational Rehabilitation Consultant of the State of Michigan Workers’ Compensation
Agency, for without all their guidance, insight and expertise this study would not have
been possible. A very sincere thanks to Ms. Kathy Rademacher and Ms. Diane Harger
of the State of Michigan for sharing their time and expertise in analyzing and compiling

data, and Doctoral Candidate, Yun-Jia Lo for her statistical assistance.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... vi
LIST OF FIGURES ............................................................................ vii
CHAPTER 1
INTRODUCTION ................................................................................ 1
Statement of the Problem ............................................................... 2
Theoretical Framework ................................................................. 6
Purpose of Study ........................................................................ 9
Research Questions ..................................................................... 9
Deﬁnition of Terms ..................................................................... 9
CHAPTER 2
LITERATURE REVIEW ...................................................................... 13
Conceptualizing Occupational Disability ........................................... l3
Expectation of Recovery ............................................................... 15
Psychological and Psychosocial Characteristics of Disability .................... 16
Environmental and Social Considerations ........................................... 23
Outcome Studies Of Occupational Disability ....................................... 26
Injured Worker Perceptions and Experiences of Workers’
Compensation ............................................................................ 29
CHAPTER 3
METHOD ......................................................................................... 32
Participants and Instrument ............................................................ 32
Variables ................................................................................. 35
Deﬁnition of Variables ................................................................. 38
Procedure ................................................................................. 45
Data Analysis ............................................................................. 46
CHAPTER 4
RESULTS ......................................................................................... 50
Descriptive Characteristics of the Sample ........................................... 53
Logistic Regression Model ............................................................ 60
CHAPTER 5
DISCUSSION .................................................................................... 67
Signiﬁcant Predictors of Settlement in the Model .................................. 68
Non-Signiﬁcant Variables in the Model ............................................. 73
Limitations of the Study ................................................................ 74
Implications for Rehabilitation Counseling Educators ............................. 78
Implications for Future Research ...................................................... 80
REFERENCES .................................................................................... 84

LIST OF TABLES
Table 1: Predictor and Outcome Variables ..................................................... 37
Table 2: Frequency and Percentage Distribution of Variables .............................. 54

Table 3: Correlation Table of Dichotomous Categorical Variables on Settlement
Outcome ............................................................................................. 57

Table 4: Chi Square Test of Association Among Variables With 3 or More
Categories on Settlement Outcome .............................................................. 59

Table 5: Complete Model for Binary Logistic Regression Analysis: Individual and
Contextual Inﬂuences on Settlement Outcome ................................................ 61

Table 6: Binominal Probability Distribution Table of Predictor Variable on
Settlement Outcome ............................................................................... 63

vi

LIST OF FIGURES

Figure 1: Conceptual Framework of the Study ............................................... 8

vii

CHAPTER 1
Introduction

According to the Bureau of Labor Statistics (2007), a total of 4.2 million work-
place injuries were reported within the United States in 2005, nearly half of which
resulted in days away from work, job transfer, or restriction. In many cases, workers
required recuperation away from work, transfer to another job, restricted physical duties
at work, or a combination of these actions (BLS, 2007). Previous studies have shown
that a signiﬁcant portion of workers who require recuperation away from work never
return to the workplace (Hester, Decelles & Gaddis, 1986; National Institute of
Handicapped Research [NIDR], n.d.; Tate, 1992; Blackwell et al., 2003) and tend to
experience continued injury-related pain. Those workers who do not return to the
workplace typically rely on statutory workers’ compensation systems for both economic
and ﬁnancial support while they recover from various injuries and cope with existing
pain.

Workers’ compensation laws were adopted in the early 1900’s replacing the
common-law tort system as the method for injured workers to maintain relative income
levels when absent from work due to an occupational illness or injury regardless of cause.
Although there is great disparity in beneﬁt rates among the 50 States, on average injured
workers while receiving treatment and attending to their injury only receive about two-
thirds of their actual weekly wage. Some may argue that this was purposely embedded
into the legislation to help limit monetary exposure for the employer, as well as provide
an incentive for the employee’s to return to work as quickly and as reasonably as

possible. While the overall objective of workers’ compensation beneﬁts is to provide

free medical care and wage-replacement beneﬁts to injured workers who cannot initially
return to work (Biddle & Roberts, 2003), its purpose was not intended to provide workers
with sustainable medical and wage-replacement beneﬁts over the course of their lifetime.

While many states have statutory limitations on the amount of time an injured
worker can receive medical and wage-replacement beneﬁts, others states do not. Within
these states, like Michigan, workers who remain out of the workplace following an
occupational injury for a substantial amount of time typically ﬁnd themselves advocating
for continued economic and medical beneﬁts, while insurers and employers seek claim
resolution and closure. This adversarial environment creates an intriguing context for
analysis into the various contextual factors that inﬂuence an injured workers’ decision to
ultimately settle his or her workers’ compensation claim and forgo any future wage-
replacement or medical beneﬁts. In order to lay the foundation for understanding how an
injured worker arrives at the decision to settle his or her claim at a particular point in
time, it is necessary to gain an understanding of how disability is deﬁned within statutory
environments, as well as the settlement process. It also requires an understanding of how
settlement is perceived by the community in which the claim is administered.
Statement of the Problem

The process in which an injured worker brings an initial claim of injury to his or
her employer, and ultimately decides to settle his or her workers’ compensation claim
varies dramatically from case to case. While the decision to ﬁle a claim has been a
subject of analysis (Lakdawalla, Reville, &Seabury, 2005; Biddle &Roberts, 2003;
Rosenman et. al., 2000; Biddle, Roberts, Rosenman, & Welch, 1998), there is a distinct

lack of foundational knowledge regarding the factors that may inﬂuence the decision to

settle a claim. The stigma associated with ﬁling a workers’ compensation claim has been
a frequent subject of analysis given that many times workers struggle with the question of
whether or not to ﬁle an initial claim of injury. Studies have shown that during any given
time, up to 45% of employed workers could ﬁle a work injury or disability claim, but
choose not to (Biddle, Roberts, Rosenman, & Welch, 1998).

As is common among most State-level jurisdictions, disability is an evolving
concept that is ultimately deﬁned by case law rather than by statutory deﬁnition. In
workers’ compensation systems, disability is deﬁned either by the medical or economic
implications that arise out of occupational illness or injury. Presently in Michigan,
disability in workers’ compensation is economically deﬁned rather than by purely
medical or functional capacity determinations. In many cases, even if a worker is unable
to return to their employer following a workplace injury, they may still be capable of
earning similar wages with a different employer and/or within a different occupation thus
limiting their claim of disability.

Under the Michigan Workers’ Disability Compensation Act of 1969, the
employee shall receive and be paid compensation until they return or are offered work
suitable to their qualiﬁcations and training. Ultimately, disability under Michigan
Workers’ Compensation Law means a limitation of an employee’s wage earning capacity
in work suitable to their qualiﬁcations and training (Workers Disability Compensation
Act of 1969, 418.30). This is drastically different from biomedical models of disability
where individuals are considered disabled if they have a deﬁcit, deﬁciency,
dysfunctional, abnormality, failing or medical condition that has been shown to impact

their ability to perform previous occupational demands (Bickenbach, 1993).

In the majority of workers’ compensation claims, wage-replacement beneﬁts are
paid voluntarily by the employer or insurance provider, and upon the individual’s full
recovery or return-to-work, beneﬁts are stopped (Wolkinson and Block, 1996). And
while the majority of workers who sustain an occupational injury or illness eventually
return to their previous job, there are those non-routine cases where individuals are
unable to return to their previous job, or may never return to work due to the nature of
their disability and resulting functional limitations.

Given that most States do not allow for indeﬁnite receipt of wage-loss and
medical beneﬁts, those who are unable to return to their original employer or usual
occupation must ultimately decide how and when to resolve their claim. States such as
Michigan, that do not place a statute of limitations on medical care or wage-replacement
beneﬁts associated with occupation injury or illness, provide an additional setting for
inquiry as to why and when an individual may choose to settle their claim given that no
statutory incentive to resolve or mitigate damages exist.

Independent upon the setting, at any time disputes may arise regarding an
occupational injury claim. As a result, adversarial relationships may develop between
employers, employees, and insurance carriers that may have an impact on communication
between parties, the receipt of wage-loss beneﬁts, and/or authorization and approval for
medical treatment. These disputes typically bring the affected parties (employee and
employer) and various stakeholders (attorneys, insurance representatives, mediators, and
judges) together for settlement negotiations, but other times can result in a lengthy
litigation process. Although most injured workers ultimately return to work, a signiﬁcant

portion of this population will become displaced or unemployed due to their occupational

injury or illness, and in turn will likely struggle with maintaining their economic security
and health care coverage. Occupational disability also has an indirect impact upon the
workers’ family (children, spouse, etc.) as well as the community in which they reside. It
is this particular population that is target of this analysis.

While statutory workers’ compensation systems tend to focus on prompt medical
treatment for injured workers with the initial focus on treating ﬁinctional impairments
following an occupational injury, less focus is given to assessing the many psychosocial
factors associated with predicating behavior. Unfortunately, many of the current models
of disability determination in workers’ compensation fail to discriminate between
“anatomical” and “functional” losses as it would speciﬁcally apply to the workplace
setting (Schultz and Gatchel, 2005), do not pay particular attention to the contextual
factors associated in determining disability, and are insufﬁcient in accounting for various
psychosocial factors that inﬂuence a workers decision to ultimately settle their workers’
compensation claim.

Ultimately the question remains, is settlement an acceptable outcome for the
multiple stakeholders involved in complex or non-routine occupational disability claims?
When claims remain unresolved or open for extensive periods of time, many workers
begin to experience social and ﬁnancial difﬁculties due to only receiving partial wage-
replacement beneﬁts, limited medical care, and potential job-loss. Employers experience
losses in productivity, expensive disability lost-time costs, and increases in risk insurance
premiums. Many would argue that settlement is the desired outcome in states like

Michigan where no statue of limitations on workers’ compensation beneﬁts exist. This

process encourages all parties to negotiate monetary demands, i.e. settlement, that would
allow them to meet their goals and limit future liability, risk, and exposure.
Theoretical Framework

Dembe (2001) presents a useful model of the social context in which occupational
injury and illness and can be conceptualized (Figure 1). This model illustrates the
contextual factors associated with various individuals, groups, and social institutions that
contribute to the overall experiences of the injured worker. More speciﬁcally, it
describes how the relationships between the injured individual, family, work, and
community impact society, and provides a useful tool in conceptualizing the many
contextual factors that inﬂuence individual decision-making within the workers’
compensation setting. While the intended use of this model is to examine a wide range of
social consequences posed by occupational injury, its usefulness for this particular study
is its comprehensive design. Although the relationships depicted within this theoretical
framework are much more dynamic and complicated in practice than portrayed by this
ﬁgure, this model provides a useful framework for examining relationships between the
injured worker and various contextual characteristics that may inﬂuence one’s decision to
settle their workers’ compensation claim.

According to Dembe’s model, the injured worker’s behavior is directly inﬂuenced
by medical professionals, their workplace, work peers, their community, family and
friends, health and social beneﬁt systems, and disability policy and law (workers’
compensation insurance). Additionally, workers are indirectly inﬂuenced by
uncontrollable sociological, economic, and political conditions. The relationships

between various contextual factors and consequences depicted in this ﬁgure are not

mutually exclusive (Dembe, 2001), in fact, they are tightly intertwined making it truly
representative of the complexity of human and social interactions. It is these interactions
that create substantial difﬁculties for future researchers to control for variability in
predicting worker behavior. The dark lines within the model represent the casual
connections between entities and depict the directionality between factors. For example,
there is direct and inﬂuential interaction between the workplace and the injured worker;
however, the worker may only experience the indirect effects of prevailing economic
conditions moderated by his or her employer.

It is hypothesized that an injured workers behavior is signiﬁcantly inﬂuenced
through their interactions and experience within the workers compensation setting. More
speciﬁcally, the decision an injured worker makes regarding the settlement of his or her

claim is inﬂuenced not only by the nature of his or her injury, but is guided by various

 
      
  

 
      
  

Macroeconomic
Conditions

Macrosociological
Considerations

 
    

 
    

(race, income, gender
roles, social class, etc.)

(economic and political
conditions, legal system)

 

 

 

 

 

 

 

 

 

The Workplace Workers’ Groups
(Interaction among workers, ‘ ﬂ (labor unions,
employer, and working advocacy groups,
conditions) co—workers, etc.)
Workers’
Compensation
Insurance

 

   
     
 

Medical
Profession

(medical examiners,

medical treating
doctors, case
managers, healthcare
plans and systems,

Worker Sustains
Occupational
Inj ury/Illness

(claims adjustors,
lawyers, courts,
investigators, etc.)

 

 

 

 

 

 

 

     
     

 

 

 

 

 

 

 

 

 

  
    
       
   
 

 

Behavioral Vocational etc.)
Response Response
Other Beneﬁt 3
Plans 4
(SSDI, LTD, Functional Community
SS], etc.) Response
(schools, church,

 

 

 

housing,
1 education, media,
technologies, etc.)
Worker’s Family and
Friends /
(children, spouse, parents,
friends, and colleagues)

 

 

 

Figure 1. Social Context of Occupational Injury and Illness

sociological and sociopolitical conditions external to the ‘workers’ compensation system.
Due to the overwhelming comprehensiveness of contextual factors associated with this
model, this study will only focus on investigating a variety of environmental (contextual)
variables associated with the interaction between the occupationally injured worker,
health and medical implications, and the workers’ compensation setting itself. Additional
investigation into various individual (demographic) and socioeconomic factors will also
be useful in helping to control for anticipated variability.
Purpose of the Study

The purpose of this study was to investigate what factors or characteristics
inﬂuence an injured worker’s decision to settle his or her workers’ compensation claim.
Research Question

What impact do various individual and contextual variables have on the decision
to settle a workers’ compensation claim?
Deﬁnition of Terms

Administrative Law Judge: See Magistrate

Collective Bargaining Agreement: An agreement negotiated between a labor
union and an employer setting forth all terms of employment for the employees who are
members of that labor union.

Defense Attorney: Licensed legal professional specializing in defending workers’
compensations claims on behalf of self-insured employers and insurance companies.

Disability: A limitation of an employee’s wage earning capacity in work suitable
to his or her qualiﬁcations and training resulting from a personal injury or work related

disease.

Employed: Persons 16 years and over in the civilian noninstitutional population
who, during the reference week, (a) did any work at all (at least 1 hour) as paid
employees; worked in their own business, profession, or on their own farm, or worked 15
hours or more as unpaid workers in an enterprise operated by a member of the family;
and (b) all those who were not working but who had jobs or businesses from which they
were temporarily absent because of vacation, illness, bad weather, childcare problems,
maternity or paternity leave, labor-management dispute, job training, or other family or
personal reasons, whether or not they were paid for the time off or were seeking other
jobs. Each employed person is counted only once, even if he or she holds more than one
job. Excluded are persons whose only activity consisted of work around their own house
(painting, repairing, or own home housework) or volunteer work for religious, charitable,
and other organizations (BLS, 2009).

Indemnity: Recompense for loss, damage, or injuries; restitution or
reimbursement.

Insurance Company: A proprietary organization that traditionally administers all
aspects of workers’ compensation claims for an employer for a fee.

Labor Force: See employed.

Magistrate: An individual employed by the Workers’ Compensation Agency to
make determinations regarding workers' compensation beneﬁts, oversees redemption
settlement hearings, writes legal opinions, and ensures compliance with legal and
statutory requirements.

Nature of Injury: The nature of injury or illness identiﬁes the principal physical

characteristic(s) of the work-related injury or illness.

10

Occupational Injury or Illness: An injury or illness that occurred or is alleged
having occurred at work while in the normal performance of associated work duties.

Occupationally Injured Worker: An individual who receives a personal injury
arising out of and in the course of employment (Workers Disability Compensation Act of
1969, Section 418.30).

Plaintiﬂ Attorney: Legally licensed professional who specializes in personal
injury cases and in representing occupational injured workers in bringing forth workers’
compensation claims.

Settlement: The legitimate termination of wage-loss beneﬁts achieved through a
process of reaching a voluntary agreement between the employee and the employer under
contested or non-contested circumstances.

Redemption: Otherwise referred to as redemption settlement. See Settlement.

Rural- Urban Continuum Codes: Rural-urban continuum codes form a
classiﬁcation scheme that distinguishes metropolitan (metro) counties by the population
size of their metro area, and nonmetropolitan (non-metro) counties by degree of
urbanization and adjacency to a metro area or areas.

Self-Insured Employer: A plan in which the employer itself is responsible for
paying all associated workers’ compensation costs, including wage-loss beneﬁts and
associated health care costs for employees and family members. Self-insured employers
may administer all workers’ compensation payments, or utilize third-party administrators.

Specific Loss: Loss of a body part or appendage as a result of a work related

illness or injury.

11

T hird—party Administrators: An outside ﬁrm i.e., not afﬁliated with the employer,
that maintains records of injured persons covered under a particular insurance plan,
administers workers’ compensation claims, issues notices and provides a variety of other
administrative services in lieu of the self-insured employer.

Unemployed: Persons aged 16 years and older who had no employment during
the reference week, were available for work, except for temporary illness, and had made
speciﬁc efforts to ﬁnd employment sometime during the 4-week period ending with the
reference week. Persons who were waiting to be recalled to a job from which they had
been laid off need not have been looking for work to be classiﬁed as unemployed (BLS,
2009)

Wage-loss: The estimation or calculation of earnings lost resulting from a
workplace injury in United States Dollars.

Wage-Loss Payments: Typically, 80% of after-tax average weekly wage.

Wage Replacement Beneﬁts: See Wage-Loss Payments

12

CHAPTER 2
Literature Review

Once an individual sustains an injury they become exposed to a multitude of new
and sometimes troubling experiences that were never anticipated. Although no two
workers experience an injury in the same way, it is vitally important that we better
understand both the subjective individual factors associated with occupational disability,
as well as the more objective empirical relationships when assessing the likelihood of
settlement. Although there appears to be no previous research that speciﬁcally addresses
the research question posed in this study, a review of related literature on occupational
injury outcomes was undertaken to provide an increased understanding of workers’
experiences within workers’ compensation systems. This provides a useful foundation of
knowledge relative to the purpose of this study and future inquiry into investigating the
individual decision-making within this context.

This literature review will cover a variety salient issues regarding impairment and
occupational disability that include conceptualizing occupational disability, expectation
of recovery, psychological and psychosocial characteristics of disability, environmental
inﬂuences of work disability, outcome studies of occupational disability, and the
experiences and perceptions injured worker’s have regarding the workers’ compensation
system.

Conceptualizing Occupational Disability

The impact that work-related or occupational disabilities have on society is

important to acknowledge. According to the Social Security Administration (2003), a 20

year—old individual who is engaged in work has about a 30% chance of becoming

l3

disabled before the age of retirement. It has also been argued that disability, in general,
has reached such proportions within the United States that it could presently be classiﬁed
as an epidemic (Schultz & Gatchel, 2005). Given this context, it appears useful to gain
an understanding of the types of demographic and environmental variables that inﬂuence
an individual’s decision making following a disabling injury or illness.

The term disability has many connotations and interpretations, and as a result, it is
very difﬁcult to universally deﬁne. Marshaw and Reno (1996) document over 20 current
deﬁnitions of disability that are utilized for purposes of public and private transfers,
eligibility for medical and vocational services, or statistical analysis. This poses a
substantial limitation when attempting to analyze a speciﬁc problem or social epidemic
connotations, such as the impact that occupational injury has on society or an individual’s
behavior. When attempting to investigate an individual’s behavior or decision-making
process, one needs to utilize a framework of disability that is able to encompass the
multiple aspects of individual interactions. Disability is context-speciﬁc and inherent to
the individual (Melhom et. al., 2001) as many perceive, and is at times conditional upon
one’s ability to meet social expectations or function within normalized environmental
conditions (Brandt Jr. & Pope, 1997).

One of the most general, yet useful deﬁnitions of disability comes from Nagi
(1969), in which disability is deﬁned as individual difﬁculty in performing socially
expected activities such as work. This conceptualization of disability recognizes the
importance of the many interactions between the environment and the impairment,
illness, or disability that cause or intensify disabilities (Nagi, 1969), such as workplace,

family, and various insurance system characteristics. This deﬁnition closely supports

14

Dembe’s (2001) comprehensive model of the social context in Which occupational injury
and illness can be conceptualized.
Expectation of Recovery

An injured worker’s expectation of recovery not only appears to inﬂuence their
decision to ﬁle a workers’ compensation claim, but also may actually inﬂuence the time
in which they choose to settle their claim. It is important to address the potential impact
that both psychological and psychosocial characteristics have on injury reporting, as well
as the duration of occupational injuries. For example, studies have shown that acute
conditions related to the current job are no more likely to lead to injury claims than
chronic conditions with long latency periods between exposure and development of
disease (Biddle & Roberts, 2003).

One of the most widely assessed psychosocial characteristic or trait associated
with expectation of recovery has been motivation. It should be noted that motivation is a
very difﬁcult concept to quantify or operationally deﬁne and requires further empirical
validation (Schultz et al., 2004). It has been found, however, that motivation to return to
work following an occupational injury has been shown to have close associations with
expectation of recovery (Schultz, 2005). It therefore remains plausible that an
individual’s expectation to recover from an occupational injury and return to work would
be inﬂuenced by their experiences within the medical care system and the workers’
compensation setting.

It can be further argued that an injured worker’s motivation to resume work
following a low back injury can be inﬂuenced by expectation of recovery, coping skills,

and perceived value of work relative to the associated costs and social impact of the

15

injury. Schultz et al. (2003) has shown that expectation of recovery was found to be
predictive in relation to work status, duration of disability and claims cost. Cole et a1.
(2002) found recovery expectations to be predictors of total time receiving beneﬁts.
Further research has shown that “positive expectations of recovery were associated with
reduction in pain grade and enhancement of functional status” and provided strong
evidence of the importance of individual behavioral factors in the prediction of pain-
related occupational disability (Schultz et al., 2003).

Psychological and Psychosocial Characteristics of Disability

Workers with occupational injuries experience numerous long-term consequences
due to their injuries that include a physical, psychological, and psychosocial conditions.
Research has shown that “perceptions of disability, particularly related to health change,
were predictive of disability including return to work and duration of disability” (Schultz
et al., 2004). According to Linton (2000) the evidence indicated that, “self-perceived
poor health is related to chronic pain and disability”. Several studies and systematic
literature reviews conﬁrm this conclusion (Carey et al., 2000; Crook et al., 1998; Crook
et al., 2002; Gatchel et al., 1995; Tate et al., 1999). Schultz et a1. (2004) also found that
“perception of health change due to back pain was a more powerful and consistent
predictor of outcomes than overall perception of disability.”

The causes of chronic pain disability remain controversial and evidence points to
psychological factors as more important than physical factors (Gallagher et al., 1989).
Gallagher et a1. (1989) explored the effects of compensation and litigation on outcome in
chronic pain patients in order to investigate whether or not signiﬁcant relationships were

present in a sample of unemployed individuals with chronic low back pain. One purpose

16

of this study was to investigate prediction of outcome and change in compensation status
over time, which would allow a researcher to identify if a signiﬁcant relationship existed
between wage-loss compensation and return to work. Gallagher (1989) found that
compensation was not signiﬁcantly associated with return to work, which opposes the
moral hazard theory inferring that in-kind beneﬁts or wage replacement compensation
fosters malingering and dishonestly among workers to exaggerate the extent of their
injuries, and provides an economic disincentive to return to work (Butler & Worrall,
1991)

Research has also shown that certain psychosocial factors also appear to be risk
factors for work disability and have identiﬁed that workers have less favorable outcomes
when they blamed work factors for their condition, rated their relations with co-workers
as poor, reported on-the-job psychological distress, or had low expectations of recovery
(Schultz & Gatchel, 2005).

It has been argued by many researchers that no single linear model can be applied
to fully encompass one’s lifetime adjustment to disability regardless of the context in
which the disability occurs (Kendall & Buys, 1998). Contemporary research has also
shown that there are dramatic differences in the direction and duration that one moves
through the psychosocial adaptation process (Kendall & Terry, 1998). Based on the
complex nature of the individual personality and the unique occupational characteristics
associated with employment, it would be difﬁcult to subscribe to a single linear model of
psychosocial adjustment to work disability. Given the added variability of workers’

compensation laws across states and varying degrees of stakeholder cooperation and

17

participation in the system, it would be increasingly difﬁcult to support a linear
interpretation or model of psychosocial adaptation within this setting.

According to Livneh (2001), no evidence has been found to support the existence
of a linear model of psychosocial adjustment to back (spinal cord) injury. Whether
addressing spinal cord injuries or chronic illness, advocates of psychosocial adjustment
models believe that human adaptation to unexpected disability involves continuing
processes of assimilation to the resulting changes in body functioning, self-concept, and
person-environment interaction. While some psychosocial adjustment models subscribe
to an unordered and individually unique adaptation process of learning to recover (Livneh
1997), occupational disabilities introduce additional variability that poses signiﬁcant
challenges for the injured worker that may not easily be assessed.

Reactions and Responses. Unlike for those individuals who sustain injuries
outside of the workplace, those who are injured out of and in the course of employment
will be involuntarily exposed to models of medical treatment that have been developed
by policy makers within the workers’ compensation system. Following an injury, the
worker is immediately introduced to a variety of employer-guided rehabilitation efforts
that may include, but are not limited to, independent medical examinations, nursing visits
and managed care, vocational assessments or rehabilitation efforts, mediations, and
possibly formal hearings regarding their claim. With limited understanding and relatively
little decision-making authority, more severely injured workers will many times seek
legal representation. This introduces additional stakeholders into the process increasing

the likelihood that the workers’ compensation setting will become more complex and

18

adversarial. It is possible that this adversarial setting introduces additional variability
into the psychosocial adjustment phase of adaptation to disability.

Anxiety. Anxiety for injured workers is often associated with the unknown nature
of how a particular injury or disability is going to impact work-life. Anxiety responses
for persons who have sustained long lasting or chronic medical conditions following an
occupational injury will face an undetermined level of uncertainty about their future. As
Livneh (2001) describes, anxiety responses may be associated with abrupt or unexpected
changes in lifestyle, anticipation of medical procedures, and recurrence of symptoms. An
additional psychosocial variable that may impact subsequent levels of a workers’ anxiety
is that of expectation of recovery. This psychological factor was intensively explored by
Schultz et a1. (2004), where it was found to be directly associated with anxiety and
predictive in identifying relation to work status and duration of disability.

As injured workers progress through the workers’ compensation system, there is
increasing emphasis placed on timely recovery and return to work. Some of the most
difﬁcult claims involve injured workers whose expectations of recovery have never been
fulﬁlled. It has been shown that workers typically perceive work-related back problems
as life-long problems (Tarasuk & Eakin, 1994); therefore, it is likely that as pain
conditions associated with occupational disability progress, anxiety will directly
contribute to the workers’ perception of his or her disability (Schultz, 2004).

There is no predicting the amount of anxiety an injured worker will experience as
a result of their disability. There is no guarantee that injured workers will ever achieve
their post-injury expectations. While some workers are quickly able to adapt to their

disability and move forward with their life, others ﬁnd themselves making more difﬁcult

19

transitions within the psychosocial adaptation process. This is evident in cases where two
individuals (Worker A and Worker B) of similar age and body-type sustain similar
injuries and undergo the same course of medical treatment. Where Worker A
experiences a successful recovery and returns to work, Worker B never returns to work
and experiences lifelong limitations in social functioning.

The adversarial nature of the workers’ compensation setting can be stressful for
the injured worker. Anticipating mediations and court appearances, visiting unfamiliar
medical examiners and doctors, and the search for new employment may all contribute to
added levels of stress and anxiety. And although the workers’ compensation system was
devised to protect and treat workers who had sustained workplace injuries, recent studies
have shown many workers’ experience additional anxiety and frustration (Strunin &
Boden, 2004).

Ethnographic interviews were conducted with 402 randomly selected workers
from Florida and Wisconsin, all of who were receiving workers’ compensation beneﬁts
following an occupational lower-back injury (Strunin & Boden, 2004). Although the
study showed that each of the workers reported different experiences within the workers’
compensation system, many reported having little or no control over their case, had little
understanding of the system, received pressure to return to work prematurely, and had
demeaning relationships with some their claims representatives (Strunin & Boden, 2004).
It has been suggested that the resulting stress related to the limitations imposed by a
worker’s occupational injury or disability, and their general lack of understanding about
the workers’ compensation system, could be directly related to levels of anxiety

experience by injured workers (Strunin & Boden, 2004).

20.

Depression. Some of the most extensive research regarding those who have
sustained back injuries has been focused on depression (Livneh, 2004). Workers who
have sustained an occupational injury have shown to experience a multitude of long-term
consequences speciﬁcally related to their injuries (Pransky et al., 2000). Research further
suggests that occupational injuries have a variety of effects on psychological well-being
and behavior (Dembe, 2001). Contributing to these ﬁndings, self—reported feelings of
depression were also found to be higher among individuals with work-related repetitive
strain injuries (Keogh et al., 2000) than those who sustain injuries outside of the
employment setting.

One explanation for this phenomenon may be linked to Erickson’s stages of
psychosocial development where the middle-aged adults are consistently challenging
themselves to establish and guide the next generation through socially valued work. If,
for instance, an injured worker is unable to work and demonstrate expressions of
generativity (a concern for people besides oneself and family) to family, friends, and co-
workers, depression may develop. Another explanation may be that the lack of control
and decision-making reported by injured workers within the workers’ compensation
system fosters feelings of hopelessness and uncertainty.

Additional studies have shown that the development of depression following a
work-related injury may delay physical recovery and the return of normal ﬁrnctioning
(Sullivan et al., 2005, Lancourt & Kettlehut, 1992). Delays in medical treatment do not
support the underpinnings of the workers’ compensation system where the timeliness of
services and return to work are highly stressed. Workers who are unable to return to

work after sustaining an occupational injury undergo numerous psychological and

21

behavioral changes that may affect their domestic role, daily living activities, vocational
function, and impact their household economic situation. This situation can be
particularly problematic for individuals involved in litigation of their worker's
compensation claims (Greenough et al., 1994), presumably because the adversarial
approach to settling work injury disputes contributes to their overall level of distress and
duration of time away from work. This more formal process of litigation may also create
extensive delays in rehabilitation service delivery creating delays in treatment, work
absence, and lost wages.

Workers’ compensation claimants who experience more serious or complex
injuries often become alienated from their workplace. Communication with the employer
and co-workers is frequently severed during dispute resolution process leaving many
workers with an overall of feeling alienation and dejection. Many employers are unable
to accommodate the disabled worker’s physical or medical restrictions. Such an
occurrence is much more common for more severely disabled workers employed with
small employers, but is also becoming more common within large ﬁrms who are
downsizing, restructuring, or relocating. Various provisions contained in collective
bargaining agreements may also preclude injured workers from returning to work under
certain circumstances, and many times do not allow for individuals to return to work if
under medical restriction.

When workers sustain more severe workplace injuries, studies have shown that
they perceive their physical abilities to be more limited than before their disability and
the odds of returning to work as low (Burns et al., 1995). Injured workers tend to

develop negative thoughts about their post-injury self due to lack of encouragement and

22

may ﬁnd it more difficult to accept or even understand their situation. These behavioral
characteristics can often have a greater inﬂuence on the length of recovery than factors
such as the severity of the injury (Currie & Wang, 2004).

Depression among occupationally injured workers appears to be more commonly
associated with those individuals who are experiencing chronic low back pain as
compared with other types of injury or illness (Currie & Wang, 2004). Ash and
Goldstiein (1995) found that workers’ compensation claimants who had sustained
occupational back injuries were signiﬁcantly less likely to return to work following
vocational rehabilitation efforts than those with less severe depression. In another study,
Hildebrandt et a1. (1997) found that depression appeared to be associated with both acute
and chronic pain, and had a negative inﬂuence on return to work. Depression may also
be more common among those injured workers who are unable to return to their previous
employer and are involved in adversarial workers’ compensation claims that have a
negative impact on their economic situation.

Environmental and Social Considerations

The social signiﬁcance of acquiring a work disability has been a frequent topic of
discussion and research in occupational and industrial medicine literature; however, the
topic has received less attention in the rehabilitation counseling literature in recent years.
What appeared to be absent from the literature reviewed is the many contextual factors of
psychosocial adaptation to disability that may be present within adversarial rehabilitative
settings. Disability and chronic pain not only affects an injured worker’s prospective
work status, but impacts their domestic and societal situation as well. Acquired

occupational disabilities in physical or cognitive functioning inﬂuence the direction and

23

duration of the psychosocial adjustment process. Previous studies have shown that
occupational injury and illness can substantially inﬂuence one’s behavior and well-being
(Dembe, 2001) throughout their entire lifetime.

Stone (1984) and Dembe (2001) have identiﬁed work disability as a complex and
socially predicated event that results from an occupational injury or illness, that not only
is determined by physical limitation, but by a variety of social considerations as well.
Livneh’s (2001) theory identiﬁes these social considerations as contextual factors that are
associated with the overall psychosocial adjustment to disability regardless of disability
type. By applying these conceptual models and theories to the context of workers’
compensation and occupational disability, it may be possible to identify speciﬁc job
characteristics that can help us better understand the impact adversarial settings may have
on the duration of work disability.

Socio-Demographic and Job Characteristics. Much less information is available
regarding the inﬂuence that particular job duties or employment settings have on an
injured worker’s psychosocial adaptation to work disability. It is plausible that certain
environmental and job characteristics will impact an individual’s psychosocial adaptation
to work disability and eventual return to work. Some of the most frequently referenced
research articles regarding socio-demographic inﬂuences on work disability and duration
of disability within workers’ compensation systems are by Allen Cheadle and colleagues
(1994). These researchers found that one consistent and signiﬁcant predictor of duration
of disability, deﬁned as time away from work, has been the size of the employer.

Employers of ﬁfty (50) employees or less have much higher incidence rates of worker

24

absence due to disability (Cheadle et al., 1994) suggesting that larger employers have
shorter durations of worker disability.

It was also found that injured workers from rural areas returned to work less often
than those workers who had access to metropolitan labor markets (Cheadle et al., 1994).
Longer durations of disability were much more common in areas with high
unemployment rates (Volinn et al., 1988) suggesting that persons with work injuries or
disabilities experience more difﬁculty in ﬁnding employment within competitive labor
markets than those without functional limitations.

Signiﬁcant and positive relationships have been shown to exist between older
workers who sustain an occupational injury and time away from work Cheadle et al.
found that the effect on age is due both to the reduced ability of older workers to
physically recover following any injury, and the reduced ability to ﬁnd suitable
employment once recovered (1994). Additionally, a cohort longitudinal study of non-
speciﬁc occupational back injury samples was able to signiﬁcantly indentify the number
of years with employer’ as a key indicator of longer durations of disability (Schultz et al.,
2004.). While some studies have been able to identify positive correlations between age
and time away from work due to disability, it is difﬁcult to identify empirical research
that addresses how a worker’s age affects their psychosocial adaptation and decision
making relative to occupational disability.

Research into the marital status of injured workers who were receiving worker’s
compensation beneﬁts have shown that incidences of longer periods of work disability
are higher among those who are divorced than those who are married at the time of injury

(Cheadle et al., 1994). This ﬁnding has also been supported in other studies that have

25

shown signiﬁcant relationships between the duration of disability and martial status of the
injured worker (Bolton, 1974; Volinn et al., 1991). These results may suggest that
married workers who have sustained an occupational injury may have spousal support
that aids in their psychosocial adjustment process, unlike divorced workers who must
manage their work disability independently.
Outcome Studies of Occupational Disability

Only two previous studies could be identiﬁed that investigated settlement
probability as a dependent variable and empirically examined a sample of workers’
compensation claims originating in New York State. The researchers investigated the
probability that the claimant and insurer (insurance company or self-insured employer)
would negotiate a settlement between parties, rather than proceed to trial for a disposition
(Thomason & Burton, 1993). Overall, it was identiﬁed that certain insurer adjustment
activities (e.g., reductions, suspensions, and/or ﬂuctuations in wage-loss beneﬁts,
approval or disapproval of medical procedures, monitoring, etc.) increased settlement
probability. The results of the regression model in predicting settlement probability
identiﬁed pre-injury weekly wage and time away from work as positively inﬂuencing the
amount of settlement. Interestingly, age was found to be positively correlated with
settlement amount implying that the older an individual is at the time of injury, the higher
the negotiated settlement.

This same study also produced results that indicate that attorney involvement
increases the amount of insurer adjustment activity and in turn increases the probability
of settlement, although decreases overall settlement size (Thomason & Burton, 1993).

Previous outcome research in workers’ compensation has found that settlement

26

probability is negatively correlated to expected settlement award size (Thomason, 1989),
implying that the more money demanded by the injured worker or attorney, the less likely
they are to reach a voluntary settlement agreement.

Other studies have shown a lack of association with post-settlement outcomes and
case resolution. Theoretically, case resolution should predict post-settlement outcomes,
treatment and temporary disability costs should reﬂect severity of injury, and disability
rating and settlement award should reﬂect residual disability (Chibnall et al., 2005). In
the Chibnall et al. study, case resolution was found to be predictive of self-reported
disability (2005). As previous studies support, case resolution appears to reﬂect levels of
medical testing and treatment decisions (e.g., rehabilitation, surgery, etc.) rather than
injury severity and residual disability (Chibnall et al., 2000).

Previous research exploring the relationship between litigation status and the
symptoms of the injured worker has been inconsistent and limited by methodological
difﬁculties. Prior to 1995, the relationship of workers' compensation and litigation to low
back pain outcome was not established in the literature (Gallagher et al., 1995). Suter
(2001) surveyed two hundred (200) chronic back pain participants at a pain center to
further investigate the relationship between being involved in litigious settings and pain
symptomology. Patients were divided into four groups, namely a non-litigating/non-
working group, a non-litigating/working group, a litigating/non-working group and a
litigating/working group. Overall, participants who were working scored lower on all the
measures than did participants who were not working (Suter, 2001). On the other hand
participants who were involved in litigation, scored higher on all the measures of

disability than did participants who were not litigating (Suter, 2001). The researcher

27

concluded that the results imply that injured workers were more likely to litigate if they
were experiencing higher levels of pain than normal, or were more likely to exaggerate
their pain levels while involved in litigation to maximize monetary awards. As such, this
study produced ﬁndings that supported both litigation and employment as signiﬁcant
factors inﬂuencing worker decision-making and recovery from injury.

It has also been shown that successful return to work by injured workers is
seriously endangered by the complexity of factors and dynamics associated with existing
disability systems (Tate, 1992). Those who were rejected by their employer following a
workplace injury engendered hostility, and fostered emotional resistance to rehabilitation
(Eaton, 1979). Additional studies have shown that locus of control or the level of control
perceived by worker over a particular situation, has been shown to be related to return to
work success (Beck, 1988), and injured workers who are receiving workers’
compensation beneﬁts show greater resistance to physical recovery than do those
individuals with similar injuries not receiving workers’ compensation (Brink, 1989).

Earlier ﬁndings suggest a number of factors that might inﬂuence successful return
to work outcomes for workers injured on the job. Factors such as workers’ compensation
beneﬁt systems, timelines of vocational rehabilitation systems, worker injury types,
demographic characteristics, and attorney involvement (Hester et al., 1986; Gardner,
1991; Blackwell, 2003 in Ash & Goldstein, 1995; Gurnerman, 1998) have all been shown
to have varying impact on return to work success following occupational injury. It has
also been shown that those injured workers who lost more work time than injured
counterparts (co-workers) experienced signiﬁcantly more economic problems due to their

injury (Pransky et al., 2000).

28

Injured Worker Perceptions and Experiences of Workers’ Compensation

Several studies, both qualitative and quantitative, have documented the
experiences of injured workers receiving or attempting to receive beneﬁts through the
workers' compensation system (Dembe, 2001). One of the most extensive qualitative
studies into the workers’ interactions with the workers’ compensation system was
conducted by Strunin & Boden (2004) who found that while experiences can be positive,
and beneﬁts to injured workers can be provided in a timely manner by considerate and
efﬁcient processes, there was an overall consensus among workers that the workers’
compensation system was experienced as “cumbersome, frustrating and demeaning.”

Studies have shown that workers generally report negative experiences within the
workers’ compensation system, a system that has been perceived by many as uncaring,
unfair, and adversarial (Reid et al., 1991; Imershen et al., 1994). Additional studies have
shown that the stress that injured workers’ experience related to their occupational injury
or illness can be caused by emotional instability related to a lack of understanding of the
workers’ compensation system and their relationships with compensation claims
representatives (Strunin & Boden, 2004).

Additional data indicates that injured workers may be reluctant to report
occupational injuries and illnesses for a variety of reasons. Examples include, but were
not limited to, fear of reprisal, belief that pain is an ordinary consequence of work
activity or aging, lack of management support after prior reports, and a desire not to lose
their job (Jefferson and McGrath, 1996; Silverstein et al., 1997; Pransky et al., 1999;

Rosenman et al., 2000). Studies have shown that many workers voluntarily choose not

29

to ﬁle a workers’ compensation claim at all based on fear of employer reprisal (Pransky,
2001)

Biddle & Roberts (2003) explored the determinants of the decision to ﬁle for
workers' compensation using a sample of individuals identiﬁed by a physician as having
a work-related illness. They utilized a Michigan governmental database of reports of
known or suspected cases of occupational illness and matched them with workers'
compensation claims data. They found that the percentage of reported occupational
injuries for which workers also ﬁled for workers’ compensation beneﬁts was between 9%
and 45%. Similarly, a study conducted in Connecticut estimates that just 10.6% of
workers with work related chronic upper extremity pain ﬁle a workers' compensation
claim for their condition (Morse et al., 1998). Although workers’ compensation insurance
is purchased by employers to pay for all medical care related to work-related injuries and
illnesses, many persons suffering work-related disorders utilize their own health
insurance or other private or public assistance programs to pay for medical care and other
expenses related to their injury, and is even more prevalent in cases where their initial
claims are denied (Dawson, 1994; Texas Workers' Compensation Research Center, 1995;
Morse et al., 1998; Keogh et al., 2000). It has also been found that sometimes injured
workers incur substantial out-of-pocket expenses for treatment of job-related disorders
(Morse et al., 1998; Galizzi et al., 1998), conditions which would otherwise be covered
under workers’ compensation insurance if reported.

Lastly, studies have also shown that injured workers experience signiﬁcant
economic hardship as a result of monetary caps on compensation beneﬁts under workers'

compensation legislation (Boden and Galizzi, 1999; Reville, 1999). These caps force

30

injured workers who remain out of work for substantial periods of time to deplete savings
and borrow money, all of which can contribute to additional stress and anxiety related
disorders (Texas Workers' Compensation Research Center, 1995; Morse et al., 1998;
Keogh et al., 2000; Pransky et al., 2000). For example, a group of workers who were
injured in Pennsylvania reported that they had to wait an average of 20 months after a
workplace injury to get income replacement beneﬁts because of delays in approving and

processing their workers’ compensation claims (Dawson, 1994).

31

CHATPER 3
Method
Participants and Instrument

The population of interest in this study is the occupationally injured worker in
Michigan. “Occupationally injured” is deﬁned as an individual who receives a personal
injury arising out of and in the course of employment (Workers Disability Compensation
Act of 1969, 418.30). While it is the intention of this study to provide empirical
explanations regarding those who have voluntarily resolved their workers’ compensation
claim, it requires an examination of a sample of workers whose claims have been
legitimately terminated (e.g., settlement, returned to work with no wage-loss, recovered
from disability, end of speciﬁc loss, deceased, lost claim etc.), as well as those who have
not returned to work and continue to receive wage-loss beneﬁts.

For purposes of this study, the qualifying term “settlement” refers to the
legitimate termination of wage-loss beneﬁts. This process of reaching a settlement
agreement between the employee and the employer may occur under either contested or
non-contested circumstances, and are typically negotiated between the employer and the
employer, or their representatives (attorneys, insurance companies, third-party
administrators, etc.). They may also be facilitated by mediation hearings within local
jurisdictions nearest the injured worker’s residence. If a claim is not concluded through a
redemption settlement under non-contested circumstances, it is brought before a special
magistrate or administrative law judge for determination.

Ultimately, the overall function of a settlement is to specify the terms and

conditions of a ﬁnal settlement agreement, including all aspects of an injured worker's

32

claim. This includes disputed wage-loss as a result of the injury and any present and/or
future medial costs. The settlement contract addresses all elements of the existing or
alleged claim, and once the settlement agreement is executed, the worker can no longer
seek additional compensation from the employer. In cases where the employee and
employer are involved in some dispute regarding the claim, it is likely that the employee-
employer relationship will be terminated. In non-adversarial claims, the employment
relationship is preserved in most cases. For purposes of this study, this includes all cases
where the injured employee returns to work with no wage loss and has been medically
cleared to return to work (i.e., recovered from injury or illness).

The State of Michigan requires that an employer shall report immediately report
all injuries, including diseases, which arise out of and in the course of the employment, or
on which a claim is made, and which result in disability extending beyond seven (7)
consecutive days, not including the date of injury, death, or speciﬁc losses. This
information is initially documented by an employer on a form titled: Employer’s Basic
Report of Injury and contains a variety of employee, employer, injury, medical,
occupational, and wage data.

All injury data reported to the State of Michigan is maintained in a variety of
administrative databases used for maintaining authority and guidance over each claim.
These databases are primarily used to collect vast amounts of information regarding past
and present workers’ compensation claims for program evaluation and public policy
development and include the identiﬁcation of all workers who were determined eligible
and successful in ﬁling a workers’ compensation claim regardless of wage-loss or

medical beneﬁt determination of liability. These data contain valuable and unique

33

information about the characteristics, activities, and results of each workers’
compensation claim that can be effectively used for research purposes.

This study utilizes a retrospective design to investigate the decision-making
process of injured workers in Michigan who reported a work-related injury resulting in a
claim for workers’ compensation beneﬁts during the 2007 calendar year. As of January
2010, a total of 42,708 individuals had reported a work-related injury having occurred in
the year 2007. This period encompasses the timeframe of claim inception (1/1/07 through
12/31/07) to the date of data extraction (1/ 1/ 10). This provides a useful window-in-time
to examine a sample of individuals over a period of 2 to 3 years in which generalizations
to the larger population can be explored. This also provides the means to collect a
sample of non-routine or complex occupational disability claims where the worker has
been absent for work for an extended period of time due to their occupational injury or
illness.

In order to gather data for this study, an initial dataset was produced by
identifying all individuals who reported a date of injury occurring in 2007 (N=42,708).
Next, all claims ﬁled in which the worker did not incur wage loss or lost time away from
work during the claim period (e.g. medical-only, no-time-lost, etc.) were eliminated
(N=17,365). Following the elimination of no-lost time and no wage-loss claims, a total
of 25,343 claims remained. Since this study is most concerned with examining the
characteristics and implications of individuals who required recuperation or recovery
away from the workplace following an occupational injury or illness, only claims in
which a settlement agreement had been executed (N=l311) and claims that were

continuing to pay wage-loss beneﬁts at the time the sample data was extracted (N= 1843)

34

were retained. The ﬁnal dataset consisted of a total sample size of 3,111 participants who
had either settled their claim or were continuing to receive wage-loss beneﬁts as of
1/1/ 10.
To improve the homogeneity of the sample and control for anticipated variability,
all sample members must have met the following criteria:
1) reported a compensable work-related injury that occurred in the
calendar year of 2007,
2) ﬁled a Form BWC-100 (Basic Report of Injury) with the State of
Michigan for wage-loss beneﬁts,
3) was absent from the workplace due to the reported injury for a
minimum of 7 days, and
4) are continuing to receive wage-loss beneﬁts, or have settled their
workers’ compensation claim
Variables
A variety of variables were selected for purposes of this analysis because they
were considered relevant in terms of identifying salient differences and commonalities
between claimants who decided to settle their workers’ compensation claim, and those
whose did not. This study will focus on investigating both categorical and continuous
variables that include various individual and contextual characteristics (see Table 1).
Additional covariates were also explored, such as amount of time elapsed between the
date of injury and the last day worked, wage loss beneﬁt, and average weekly wage.
The dichotomous outcome variable is the decision to settle 3 workers’

compensation claim, classiﬁed by a unique closure code within the sample dataset, which

35

allows the researcher to identify those who have settled their claim, coded as “R”. This
allows for the creation of a dichotomous outcome variable (R=1, Other=0), which
distinguishes between the groups of interest. Similarly, each claim is assigned a unique
claims status code that allows the researcher to identify those claims that are continuing
to pay wage-loss beneﬁts (Coded as PAY), versus those who have resolved their claim
(coded as PAID). Based on the rationale associated with the purpose of this study and
sample dataset creation, these two mutually exclusive groups were combined together
(N=3111) to create the ﬁnal research sample of those who settled their claim and those

who continue to receive wage-loss beneﬁts 2 to 3 years post-injury.

36

Table 1: Predictor and Outcome Variables

Individual (Demographic) Characteristics
Age (Under or Over 55)
Gender
Marital Status
Number of Dependents
Body Part of Injury
Nature of Injury

Contextual (Socioeconomic and Environmental) Characteristics
Pre-Injury Average Weekly Wage (AWW)
Timeliness of Wage Loss Beneﬁt (in days)
Geographic Area
Attorney Involvement

Outcome Variable

 

Settlement (Yes or No)

37

Deﬁnition of Variables

As explained in detail below, a variety of recoding procedures were carried out
prior to ﬁnal model development in order to group these data into categories that make
sense in the context of this research.

Individual Characteristics.

Age. An independent, continuous variable identifying the amount of time elapsed
in whole years between and individuals recorded date of birth and date the sample dataset
was created (January 1, 2010). The age variable was also coded as an independent,
dichotomous variable identifying workers under the age of 55 (coded as 0), and workers
over the age of 55 (coded as 1). This allows the researcher to make comparisons between
individuals of average working age and older workers at or above the age of 55.

Gender. An independent, dichotomous variable designating male versus female
as recorded on the individual’s Form BWC-100 (Basic Report of Injury). Gender was
coded 1 for male, and 2 for female.

Marital Status. An independent, dichotomous variable differentiating between
those who ﬁlled as married or single on the Form WC-701 (Notice of Compensation
Payments) as of the date of the reported injury. Each Form WC-701 is completed by the
party providing wage loss compensation (e. g., self-insured employer, insurance company,
third party administrator, etc.) and identiﬁes the worker’s tax ﬁling status among four (4)
categories of single, single/head of household, married/ﬁling joint, or married/ﬁling
separately. This allows the researcher to create a dichotomous variable of those who are

Single (coded as 0) and those who are married (coded as 1).

38

Number of Dependents. The original designation of number of dependents is a
continuous variable ranging from zero (0) to twelve (12) within the sample dataset. An
independent, dichotomous variable differentiating between those who reported zero (0) or
one (1) dependents (coded as 0), and those who reported two (2) or more dependents
(coded as 1) as identiﬁed on the form WC-701. This ﬁgure was identiﬁed as of the date
of injury and cannot account for changes in the number of dependents that may have
occurred (e.g., birth of child, divorce, death of spouse, etc.) subsequent this date. This
categorical variable was created in order to investigate possible differences or
considerations in the decision-making between injured workers with dependent children
(claiming 2 or more dependents) and those without.

Body Part of Injury. An independent, categorical variable consisting of eight (8)
categories based on the Occupational Injury and Illness Classiﬁcation System (OIICS).
The OIICS was developed by the Bureau of Labor Statistics (BLS) to provide a
standardized coding system for characterizing work-related injuries and illnesses (Bureau
of Labor Statistics, 2007). The part of body affected code structure is arranged in order
from the top of the body (Division 0, Head) to the bottom of the body (Division 4, Lower
Extremities). Division 5 classiﬁes body systems. Division 8 classiﬁes multiple body
parts from two or more divisions. Division 9 classiﬁes prosthetic devices and any human
parts of the body which are not classiﬁed or listed under any of the other divisions (BLS,
2007)

Each part of body was classiﬁed according to OIICS thus allow for the creating of
7 dummy variables (1 = neck, including throat, 2 = trunk, 3 = upper extremities, 4 =

lower extremities, 5 = body systems, 8 = multiple body parts, 9 = other body parts, and

39

9999 = nonclassiﬁable). The variable “head” was designated the constant, coded as 0.
For purposes of this study, the categories “other body parts” and “nonclassiﬁable” were
combined into one category due the limited number in each category (N=34). The part of
body affected identiﬁes the part of the body directly affected by the previously identiﬁed
nature of injury or illness (OIICS, 2007).

Nature of Injury. An independent, categorical variable originally consisting of
eight (8) categories based on the Occupational Injury and Illness Classiﬁcation System
(OIICS). The nature of injury or illness code structure is arranged so that traumatic
injuries and disorders are listed ﬁrst (category 0) while diseases are listed in categories 1
through 8 (1 = systemic diseases or disorders, 2 = infectious and parasitic diseases, 3 =
neoplasms, tumors, and cancer, 4 = symptoms, signs, and ill-deﬁned conditions, 5 = other
conditions or disorders, and 8 = multiple diseases, conditions, or disorders). Category 8
classiﬁes multiple physical characteristics whose individual codes are found in categories
1 through 5, and those coded 9999 are nonclassiﬁable (OIICS, 2007).

For purposes of this study the nature of injury variable was reduced and recoded
into three (3) categories based on OIICS thus allowing for the creating of two (2) dummy
variables (i.e., 1 = systems diseases or disorders and 2 = all other). The variable
traumatic injuries and disorders was designated as the constant. This variable includes all
traumatic injuries and disorders, effects of external agents, and poisoning is the result of a
single incident, event, or exposure (OIICS, 2007). System diseases or disorders refer to
toxic and non-toxic diseases or disorders affecting systems of the body that may affect
the whole named body system, or more commonly, affect only a part of the named body

systems at the time of diagnosis. Given the limited number of claims involving nature of

40

injuries involving categories 3 through 5, 8, and 9999, these various natures of injury
were combined to account for the category “all other”.
Contextual Characteristics

Pre-Injury Average Weekly Wage (A W W). A continuous, independent variable
specifying the calculated total gross weekly wage (highest 39 of 52 weeks) that a worker
receives in the normal performance of their work in rounded United States Dollars
(USD). As deﬁned by the Michigan Workers’ Disability Compensation Act of 1969,
AWW refers to the weekly wage earned by the employee at the time of the employee’s
injury in all employment, inclusive of overtime, premium pay, and cost of living
adjustment, and exclusive of any fringe or other beneﬁts which continue during the
disability (§418.371).

If the employee worked less than 39 weeks in the employment in which the
employee was injured, the AWW shall be based upon the total wages earned by the
employee divided by the total number of weeks actually worked. If an employee sustains
a compensable injury before completing his or her ﬁrst work week, the average weekly
wage shall be calculated by determining the number of hours of work per week
contracted for by that employee multiplied by the employee’s hourly rate, or the weekly
salary contracted for by the employee. If the hourly earnings of the employee cannot be
ascertained, or if the pay has not been designated for the work required, the wage, for the
purpose of calculating compensation, shall be taken to be the usual wage for similar
services if the services are rendered by paid employees (§ 418.371).

If there are special circumstances under which the AWW cannot justly be

determined by applying the previous rules, an AWW may be computed by dividing the

41

aggregate earnings during the year before the injury by the number of days when work
was performed and multiplying that daily wage by the number of working days
customary in the employment, but not less than 5 days (§ 418.371). This amount is
initially identiﬁed on the Form WC-100.

The AWW variable was used to create an independent, categorical variable
determined by percentage of income (low, medium, and high). While there are numerous
methods and taxonomies in deﬁning earning levels in the United States, many deﬁne
low-wage jobs as those that exist within the bottom third of national wage distribution.
According to Boushey et a1. (2007), approximately one-third of the US. population falls
into the low-wage category. The Federal Bureau of Labor Statistics commonly ranks
jobs and associated wages by quartiles. As such, very low wage earners account for the
ﬁrst quartile, and low wage earners account for the second quartile.

For purposes of this study, the sample wage distribution was divided into thirds to
account for low wage earners (N=1,025) who earned less than $497 per week, and also
encompasses all individuals who earned less than two-thirds the national median wage of
$12.74/hr (SSA, 2008). All individuals that earned between $498 and $841 per week
were placed into the middle-third category and represents middle income workers. The
highest third, those workers earning over $841 per week, constituted the group of high
wage earners. This process allows the researcher to make assumptions about a particular
population and address the implications regarding the socio-economic status of an
individual or family and their decision to settle a workers compensation claim.

Timeliness of Wage-Loss Benefit. This variable accounts for the time elapsed

between the reported date of injury and the initial wage-loss payment. This was

42

originally an independent, continuous variable that accounted for the time elapsed (in
days) between the reported date of injury and when the worker receives his or her initial
wage-loss payment. For purposes of this study, it was recoded as a categorical variable
to make distinctions among those who received prompt wage-loss beneﬁts following
their reported date of injury (within 30 days) and those who did not. This allows the
researcher to address implications of prompt or delayed receipt of wage-loss beneﬁts
after exiting the occupational setting due to injury or illness, and can be effectively used
to investigate the potential impact this timeframe may having on an individuals decision-
making in workers’ compensation settings.

Geographic Residence. An independent, dichotomous variable distinguishing the
workers’ legal residence and labor market area between rural and metropolitan. For
purposes of this study, the geographic unit used to categorize this variable was the
worker’s county of residence. According to the Ofﬁce of Management and Budget
(OMB) and the Economic Research Service of the United States Department of
Agriculture (USDA), metropolitan areas are deﬁned to include central counties with one
or more cities of at least 50,000 residents, or with an urbanized area of 50,000 or more
and total area population of at least 100,000. This variable was coded at 1 within the
dataset. The OMB and USDA deﬁne rural areas as open country and small settlements of
less than 2,500 persons. This variable was assigned as the constant variable, and was
coded at 0 within the dataset.

Attorney Involvement. An independent, dichotomous variable that designates a
claim as either adversarial (e.g., litigated, disputed, or challenged by the employer), or

non-adversarial (e.g., routine or non-disputed) in nature as of January 1, 2010. This

43

delineation between litigated and non-litigated claims provides and effective method of
estimating attorney involvement in the process of settlement. For example, should a
claim become contested or disputed, the likelihood of attorney involvement is high.
Conversely, should a claim remain uncontested or non-disputed by the employer,
attorney involvement in the decision making and claims processes is minimal or very
low. To make observations regarding the adversarial nature of worker’s compensation
and settlement outcome, the researcher must be able to differentiate between litigated and
non-litigated claims. While these data do not allow for the determination of when a
particular claim has become disputed, the use of various claims status variables and dates
of claim inception, wage-loss payment initiation, and variable coding procedures allows
the researcher to accurately distinguish between litigated and non-litigated claims.

For example, all cases within the dataset were originally designated as CONT,
NCONT, RES, or APP. The variable CONT implies the claim was either historically or
is presently contested. The variable RES implies that a court order disposition regarding
the claim was reached prior to January 1, 2010. The variable APP implies that the claim
is in the appeals process and was historically litigated, therefore contested status is
implied. As such, all of these claims were or had been contested claims as of January 1,
2010 and were coded as 1 within the sample set, while all other values were coded as 0
and were never contested claims.

Settlement. The dichotomous, dependent variable that identiﬁes if a claim has
settled (0 = No and 1 = Yes). For purposes of this study, settlement refers to the

legitimate termination of wage-loss beneﬁts.

44

Procedure

A retrospective research design was utilized for this study by selecting all of the
2007 claims of injury in the State of Michigan. Once all claims were selected and
veriﬁed, all personal identiﬁers were removed from the research dataset as to protect the
identity of each subject. This simple ex-post facto design provides convenient and ethical
manner in which to thoroughly investigate the research question: What impact do various
individual (age, gender, marital status, number of dependents, type of injury, and part(s)
of body affected) and contextual variables (pre-injury average weekly wage, timeliness of
wage-loss beneﬁt, geographic residence, and attorney involvement) have on the decision
to settle a workers’ compensation claim?

The effectiveness of the ex-post facto design shall be reliant on the fact that both
the predictor (independent) and outcome (dependent) variables are all clearly identiﬁed.
Confounding variables were thoroughly sought out and investigated in an attempt to
explain all possible relationships that are hypothesized. The most signiﬁcant limitation
of this design is that the researcher is unable to make cause-and-effect conclusions since
the cause element has already occurred. Due to the large sample size, a casewise
deletion technique was utilized to eliminate cases where one or more variables were
missing, or where cases failed to match the speciﬁc sample criteria. Casewise deletion is
preferred method of deletion when sample size is large in relation to the number'of cases
that have missing data (Allison, 2001). Various statistical methods were utilized to
examine and draw hypothetical conclusions regarding this study as described in the

following section.

45

Data Analysis

The Statistical Package for the Social Sciences (SPSS) version 15.0 was the
statistical software package utilized for the data analysis (SPSS Inc., 2007). This study
provides tables of descriptive statistics, such as frequency distributions of data, means,
standard deviations, ranges, and variances of scores to provide the reader with important
information regarding the sample set utilized in this study.

Since no study could be identiﬁed that has empirically examined the decision to
settle a workers’ compensation claim, a series of exploratory statistical procedures were
carried out in order to identify any confounding or moderating effects between all
variables of interest, as well as provide the reader with a variety of descriptive statistics,
frequency distributions, and cross-tabulations of data.

The statistical method of choice for analyzing the relationships between the
categorical variables containing more than two (2) groups of interest is chi-square
analysis. This non-parametric test of statistical signiﬁcance has been shown to be
appropriate for evaluating dichotomous independent and dependent variables (Huck &
Cormier, 1996), and will provide an effective method for examining whether or not
relationships exist between the decision to settle a workers’ compensation claim and an
individual’s pre-injury average wage, timeliness of wage—loss beneﬁt, type of injury, and
part(s) of body affected.

The chi-square test measures the dependence between two sets of frequency
measures. It is used to analyze a contingency table consisting of ordinal variables,
containing three (3) or more categories, in order to determine if the observed cell

frequencies differ signiﬁcantly from the expected frequencies and represents the sum of

46

the contributions from each of the individual cells. Every cell in a table contributes
something to the overall chi-square statistic. If a given cell differs markedly from the
expected frequency, then the contribution of that cell to the overall chi-square is large. If
a cell is close to the expected frequency for that cell, then the contribution of that cell to
the overall chi-square is low. A large chi-square statistic indicates that somewhere in the
table, the observed frequencies differ markedly from the expected frequencies. This also
implies that two variables are likely associated to each other (Greenwood & Nikulin,
1996).

The remaining nominal variables of interest (age group, gender, marital status,
number of dependents, geographic residence, and attorney involvement) will be evaluated
by calculating a Cramer’s V statistic that allows the researcher to investigate the strength
of dependency or correlation between two or more variables or sets of variables (Cohen
and Cohen, 1983). Correlations have three fundamental dimensions: signiﬁcance,
direction, and magnitude. Cramer's V is a statistic measuring the strength of association
or dependency between two (nominal) categorical variables in a contingency table and
varies between 0 and 1. It is used as post-test after chi-square has determined
signiﬁcance. Although the chi-square statistic may indicate that there is a signiﬁcant
relationship between two variables, it does not allow the researcher to identify how
signiﬁcant or important this relationship could be. The closer the Cramer’s V statistic is
to l, the stronger the association between the variables (Cramer, 1999).

A simple correlation or Pearson’s Correlation analysis will also allow for all
dichotomous variables to be investigated as to the strength and direction of any linear

relationships, should any exist. This method of statistical analysis allows the researcher

47

to check multicolinearity, as well as directional characteristics of the data that could have
an impact on prediction modeling. The Pearson correlation coefﬁcient is a measure of
the correlation between two nominal variables, giving a value between +1 and -1
inclusive. It is widely used to measure of the strength of linear dependence between two
continuous variables. A value of 1 implies that a linear equation describes the relationship
between two variables perfectly, with all data points lying on a line for which one
variable increases as the other increases. A value of -1 implies that all data points lie on a
line for which one variable decreases as the other increases. A value of 0 implies that
there is no linear correlation between the variables (Rodgers & Nicewander, 1988; Fisher,
1915). This method was primarily used to investigate the relationships between the
original continuous variables, such as AWW, number of dependents, and timeliness of
wage-loss beneﬁts, as well as various dichotomous variables on settlement outcome.

The primary statistical method of choice for this study is binary logistic
regression. Logistic regression analysis has been recognized as an effective method of
analysis when the equation to be estimated has a dichotomous response variable
(Tabachnick & Fidell, 2001), such as settlement versus non-settlement of a workers’
compensation claim. It allows investigators to examine relationships between a binary
outcome variable and a group of predictor variables. Logistic regression also allows
researchers to calculate odds ratios, deﬁned as the ratio of the odds of an event for one
group to the odds of the same event for another group (Ott & Longnecker, 2001). This
process allows for the odds calculation between independent and dependent variables
which provide useful information about how two people may differ on one variable,

when they are the same on all other variables (Capella, 2002). The main advantage of

48

using this form of regression analysis, is that it allows the researcher to isolate the impact
of each predictor variable while holding constant all others.

For example, an Odds ratio can allow an investigator to compare the decision to
settle a workers’ compensation claim between males and females when controlling for all
other variables. Odds ratios are typically the parameters of interest when using logistic
regression due to ease of interpretation (Hosmer & Lemeshow, 1989; Capella, 2002) and
once each variable has been descriptively explored, logistic regression will be used to
investigate associations that may exist between the independent variables and the
response variable, as well as potential interactions among the variables of interest

presented as:
Logit (Pi) = BO + 61X” + + BKXKJ + E,
where (Pi) is the log odds of the dependent variable, and B0 is the constant where there

are K independent (X) variables. The notation E represents that random deviation or
random “error” which is assumed to have the mean of zero (Devore and Berk, 2007).
Given the exploratory nature of the study, a level of signiﬁcance of alpha (or) = .05 was
implemented as the minimum rejection level for all statistical analyses and used as the
guideline for assessing the overall ﬁt of the model.

Lastly, a Hosmer-Lemeshow Test will be carried out to assess the overall ﬁt of

the data to the model and estimate the amount of unexplained variability.

49

CHAPTER 4
Results

The purpose of this study was to examine various individual and contextual variables
as predictors for individuals involved in complex or non-routine workers’ compensation
claims within the State of Michigan. This unique sample represents individuals who have
remained out of the workforce, and have had open workers’ compensation claims for a
substantial amount of time. The research question relative to the study asked if various
individual and contextual variables predicted settlement outcome in more complex
occupational disability claims.

Review of variables in the analysis.

Individual Variables - age, gender, marital status, number of dependents, type of
injury, part of body affected. As previously stated, all variables within this category are
categorical in nature. Age, gender, martial status and number of dependents are
dichotomous variables, where as type of injury and part of body affected contained two
or more groups. Age was grouped into two (2) categories identifying those under the age
of 55, and those aged 55 and older. The number of dependents variable was divided into
two (2) groups: less than 2, and 2 or more. The body part of injury variable was divided
into eight (8) groups: head, neck (including throat), trunk, upper extremities, lower
extremities, body systems, multiple body parts, and other or nonclassiﬁable. The nature
of injury variable was divided into three (3) groups: traumatic injuries and disorders,
systemic, dieses or disorders, and other conditions or disorders (see Table 1).

Contextual Variables — pre-injury average weekly wage, timeliness of wage-loss

beneﬁt, geographic residence, attorney involvement. The contextual variables in this

50

study are a combination of dichotomous and non dichotomous categorical variables.
Pre-injury average weekly wage was divided into three (3) groups: those who had
reported earnings at the time of injury of $0-$497, $498-$841, or $842 or more. The
variable timeliness of wage-loss beneﬁt refers to the time between the date of injury and
the initial wage-loss payment in days, and is divided into three (3) groups: 0-30 days, 30-
90 days, and 90 days or more. Geographic residence is a dichotomous variable
differentiating between metropolitan and rural areas as classiﬁed by the OMB and
USDA. The litigated or attorney involvement variable is dichotomous and indicates
whether or not a claim was involved in litigation.

Outcome of variable. The settlement variable is a dichotomous variable
differentiating between those who have settled their claim, and those who have not. This
variable was the primary outcome characteristic utilized in the study.

Review of Methodology.

Various descriptive and logistic regression analyses were utilized as the statistical
means to investigate and assess relationships between both individual and contextual
characteristics and the dependent variable. Chi square analyses were conducted to test
the association of each independent variable on settlement outcome, for model building
considerations, and to identify the goodness-of-ﬁt regarding the ﬁnal model. Binary
logistic regression was the primary method for predicting the decision to settle a workers

compensation claim as depicted by:
Logit (P (Settlement=l)) = 0‘0 + 0‘i (Individual Characteristics)i + 0Lj (Contextual

Characteristics)j + E,

51

Or more speciﬁcally:

Logit (P (Settlement=l)) = 130 + I31 Age + I52 Gender + I33 Marital Status + 54

Number Of Dependents + I35 Type Of Injury + I36 Part Of Body + I37 Pre-Injury

Average Weekly Wage + 58 Timeliness of Wage Loss Beneﬁt + I39 Geographic

Residence + I310 Attorney Involvement + E

Given the exploratory nature of this analysis all predictor variables were included
in the ﬁnal regression model. It is common practice that only variables found to be
statistically signiﬁcant within the sample group are considered to predict outcome and
remain in the ﬁnal model (Cohen, Cohen, West, & Aiken, 2003); however, there are
unique occasions when the outcome variable being studied is not well known and
important covariates, as well as their associations with the outcome variable, are not well
understood (Hosmer & Lemeshow, 2000). As such is the case, no ﬁnal model was ﬁtted
or purported for purposes of this study and future research will serve this purpose more
astutely.

A variety of statistical measures were assessed in order to address the likelihood
of a particular variable predicting settlement. Additionally. the estimation of odds ratios
and expected probabilities were calculated to investigate the presence of interactions and
resulting implications. Odds ratios greater than 1 increased the likelihood a variable
predicting settlement. In cases where odds ratios are less than 1, settlement was less
likely to occur. A probability distribution was calculated in order to estimate the
probability of the settlement occurring among each independent variable and will serve to

ease the interpretation of results as predicted in the model. Finally, a statistical

52

assessment of how well the data ﬁts the regression model and its statistical signiﬁcance
will be carried out.
Presentation of Results
Descriptive Characteristics of the Sample

AS illustrated in Table 2, males comprised 64.9% of the sample (N= 2,008),
where as Females accounted for the remaining 35.1% (N=1,086). Over 74% of the
sample consisted of those under the age of 55 (N=2,300) and 35.7% were age 55 and
older (N= 794). Ages within the sample ranged from 17 to 87, with total of 17 missing
cases. The mean age of the sample was 47 with a standard deviation of 11.378.
Approximately 45% of the sample participants were unmarried (N= 1,404). Regarding
number of dependents, 74.1% of the sample reported having less than 2 dependents (N =
2,306), leaving approximately 25.9% claiming 2 or more dependents at the time of injury
(N= 805).

The nature of injury variable originally contained 8 categories or divisions. Given
the extreme limited number of individuals within the sample reporting infectious and
parasitic disease (.1%), neoplasms and cancer (less than 1%), symptom signs and ill
deﬁned conditions (less than 1%), other conditions and disorders (less than 1%), or
multiple diseases, conditions or disorders (0), all were classiﬁed into one category titled
other or non-classiﬁable injury or illnesses (6.2%, N= 193).

Regarding the primary body part of injury, 39.4% reported injuries existing within
the trunk region (N= 1,225), 19.1% with injuries affecting the upper extremities (N=
595), 19.4% reported injuries to the lower extremities (N= 602), and 16% reported

injuries to multiple body parts (N = 496).

53

Table 2: Frequency and Percentage Distribution of Variables

 

 

Variable N Valid %
Individual Characteristics
Age
Under 55 2,300 74.3
55 and older 794 35.7
Gender
Male 2,008 64.9
Female 1,086 35.1
Marital Status
Single 1,404 45 .2
Married 1,702 54.8
Number of Dependents
Less than two 2,306 74.1
Two or more 805 25.9
Body Part of Injury
Head 73 2.3
Neck, Including Throat 65 2.1
Trunk 1,225 39.4
Upper Extremities 595 19.]
Lower Extremities 602 19.4
Body Systems 1 8 .6
Multiple Body Parts 496 16.0
Other or Nonclassiﬁable 34 1.1
Nature of Injury
Traumatic Injuries and Disorders 2,728 88.]
Systemic Diseases or Disorders 161 5.2
Other Conditions or Disorders 208 6.7
Contextual Characteristics
Average Weekly Wage (AWW)
$0 to $497 1,025 33.0
$498 to $841 1,026 33.0
$842 or more 1,055 34.0
Time Between Date of Injury and Initial Wage Loss Payment (days)
Zero (0) to 30 2,267 73.9
30 to 90 357 1 1.5
90 or more 444 14,5
Geographic Residence
Metropolitan 1,228 39.5
Rural 1,883 60.5
Litigated (Attorney Involvement)
No 1,741 56.0
Yes 1,370 44.0
Settlement
No 1.757 56.5
Yes 1,354 43.5

 

54

The remaining reports of body part injuries consisted of less than 6% of the total sample
encompassed head (2.3%, N= 73), neck (2.1%, N: 65), body systems (.6%, N= 18), and
other or non classiﬁable (1.1%, N: 34). Traumatic injuries and disorders were vastly the
most reported feature of injury, consisting of 88.1% of the sample (N= 2728).

Interestingly, pre-injury average weekly wage was evenly distributed among each
category (low, moderate, and high wage earner) at 33% or 34% respectively. The vast
majority of individuals within the sample received their initial wage loss payment within
30 days of their reported date of injury (N= 2,267, 73.9%). Nearly two-thirds of the
individuals within the sample resided within rural areas (N= 1,883, 60.5%), as compared
to those residing within metropolitan areas (N= 1,228, 39.5%).

Lastly, 44% of the sampled population were involved in litigated claims (N=
1,3 70). The remaining 56% of workers’ compensation claims did not involve any form
of litigation (N= 1,741). Considering all claims contained within this sample (N=
3,111), 43.5% resulted in a settlement agreement between the employee and the
employer. The remaining 56.5%, or 1,757 claims remained paying claims in which the
worker continued to receive wage loss beneﬁts as of January 1‘”, 2010.

Chi Square Tests of Association

Chi square tests of association were carried out to assess the level of signiﬁcance
for all categorical variables on settlement outcome. As previously stated, this process is
used to analyze a contingency table consisting of categorical variables, in order to
determine if the observed cell frequencies differ signiﬁcantly from the expected

frequencies and represents the sum of the contributions from each of the individual cells.

55

Furthermore, a Cramer's V statistic was used to assess the variables to determine the
strength of any signiﬁcant association.

As identiﬁed in Table 3, age (df = 1, N = 3,094) had a chi square of 4.283 and
was found to be statistically signiﬁcant, indicating that there was some level of
association between age and settlement outcome. The Cramer’s V statistic of .037
indicates the strength of this relationship. Gender (df = 1, N = 3,094) had a chi square of
.178 and was found to have a relatively small relationship (Cramer’s V = .008) on
settlement outcome. The relationship between this variable and settlement as the
outcome was not found to be statistically signiﬁcant.

The variable marital status (df =1, N = 3,106) was statistically signiﬁcant (p =
.046) had a chi square of 3.971 indicating there is some measurable level of dependence
between marital status and settlement. The strength of association found was similar to
that of age (Cramer’s V = .036) which was also found to statistically signiﬁcant. The
number of dependents variable (df = 1, N = 3,111) had a chi square of 2.905 which was
not found to statistically signiﬁcant and had no discemable relationship on settlement
outcome. Similar ﬁndings were found regarding geographic residence (df = 1, N =
3,111), where there appears to be no signiﬁcant relationship between those who reside in
rural or metropolitan areas and settlement outcome (chi square = .608, Cramer’s V =
.014).

The most statistically signiﬁcant independent variable was attorney involvement
(df = 2, N = 3,111), that had a chi square statistic of 930.399 implying that a relationship

does in fact exist between attorney involvement and settlement outcome. The strength of

56

Table 3: Correlation Table of Dichotomous Categorical Variables on Settlement Outcome

 

 

Variable N Chi square Crilmer’s (51:21:51)
Age (Under/Over 55) 3094 4.283 .037 .038*
Gender 3094 .178 .008 .673
Marital Status 3106 3.971 .036 .046*
Number of Dependents 31 1 1 2.905 .031 .088
Geographic Residence 31 1 1 608 .014 .436
Attorney Involvement 31 1 1 930.399 .547 .000"

 

*. Correlation is signiﬁcant at the .05 level.
**. Correlation is signiﬁcant at the .01 level.

57

this relationship was the strongest among all independent variables as indicated by a
Cramer’s V statistic of .547.

As shown in Table 4, the body part of injury variable (df =7, N= 3,108) had a chi
square statistic of 25.615 which is signiﬁcant at the a = .05 indicating that at least one
group within this variable had a signiﬁcant difference between observed and expected
outcomes relative to settlement outcome. The strength of this relationship was found to
be quite small (Cramer’s V = .091). This was also true for the pre-injury average weekly
wage (df =2, N= 3,106) which had a chi square of 116.35 (p < .001) and timeliness of
wage-loss beneﬁt (df = 2, N= 3,068) with a chi square of 18.741 (p < .001), indicating
that all of these independent variables were statistically signiﬁcant, and at least one of the
categories within each variable was signiﬁcantly associated with settlement outcome.

The correlation between the timeliness of wage-loss beneﬁt and date of injury and
settlement outcome was .078, although there does appear to be some measurable
inﬂuence or association between pre-injury average weekly wage and settlement outcome
(Cramer’s V = .194).

The nature of injury variable (df = 2, N = 3097) had a chi square of 3.897 and was
found to be statistically signiﬁcant (p = .145) indicating some that there is some level of
association with the dependent variable, although the correlation between nature of injury
and settlement outcome was .035 indicating relatively no relationship exists between the

variables.

58

Table 4: Chi Square Test of Association AmoniVariables With 3 or More Cattﬂiries on Settlement Outcome

 

Cramer’s Chi-

 

Variable N No Exp. Yes Exp. V square
Body Part of Injury 3108 .091 25.615"
Head 43 41.2 30 31.8
Neck, Including Throat 42 36.7 23 28.3
Trunk 660 691.7 565 533
Upper Extremities 367 336 228 259
Lower Extremities 363 339.9 239 282.1
Body Systems 7 10.2 1 1 7.8
Multiple Body Parts 260 280.1 236 215.9
Other or Nonclassiﬁable 13 19.2 21 14.8
Nature of Injury 3097 .035 3.867
Traumatic Injuries and Disorders 1543 1546.8 1 185 1 181.2
Systemic Diseases or Disorders 102 91.3 59 69.7
Other Conditions or Disorders 1 1 1 1 17.9 97 90.1
Average Weekly Wage (AWW) 3106 .194 1 16.35“
$0 to $497 477 578.5 548 446.5
$498 to $841 545 579.1 481 446.9
$842 or more 731 595.4 324 459.6
Time Between Date of Injury and Initial 3068 .078 18.741”
Wage Loss Payment (days)
Zero (0) to 30 1226 1270.2 1041 996.8
30 to 90 204 200 153 157
90 or more 289 248.8 155 195.2

 

*. Asymp. signiﬁcant at the .05 level.
**. Asymp. signiﬁcant at the .01 level.

59

Logistic Regression Model

As shown in Table 5, a complete binary logistic regression model was developed
including all individual and contextual inﬂuences on settlement outcome (N= 3,111).

The ﬁnal model is as follows:

Logit (P (Settlement=l)) = -.616 + .094 (Age) - .394 (Female) + .086 (Married)

+ .062 (Number Of Dependents > 2) - .076 ( Nature of Injury - Systemic Illness) -

.040 (Nature of Injury - Other) - .342 (Head) - .485 (Neck) + .103 (Trunk) - .032

(Lower Extremities) + .574 (Body Systems) + .198 (Multiple Body Parts) — .590

(Other) - .272 (AWW > $498, but < $842) - .890 (AWW > $842) + .032

(Timeliness of Wage-Loss Payment > 30 days, but < 90 days) - .369 (Timeliness

of Wage-Loss Payment > 90 days) - .066 (Metropolitan) + 2.584 (Attorney

Involvement = Yes) + E
A total of four (4) variables (gender, pre-injury average weekly wage, timeliness of wage
loss beneﬁt, and attorney involvement) were identiﬁed as signiﬁcant at or = .05.

Age was not predictive Of settlement outcome and had a negligible effect on an
individuals decision to settle his or her workers compensation claim ([3 = .094, p = .391).
According to the model, injured workers under the age of 55 at the time of injury were
1.099 times likely to settle when compared to those over the age of 55, and each group
had roughly a 1 in 3 chance in settling when controlling for all other variables in the
model ( prob. (under 55) = .372, prob. (55 and over) = .351).

Females, as compared to males, were 1.414 (1 / Exp (13)) times less likely to settle

a workers’ compensation claim when controlling for all other variables within the model

60

Table 5: Complete Model for Binary Logistic Regression Analysis: Individual and Contextual Inﬂuences

on Settlement Outcome (N=31 1 1)

 

 

. Odds Conf. Interval
Variable B SE p—value Ratio L U
Age (compared with age 55 and older)
Under 55 .094 .1 10 .391 1.099 .886 1.363
Gender (compared with male)
Female -.346 .100 .001 ** .707 .581 .861
Marital Status (compared with single)
Married .086 .098 .380 1.090 .900 1.320
Dependents (Compare with less than 2
dependents)
Two (2) or more .062 .1 12 .580 1.034 .854 1.326
Body Part of Injury (Compared with
upper extremity)
Head -.342 .315 .277 .710 .383 1.316
Neck, Including Throat -.485 .333 .144 .615 .321 1.181
Trunk .103 .128 .424 1.108 .861 1.425
Lower Extremities -.032 .149 .831 .969 .724 1.296
Body Systems .574 .626 .360 1.775 .520 6.057
Multiple Body Parts .198 .155 .201 1.220 .900 1.653
Other or Nonclassiﬁable -.590 .724 .416 .555 .134 2.293
Nature of Injury (Compared with
traumatic injuries)
Systemic Diseases or Disorders -.076 .218 .727 .927 .605 1.420
Other Diseases or Disorders -.040 .190 .835 .961 .662 1.395
Average Weekly Wage (Compared with
group $0 to $497)
$498 to $841 -.272 .112 .015“ .762 .612 .948
$842 and above -.890 .120 <.001** .41 1 .324 .520
Time Between D01 and Initial Wage
Loss Payment
(Compared with 0 to 30 days)
30 to 90 days .032 .145 .826 1.032 .777 1.371
90 days or more -.369 .136 .007" .692 .530 .902
Geographic Residence (Compared with
rural)
Metropolitan Area -.066 .093 .474 .936 .781 1.122
Litigated (Compared with no)
Yes 2.584 .093 <.001** 13.255 11.056 15.892

 

Note: D0] = Date of Injury.
*p<.05.**p<.01
L = Lower CI; U = Upper CI at 95%

61

(B = -.346, p < .001). As depicted in Table 6, the estimated probability of females to
settle is .276 or roughly one chance in four (4) when controlling for all other variables.

An individual’s martial status was not statistically signiﬁcant in predicting
settlement in the model (p = .380). When controlling for all other variables, the model
indicates that individuals who are married do, however, have a slightly larger probability
of settlement (B = .086), than those who are single at the time of injury. Similarly, there
was no statistical signiﬁcance detected in the model relative to the number of dependents
an individual reported at the time of their reported injury (B = .062, p = .580). The
coefﬁcient is positive indicating that those persons with 2 or more dependents have a
larger probability of settling their claim when controlling for all other variables. The
calculated probability is .365 which is slightly greater than a 1 in 3 chance of settling
when all remaining variables remain constant, whereas those persons with less than 2
dependent had a .351 probability of reaching a settlement agreement within the sample
timeframe of three years post-injury.

The body part of injury variable, as compared with upper extremities, contained 7
distinct injury categories, none of which yielded signiﬁcant inﬂuence within the model at
or below or = .05. When controlling for all variables in the model, those workers who
reported sustaining body systems injuries and illnesses (B = .062, p = .580) were 1.775
times more likely to settle their claim as compared to those with upper extremity injuries.

Individuals reporting head, neck, lower extremity, or nonclassiﬁable/other injuries and

62

Table 6: Binominal Probability Distribution Table of Predictor Variable on Settlement

 

 

Variable B P (X)
Intercept -0.616 0.3 50691 73
Age Under 55 0.094 0.37238469
Gender Female -0.346 0.27647794
Married Status Married 0.086 0.37051689
Dependents 2 or more 0.062 0.36493688
Body Part oflnjury Head ~0.342 0.27727881
Neck -0.485 0.24955257
Trunk 0.103 0.3 7449052
Lower Extremities -0.032 0.34344037
Body Systems 0.574 0.48950154
Multiple Body Parts 0.198 0.39699543
Other -O.590 0.23040957
Nature of Injury Systemic Diseases -0.076 0.3335883]
Other Diseases -0.040 0.34163873
Pre-Injury Weekly Wage $498-$841 -0.272 0.29152273
$842 and above -0.890 0.18153235
Time Of wage L055 3090 days 0.032 0.3580127]
Beneﬁt
90 days or more -0.369 0.27190080
Geographic Residence Metropolitan Area -0.066 0.33581507
Attorney Involvement Yes 2.584 0.87739613

 

63

illnesses had a lesser probability of settlement than those who experienced trunk, body
systems, and multiple body systems injuries or illness when controlling for all remaining
variables. Individuals who reported injuries that could not be classiﬁed or categorized
were nearly 1.694 times less likely to settle 3 workers’ compensation claim when
controlling for all other variables within the model (B = -.590, p = .555).

Regarding the nature or type of injury variable, no signiﬁcant relationship was
identiﬁed relative to settlement outcome. Compared with traumatic injuries and illnesses,
both systemic (B = -.076) and nonclassiﬁable injuries/illnesses (B = -.040) yielded a lesser
probability estimate indicating that individuals who report these types of injuries have a
lower probability of settlement than those who reported breaks, strains, or sprains
(traumatic injuries). The probability among each of category of nature of injury on
settlement falls between .333 and .351 indicating an individual has a 1 in 3 chance of
settlement regardless of type of injury when all variables in the model remain constant.

Pre-injury average weekly wage was found to be signiﬁcantly related to the
probability of settlement in the model. Compared with individuals with earnings of less
than $497 per week, those earning between $498 to $841 (B = -.272, p = .015), as well as
those earning above $842 per week (B = -.890, p < .001), indicate negative associations
on settlement outcome when controlling for all other variables in the model. This would
indicate that those who earn higher pre-injury wages have a lower probability of
settlement than low wage earners (less than $498/week). The estimated probability of
settlement of those who earned greater than $842 per week was .182, indicating that high
wage earners are 2.433 times less likely to settle than those who earn $497 or less per

week (Odds Ratio = .411) when controlling for all other variables within the model.

Those who earn between $498 and $841 per week were 1.312 less likely to less likely to
settle than those who earn $497 or less per week (Odds Ratio = .762) when controlling
for all other variables within the model.

Individuals who did not receive initial wage loss payments after their reported
date of injury until 90 days or more (B = -.369, p = .007) were less likely to settle when
controlling for all other variables. This was statistically signiﬁcant at the .05 level. As
such, it appears that the longer the duration of time between the reported date of injury
and initial payment of wage-loss payment, the lower the probability of settlement. The
estimated probability of settlement regarding those who did not receive wage-loss beneﬁt
until after 90 days, or more following their reported date of injury is .272, or roughly 1
chance in 4 when controlling for all other variables. This is compared to the those who
did receive wage-loss beneﬁts within 90 days (or less) following their date of injury,
where settlement is estimated to occur 35% of the time (prob. = .350 to 358) or roughly 1
out of 3 times when controlling for all other variables.

Geographic residence was not signiﬁcantly related to probability of settlement in
the regression model (B = -.066, p =.474). The negative coefﬁcient implies that although
small, there is possibly a negative association between those who reside in metropolitan
areas and settlement outcome. This would indicate that those who reside in rural areas
are more likely to settle.

The attorney involvement variable (B = 2.584, p < .001) was statistically
signiﬁcant in predicting settlement when controlling for all other variables. This variable

accounted for the most substantial amount of variance within the regression model,

65

whereas the odds ratio predicts that individuals with attorney representation are 13.2
times more likely to settle than those without attorney representation.

The overall ﬁt of the logistic regression model is not statistically signiﬁcant and
does not ﬁt the data well (chi square = 7.62, p = .447). The R square value of .409
indicates that approximately 41% of the variability within the model has been accounted
for in predicting settlement. Nearly 60% of the variability remains unexplained. The

overall percentage of successfully predicted outcome values in the model was 78.8%.

66

CHAPTER 5
Discussion

It has been argued that if an individual is unable to return to his or her previous
job as a result of an occupational injury or illness, settlement is the primary goal in any
workers compensation system. It would then be valuable to identify what variables
signiﬁcantly inﬂuence an individual’s decision-making within this complex system. The
purpose of this study was to investigate what variables may have an impact on an
individual’s decision to settle their workers compensation claim where a permanent or
persistent disability is perceived to exist. Four (4) of the predictor variables identiﬁed by
this researcher have possible factors in inﬂuencing an occupationally injured workers
decision to settle his or her claim, and were identiﬁed as signiﬁcantly contributing to the
logistic regression model. These predictor variables were gender, average weekly pre-
injury wage, timeliness of wage-loss beneﬁt (time elapsed between date of injury and
initial wage loss payment), and attorney involvement. The model was successful
predictor of settlement 78% of the time although could only account for approximately
41% of the variance in claims resulting in settlement. Given the amount of unexplained
variance and lack of ﬁt of the model to the data (chi square = 7.62, p = .447) caution
should be used when interpreting these results.

Given the large number of participants involved in workers’ compensation
systems, and the numerous organizations impacted by workers compensation costs, it was
surprising to ﬁnd very few empirical studies that investigated why an injured worker
would choose to settle his or her claim, once a routine return to work seemed unlikley. It

appears that the most common dependent variable used in workers’ compensation

67

outcome research is return to work. While this outcome variable is commonly utilized
among medical and social science researchers, return to work lacks a consensus of
deﬁnition. These studies do however, allow us to compare and interpret ﬁndings, as well
as better understand the limitations of such research. Given the limited amount of
empirical inquiry into settlement probability, the discussion of these results are somewhat
limited and further research is warranted.
Signiﬁcant Predictors of Settlement in the Model

The most positive and signiﬁcant predictor of claims resulting in settlement was
the attorney involvement variable. As previously discussed, this variable implies that the
claim, at some point, was disputed resulting in attorney involvement and litigation. The
calculated odds ratio indicates that individuals who are represented by attorneys are
essentially thirteen (13) times more likely to settle their workers’ compensation claim
when accounting for all other variables in the model. In fact, of all non-litigated claims
(N= 1,741) contained within the sample, 80.5% (N= 1,402) did not settle and continued
to receive workers’ compensation beneﬁts, whereas 19.5% (N= 339) had settled their
claim. Conversely, of the 1,370 litigated claims within the sample, 74.1% resulted in
settlement (N= 1,015) and approximately 25% (N= 355) continued to receive workers’
compensation beneﬁts up to three (3) years later. These ﬁndings are consistent with
Thomason & Burton’s (1993) ﬁndings where those claims that had attorneys involvement
were more likely to settle.

Given that this research sample involves more problematic workers’
compensation claims, or claims in which individuals were absent from the workplace for

a substantial amount of time, it is not surprising that attorneys specializing in these

68

matters would be involved. There are attorneys who represent injured workers and
specialize in advising injured parties about their legal rights and entitlements, the process
of disability determination, and seek to obtain maximum awards for their clients. These
claims traditionally involve individuals who are unable to return to their previous work or
usual occupation a result of a disability or functional limitation. There are also a
substantial portion workers who have been absent from the workplace for a signiﬁcant
period of time due to a reported occupational injury or illness require additional medical
care.

There are also attorneys who are retained by employers who specialize in
defending claims of injury brought forth by employees. These attorneys specialize in
various aspects of employment law, workplace injury and liability issues, and assist in the
defense of workers’ compensation matters. Their overall goal is to reduce risk and
exposure for employers, limit the liability of new and existing claims, and consult with
employers and insurance companies on settlement strategies. This requires navigating a
complex legal system specializing in disability determination and compensation, and as
attorneys become involved in managing or litigating disputed claims, injured workers
appear more likely settle their claim. Therefore, under these more adversarial
circumstances settlement becomes a viable option for all parties, although it contingent
upon a variety of factors such as, successful negotiation between the employer and
employee, and ultimately a magistrates approval. Consequently, there is a great deal of
speculation about what occupationally injured worker will do after they settle their claim.

Another positive predictor variable was pre-injury average weekly wage which

was utilized to compare between high wage earners and low wage earners on settlement

69

outcome. While previous studies have shown that the level of pre-injury compensation
is not signiﬁcantly associated with return to work outcomes (Gallahger, 1989), it appears
to have a measurable impact on the decision to settle a workers’ compensation claim,
especially in cases where the worker has been absent from the employment setting for a
substantial amount of time. It was not surprising to ﬁnd that high wage earners, or those
who earned in excess of $842 per week were much less likely to settle their workers
compensation claim when compared to those who make less. In fact, individuals who
earned less than $497 per week were more than two (2) times likely to settle their
workers compensation claim as compared to higher wage earners when controlling for all
other variables.

It can be hypothesized that individuals who have higher earnings have more
monetary savings which can be accessed to meet ﬁnancial obligations when not working.
Higher wage earners may have more capital assets that can be liquidized or sold to
generate needed income when unable to generate income through employment. This
would imply that the more money one makes, the more likely they would be able to
sustain ﬁnancial obligations and hold out for the most favorable settlement. It is also
supportive of previous research that identiﬁes an inverse relationship between settlement
probability and settlement expectations (Thomason, 1989), implying that high-wage
earners are likely to seek higher settlements due to the larger amount of lost wages.
Additionally, previous research models aimed at predicting settlement probability have
identiﬁed both pre-injury weekly wage and time away from work due to injury as

positively inﬂuencing the amount of settlement (Thomason & Burton, 1993).

70

Interestingly, these ﬁndings are somewhat contradictory to the theory that high
wage earners would have a more difﬁcult time sustaining their lifestyle and meeting
ﬁnancial obligations if not working, especially in disability insurance systems where
wage-loss beneﬁts are capped in accordance with workers’ compensation laws. For

example, as of June 30, 2007, the maximum weekly beneﬁt that any one person could

receive was $739.00 regardless of pre-injury earnings (§ 418.355). There is also the
assumption that low-wage earners are more readily capable of making economic and
lifestyle adjustments based on weekly wage-loss beneﬁts given that the difference
between their pre-injury income and wage-loss beneﬁt is much less than that of high
wage earners.

The timeliness of wage-loss beneﬁt variable compared three (3) groups of
workers who experienced varying durations of time between their reported date of injury
and receipt of their ﬁrst wage-loss payment. The ﬁndings indicate that the longer the
duration between the date of injury and initial wage loss payment, the less likely an
individual is to settle his or her workers compensation claim. The variable that represents
the category of workers who did not receive workers compensation beneﬁts until 90 days
or more were only half as likely to settle their workers compensation claim than those
who received wage-loss beneﬁts within 3 months. In other words, these individuals have
a 1 in 4 chance of settlement when controlling for all other variables.

One explanation for this ﬁnding is that individuals who do not receive prompt
wage replacement beneﬁts following a workplace injury may tend to become disgruntled,
alienated, and may feel that their integrity has been called into question. Previous studies

support this ﬁnding whereas injured worker’s who are unable to return to retum to work

71

 

in a timely fashion, may be forced to borrow money, deplete savings, and experience
economic hardship (Morse et al., 1998; Keogh et al., 2000; Pransky et al., 2000). As
such, these claims may tend to become more adversarial in nature resulting in longer
claim duration. This long period until payment may also represent evidence of an
employer or insurance companies unwillingness to acknowledge the legitimacy of a
claim. In any event, the longer the duration of time between an individual’s report of
injury and the date of wage-loss payment, the more likely the claim is to be disputed.

The last signiﬁcant predictor identiﬁed within the regression model was gender.
When accounting for all other variables, females were less likely to settle their workers
compensation claim than males, and had an estimated probability of 1 in 4 in resulting in
settlement when controlling for all other variables. The distribution among females
versus males within sample set was approximately 2:3, which is interesting given that this
is less than the female to male representation in the labor force which typically rests
between 43 and 46 percent or closer to 1:1 (BLS, 2009). As indicated by these ﬁndings it
appears that women are not only less likely to have these troublesome claims, they also
appear more reluctant to settle them when involved.

The most signiﬁcant implication regarding gender is that female presence in this
sample is not proportionate to their participation in the labor force. This may imply that
females are less likely to report injuries that occur at work. This ﬁnding may also be
inﬂuenced by the nature of work that females traditionally perform, which is likely to be
more administrative, ofﬁce-oriented, or customer serviced based, rather than heavy
manual labor. There is also the consideration of risk behavior, whereas males have been

traditionally more likely to take physical risks than females. Given that there are two (2)

times as many males ﬁling workers’ compensation claims than females, and no empirical
studies could be identiﬁed that could be speciﬁcally investigated as to why females
would be less likely to settle their workers compensation claim versus males, this variable
alone in predicting settlement outcome warrants ﬁrrther investigation.

Non-Signiﬁcant Variables in the Model

The remaining variables were not found to be consistently signiﬁcant within the
regression modeling. For example, there was no statistical difference between those
persons under the age of 55 and those 55 years of age and older and their decision to
settle their workers’ compensation claim. While previous studies have shown positive
relationships to exist between older workers’ who sustain an work injury and time away
from work (Cheadle, 1994), older workers do not appear any more likely to settle their
claim than younger workers given the parameters of this study. Future studies may serve
to address additional categorical variables amongst age (e.g., 15-25, 26-49, 50-62, and 63
and over) to account for various changes in economic and social statuses among
workforce participants.

While previous studies have shown that single or divorced workers to have
longer periods of occupational disability and time away from work (Bolton, 1974; Volinn
et al., 1991; Cheadle et al., 1994), this study found that one’s marital status had no impact
on a person’s decision on a workers compensation claim when controlling for all other
variables, and the odds ratio was very close to one implying that those who were married
had essentially the same likelihood of settling than those who were single.

Neither body part of injury or nature of injury yielded any signiﬁcant difference

between those who chose to settle their claim and those who did not when accounting for

73

all other variables. It is important to note, however, that the vast majority of injuries to
body parts consisted of upper extremity, lower extremity and trtmk injures, while the
overwhelming majority of injuries (more than 80%) resulted from traumatic illnesses or
injuries, sprains, strains and breaks as compared to all other injury types.

Lastly, although nearly two-thirds of sample participants resided within rural
communities, there was no signiﬁcant difference found between those who resided in
metropolitan areas and those within rural communities and their decision to settle. While
previous studies have shown that injured workers who resided in rural areas were less
likely to return to work (Cheadle et al., 1994), this sample of workers, as a whole, were
all less likely to return to work given the nature of their claim. These ﬁndings are
inﬂuenced by how a particular researcher deﬁnes rural versus metropolitan; however, if
one were to assume that workers from rural areas are less apt to return to work, they may
be less likely to settle their claim. Further research into the potential impact that
geographic location and economic conditions may have on individual decision making in
workers’ compensation is also warranted given our limited understanding.

Limitations of the Study

The primary limitation of this study is that the ﬁndings are heavily reliant upon
retrospective data and the researcher is unable to make cause-and-effect conclusions
since the “cause” element has already occurred.

The data used to construct the sample is also heavily reliant upon self-reported
information. The validity of these data is also contingent upon the accuracy in which the

self-reported data was entered and maintained. The research data and presentation of

74

ﬁndings, assumptions, and purported implications cannot account for claims that were
mistakenly deleted or falsely classiﬁed.

It is also possible that a portion of these data may be missing key demographic or
injury related variables and would eliminate population members and or variables from
the analysis. This should be carefully considered with generalizing about these results
among different years, whereas changes in legislative practices, public policies, labor
market conditions, medical technology, and the nature of work may have taken place.

While the chi-square analysis is a useful statistical technique for evaluating
categorical variables, it may not fully account or control for other relevant variables that
are inﬂuencing the outcome (Glass & Hopkins, 1996). This limitation is somewhat
mitigated by the large sample size, however, should be carefully considered when
interpreting this research. Although the model was successful predictor of settlement
78% of the time it could only account for approximately 41% of the variance in claims
resulting in settlement (chi square = 7.62, p = .447) implying that additional predictor
variables on settlement outcome are explained. Examples of additional variables of
interest that may have an impact on predicting settlement outcome not present in this
study include participation in vocational rehabilitation, settlement amount, date of
settlement versus date of injury, union membership, severity of injury or illness, worker
occupation, industry classiﬁcation, insurer actions (e.g., beneﬁt rate changes, medical
examinations, forensic evaluations, etc.), and legal proceedings variables, such as the
occurrence of mediations or trials and their potential impact on injured worker decision

making.

75

Given many of the previously stated considerations, the statistical conﬁdence
interval chosen for the study (a = .05) may inﬂuence the probability of a Type I error, or
in other words, allowing the researcher to incorrectly reject the null-hypothesis when
comparing or analyzing individual variables. This does not appear to be a substantial
concern, given that all of the signiﬁcant predicator variables in the model were signiﬁcant
at the .01 level, however there were individual independent variables that were found to
be signiﬁcantly correlated with settlement outcome that remained in the model. While
the degree of these correlations were small overall, there appears to be some confounding
effects that could not be accounted for in this study. For example, the variable gender
was not found to signiﬁcantly associated with the settlement outcome (chi square = .178);
however, it was found to be signiﬁcant predictor in the regression model. This would
imply that gender’s effect within the model is being inﬂuenced by another variable in the
model, or a characteristic inﬂuencing settlement outcome that exists within the
unexplained variance.

Unfortunately, any study conducted within workers’ compensation setting lends
itself to generalizability concerns since each State jurisdiction has different rules and
regulations regarding disability determination and beneﬁt provisions. Additionally, this
study focuses on more problematic or complex disability claims in workers’
compensation, and as such would not be appropriate to make generalizations about these
ﬁndings across the vast majority of claims that resolve under normal or normally
anticipated circumstances. As such, it would not be appropriate to assume this sample

would be representative of any or all other injured worker populations without additional

76

research into the administrative rules and procedures that govern a particular jurisdiction,
and the speciﬁc circumstances and nature of a particular claim.

Given the purpose of the study, a variety of continuous variables were recoded in
order to address theoretical questions regarding individual decision-making. As such,
there was a change in the distribution of values within each of these variables. This
change will affect the measures of association that are used in methods of statistical
analysis among categorical variables. The recoding procedure, while useful to produce
meaningful interpretations of variables speciﬁc to a particular industry, environment, or
population, can produce unpredictable effects that need to be interpreted carefully when
making assumptions about the signiﬁcance of any particular variable. This too could
potentially inﬂuence the occurrence of a Type I or Type 11 error as previously discussed
above.

As stated in the results section, there were a variety of statistically signiﬁcant
correlations identiﬁed between three (3) predictor variables and settlement outcome,
which could have affected the outcome model. These multicollinearity characteristics
could have an impact on signiﬁcant ﬁndings as identiﬁed within the logistic regression
model. Two of the variables had very weak, negative correlations and do not appear to
have much association with the outcome variable. Attorney involvement, however, had
a moderately positive correlation on settlement outcome and warrants caution when
interpreting this variable in the model.

A signiﬁcant limitation regarding the timeliness of wage-loss beneﬁt variable
worth noting. These data do not allow the researcher to account for a workers’ last day

of work, and as a result, it is possible that an individual may have continued to work

77

subsequent their date of injury. For example, it is possible that a person may have chosen
to continue to work and not report an injury until many days, weeks, or months
subsequent the injury event. Therefore, some cases may not accurately represent the
person being away from work between the reported date of injury and when wage-loss
beneﬁts initiate.

Implications for Rehabilitation Counselor Education

Severe workplace injuries traditionally result in a medical recovery periods that
require either a short-term or long-term period away from work (Tate, 1992). It is this
within this undetermined period of time in which the injured worker will begin to
psychosocially adapt to their occupational injury, illness or disability. While most
workers substantially recover from their injuries and return to work, others experience a
number of long-term consequences (Pransky et al., 2001) that may inﬂuence their
decision-making and overall ability to psychosocially function.

The availability of conceptual models such as Livneh’s Model of Psychosocial
Adaptation to Chronic Illness and Disability (2001) or Dembe’s Model of the Social
Context of Occupational Injury or Illness (2001) provide useful information regarding the
contextual inﬂuences and reactionary responses of those who have sustained a work
disability and are receiving workers’ compensation beneﬁts. More speciﬁcally, it is the
application of these models that allows educators and students to better understand the
complex nature of workers’ compensation and participants. The extent at which injured
workers are able to successfully adjust and adapt to their occupational disability appears
to be inﬂuenced by their ability to manage varying levels of stress, anxiety, and

depression as experienced within the workers’ compensation setting, and may

78

additionally be affected by the size of employer, relative labor market conditions, marital
status, and duration of unemployment.

Continuing research in the area of occupational injury and return to work and
settlement outcomes can assist rehabilitation counseling educators prepare counselors for
entry into private sector rehabilitation, a service setting that demands specialized skills in
rehabilitation consulting, disability management, and vocational evaluation. This
research provides introductory support that there are signiﬁcant factors affecting an
individual’s decision to settle his or her workers compensation claim. These ﬁndings, in
combination with previous rehabilitation outcome research, provide evidence that there
are measurable characteristics that inﬂuence individual decision making. Licensed and
certiﬁed professionals who provide medical and vocational rehabilitation services to
injured workers in workers’ compensation settings are often faced with challenging
circumstances. There remains a need for competent and highly trained vocational
rehabilitation counselors who specialize assisting occupational injured persons return to
work. There is a similar need of improving upon the education of rehabilitation
counselors who perform forensic evaluations in workers’ compensation, and the
associated methods used to assess an individual’s rehabilitation potential, employability,
and wage earning capacity.

For example, studies such as this can also be incorporated into curriculum design
for graduate-level education in private sector rehabilitation, and may be used in part for
class discussion, presentation, or case studies regarding vocational outcomes in complex
disability claims. It can additionally provide rehabilitation educators with empirical

ﬁndings that can be compared among similar studies. It may also serve to supplement a

79

continuing body of research that may be used toward developing evidence-based
practices in rehabilitation counseling in workers’ compensation settings.
Implications for Future Research

This particular study served to address particular variables associated with the
decision making process of an injured worker. We presently have access to a variety of
economic-based studies that have investigated cost containment and settlement outcome
strategies; medical research focused on duration of disability and related occupational
disability outcomes; and vocational rehabilitation research aimed at investigating
employment outcomes of those with disabilities. Very few studies have served to address
industry-speciﬁc or related occupational interests relative to predicting the costs
associated with workers compensation claims or what inﬂuences decision making in
workers’ compensation systems. While there has been a variety of qualitative studies that
have examined the experiences of injured workers involved in workers compensation,
further inquiry into the experiences and decision making process of the injured worker is
needed.

While all of these studies serve to address very important issues within existing
workers’ compensation systems, this researcher was only able to identify two studies that
empirically investigated why a worker would choose to settle his or her claim. Given that
workers’ compensation is an inherently adversarial process, involving both collective and
competing interests among employers and employees, this question continues to warrant
further investigation. Replication of similar research methodologies and samples across
different State jurisdictions and timefrarnes would be invaluable in determining

signiﬁcant predictors of settlement. Results of, these studies would be beneﬁcial for

80

insurance carriers and self-insured employers whose goal is to provide safe work
environments, yet must control costs associated with lost productivity, insurance
premiums, and indemnities that result from workplace accidents.

As previously stated, it has been argued that if an individual is unable to return to
his or her previous job, settlement is the primary goal in any workers compensation
system. It would then be valuable to identify what variables Signiﬁcantly inﬂuence an
individual’s decision-making within this complex system. While it is impossible to
account for all of the variance associated with human behavior and the decision making
process, future studies would serve to validate or challenge these ﬁndings. Additional
research in this area could help to develop best practices within workers’ compensation
that would have profound impact on vocational rehabilitation, claims management, and
help to ensure fair and just treatment for injured workers. Future research across those
States that have similar workers’ compensation systems would also offer additional
ﬁndings that would be useful in policy development and program evaluation.

Qualitative studies investigating why individuals have chosen to settle a workers’
compensation claim would also contribute in the evaluation of workers’ compensation
systems from a post-settlement perspective, and serve to investigate what injured workers
do once they exit the workers’ compensation system. Qualitative follow-up studies could
also be used to update historical studies regarding individual’s perceptions of the
workers’ compensation system and help to further understand what individual’s do after
they settle their claim.

Used in conjunction with replicated research, these data can serve to provide the

foundation for meta-analytic studies that can address positive and negative trends among

81

workers’ compensation systems and settlement outcomes. Applied in conjunction with
vocational rehabilitation outcome studies, researchers can better understand both the
incentives and disincentives associated with workers’ compensation and return to work,
and its interaction with the likelihood of successful job placement. For example, what
impact does an individual’s decision to return-to-work have on settlement outcome, or if
an injured worker wishes to settle their claim, what impact does this decision have on
vocational rehabilitation outcome?

This research can serve to provide valuable information regarding the perceived
disincentives that many social insurance and no-fault systems may have on an
individual’s decision to settle their claim or return to work. Further research in this area
can improve our understanding of competing interests and secondary gain issues in
disability determination environments. Utilized in conjunction with new and existing
medical research, these studies may be used investigate how certain types of injuries
impact an individual’s employability and wage earning capacity, duration of disability,
and rehabilitation outcomes.

While this study begins another inquiry into what factors inﬂuence injured
worker’s to settle their workers’ compensation claim, it is also important to improve our
understanding of the relationship between settlement outcome and long-term outcomes of
the injured worker. Even under voluntary, court-approved settlement agreements, injured
workers may remain unsure and unoptimistic about their future. This warrants future
research that can compare the initial perceptions, interests, and goals of the injured
worker at the time of settlement and long-term outcomes, such as quality of life, physical

and mental health, and work-status many years post-settlement.

Settlement outcome in workers’ compensation remains an interesting topic of
discussion among many professionals, policy makers, researchers, and educators.
Overall, outcome studies regarding settlement serve to improve our understanding of the
occupationally injured individual, including their experiences and perceptions of the
workers compensation system as a whole. Hopefully, this study will help spark renewed
interests in assessing present practices in workers’ compensation, and contribute toward
improving the quality of life of individuals suffering from occupational illness or injury.
Most importantly, further research in this area can help provide information that is useful
in administering various rehabilitation services to injured workers to help improve post-

injury outcomes.

83

References
Allison, P. D. (2001). Missing data. Thousand Oaks, CA: Sage Publications.

Ash, P., & Goldstein, S. I. (1995). Predictors of returning to work. Bulletin of the
American Academy of Psychiatry and the Law, 23, 205-210.

Beck, R. J. (1988). A survey of injured worker outcomes in Wisconsin. Journal of
Applied Rehabilitation Counseling, 20(1), 20-24.

Berkowitz, E. D. (1999). Supporting disability: An historical perspective. American
Rehabilitation, 25, 2-8.

Bickenbach, J. E. (1993). Physical disability and social policy. Toronto: University of
Toronto.

Bureau of Labor Statistics. (2008). Civilian labor force and participation rates with
projections: 1980 to .2016, Monthly Labor Review. [Table 575]. Retrieved
http://www.bls.gov/opub/ee/home.htm.

Bureau of Labor Statistics. (2009). Glossary of terms. Retrieved:
http://www.blsgov/bls/glossaryhtrn.

 

Bureau of Labor Statistics (2007). Occupational Injury and Illness Classiﬁcation System
Manual. Retrieved: http://www.bls.gov/iif/oshoiics.htm.

Biddle, J. & Roberts, K. (2003). Claiming behavior in workers’ compensation. The
Journal of Risk and Insurance, 70(4), 759-780.

Biddle, J, Roberts, K., Rosenman, K. D. & Welch, E. M., (1998). What percentage of

workers with work-related illnesses receive workers' compensation beneﬁts?
Journal of Occupational & Environmental Medicine, 40, 325-331.

Blackwell T. L., Leierer, S. L., Haupt, S. & Kampitsis, A. (2003). Predictiors of
vocational rehabilitation return-to-work outcomes in workers’ compensation.
Rehabilitation Counseling Bulletin, 46(2), 108-1 14.

Boden, L. I. & Galizzi, M. (1999). Economic consequences of workplace injuries and
illnesses: lost earnings and beneﬁt adequacy. American Journal of Industrial
Medicine, 36(5), 487-503.

Bolton B. (1974). Introduction to Rehabilitation Research. Springﬁeld, IL: Thomas.

Boushey, H., Fremstad, S., Gragg, R., & Wallar, M. (2007). Understanding low-wage

work in the United States. Washington, DC: Center for Economic Policy and
Research

84

Brandt, E. N. & Pope, A. M. (Eds.) (1997). Enabling America: Assessing the Role of
Rehabilitation Science and Engineering. Washington, DC: National Academy
Press.

Brink, N. E. (1989). The power struggle of workers’ compensation: Strategies for
intervention. Journal of Applied Rehabilitation Counseling, 20(1), 25-28.

Burns, J. W., Sherman, M. L., Devine, J., Mahoney, N., & Pawl, R. (1995). Association
between workers’ compensation and outcome following multidisciplinary
treatment for chronic pain: Roles of mediators and moderators. Pain, 11, 94—102.

Butler, R. J. & Worrall, J. D. (1991). Claims Reporting and Risk Bearing Moral Hazard
in Workers' Compensation, The Journal of Risk and Insurance, 58, 196-197.

Capella, M. E. (2002). Inequites in the VR system: Do they still exist? Rehabilitation
Counseling Bulletin, 45, 143-153.

Carey, T. 8., Garrett, J. M. & Jackman, A. M. (2000). Beyond the good prognosis. Spine,
25, 115—120.

Cheadle, A., Franklin, G., Wolﬂragen, C., Savarino, J., Liu, P. Y., Salley, C., & Weaver,
M. (1994). Factors inﬂuencing the duration of work-related disability: A

population based study of Washington State workers’ compensation. American
Journal of Public Health, 84, 190-196.

Chibnall, J. T., Tait, R. C. & Merys, S. C. (2000). Disability management of low back
injuries by employer-retained physicians: ratings and costs, American Journal of
Industrial Medicine, 38, 529—538.

Chibnall, J. T., Tait, R.C., Andresen, E. M. & Hadler, N. M. (2005). Race and
socioeconomic differences in post-settlement outcomes for African American and
Caucasian Workers' Compensation claimants with low back injuries. Pain,
114(3), 462-472.

Cohen, J. & Cohen, P. (1983). Multiple Regression/Correlation for the Behavioral
Sciences (2nd ed.). Hillsdale, NJ: Erlbaum Associates, 67-69, 490-497.

Cole D.C., Mondlock, M.V. & Hogg-Johnson, S. (2002). Listening to injured workers:
How recovery expectations predict outcomes: A prospective study. Canadian
Medical Association Journal, 166, 749-54.

Crook, J., Milner, R., Schultz, I. Z. & Stringer, B. (2002). Determinants of occupational

disability following a low back injury: a critical review of the literature. Journal
of Occupational Rehabilitation, 12, 277—295.

85

Crook, J., Moldofsky, H. & Shannon, H. (1998). Determinants of disability after a work
related musculoskeletal injury. Journal of Rheumatol, 25, 1570—1577.

Currie, S. R., & Wang, J. (2004), Chronic back pain and major depression in the general
Canadian population. Pain, 107, 54—60.

Dawson S. E. (1994). Workers' compensation in Pennsylvania: the effects of delayed
contested cases. Journal of Health and Social Policy 6(1), 87-100.

Dembe, A. E. (2001). The social consequences of occupational injuries and illness.
American Journal of Industrial Medicine, 40, 403-417.

Devore, J. L. & Berk, K. N. (2007). Modern mathematical statistics with applications.
Belmont, CA: Thompson Higher Education, 600-601.

Dworkin R.H., Handlin, D. S., Richlin, D. M., Brand, L. & Bannucci, C. (1985).
Unraveling the effects of compensation, litigation, and employment on treatment
response in chronic pain. Pain, 23, 49—59.

Eaton, M. (1979). Obstacles to the vocational rehabilitation of individuals receiving
workers’ compensation. Journal of Rehabilitation, 45(2), 59-63.

Fisher, RA. (1915). "Frequency distribution of the values of the correlation coefﬁcient in
samples from an indeﬁnitely large population". Biometrika. 10 (4), 507—521.

Galizzi, M., Boden, L. I. & Liu T-C. (1998). The workers’ story: results of a survey of
workers injured in Wisconsin. Cambridge, MA: Workers Compensation Research
Institute.

Gallagher, R. M., Williams, R. A., Skelly, J., Haugh, L. D., Rauh, V., Milhous, R. &
Frymoyer, J. (1995). Workers' compensation and return-to-work in low back pain.
Pain, 61(2), 299-307.

Gardner, J. A. (1991). Early referral and other factors affecting vocational rehabilitation
outcome for the workers' compensation client Rehabilitation Counseling Bulletin,
34, 197-209.

Gatchel, R. J., Polatin, P. B. & Mayer, TC. (1995). The dominant role of psychosocial
risk factors in the development of chronic low back pain disability. Spine, 20,
2702—2709.

Glass, G. V. & Hopkins, K. D. (1996). Statistical methods in education and psychology
(3rd ed.). Boston: Allyn & Bacon.

86

Greenough, C. G., Taylor, L. J., & Fraser R. D. (1994). Anterior lumbar fusion: A
comparison of noncompensation patients with compensation patients. Clinical
Orthopaedics and Related Research, 300, 30—37.

Greenwood, P.E. & Nikulin, MS. (1996) A guide to chi-squared testing. Wiley, New
York.

Gumerrnan, S. H. (1998). Examination of four potential predictors of return to work in
mid-career low back injured workers' compensation recipients. Dissertation

Abstracts International, 59 (09B). (UMI NO. AA19829785).

Hester, E. J., Decelles, P. G., & Gaddis, E. L. (1986). Predicting which disabled
employees will return to work: The Menninger RTW scale. Topeka, KS:
Menninger Foundation.

Hildebrandt, J., Pﬁngsten, M., Saur, P., & Jansen, L. (1997). Predictors from a multi-
disciplinary treatment program for chronic low back pain. Spine, 22, 990—1001.

Hosmer, D. W. & Lemeshow, S. (2000). Applied logistic regression (2nd ed.). New
Jersey: John Wiley & Sons, Inc.

Hosmer, D. W. & Lemeshow, S. (1989). Applied logistic regression. New York: Wiley

Huck, S. W., & Cormier, W. H. (1996). Reading statistics and research (2nd ed.). New
York: Harper Collins.

Imershein, A. W., Hill, S. & Reynolds, A. M. (1994). The workers' compensation system
as a quality of life problem for workers' compensation claimants. Advances in
Medical Sociology, 5, 181-200.

Jefferson, J. R. & McGrath, P. J. (1996). Back pain and peripheral joint pain in an
industrial setting. Archives of Physical Medical Rehabilitation, 77(4), 385-390.

Kendall, E. & Buys, N. (1998). An integrated model of psychosocial adjustment
following acquired disability. Journal of Rehabilitation, 64, 16-20.

Keogh, J., Nuwayhid, 1., Gordon, J. & Gucer, P. (2000). The impact of occupational
injury on injured worker and family: outcomes of upper extremity cumulative

trauma disorders in Maryland workers. American Journal of Industrial Medicine,
38, 498-506.

Krause, N., Frank, J. W., Dasinger, L. K., Sullivan, T. J., & Sinclair, S. J. (2001).
Deterrninates of duration of disability and return-to-work after work-related injury
and illness: Challenges for future research. American Journal of Industrial
Medicine, 40, 464-484.

87

Lancourt, J. & Kettelhut, M. (1992). Predicting return to work for lower back pain
patients receiving worker's compensation. Spine, 17, 629—640.

Linton S. J. (2000). A review of psychological risk factors in back and neck pain. Spine,
25, 1148-1156.

Livneh, H. & Antonak, R. F ., (1997). Psychosocial adaptation to chronic illness and
disability. Gaithersburg, MD: Aspen Publishers.

Mashaw, J. & Reno, V. (1996). Balancing security and opportunity: The challenge of
disability income policy. Report of the Disability Policy Panel: Washington DC.

Melhorn, J. M. (2000). Cost effective management of musculoskeletal disorders. In The
24’” Annual Carroll P Hungate Postgraduate Seminar on Occupational and
Environment Health (pp. 1-45). Overland Park, KS: Great Plains College of
Occupational and Environment Medicine.

Morse, T. F ., Dillon, C., Warren, N., Levenstein, C. & Warren, A. (1998). The economic
and social consequences of work-related musculoskeletal disorders: The
Connecticut upper-extremity surveillance Project (CUSP). International Journal
of Occupational and Environmental Health, 4, 209-216

National Institute of Handicapped Research (n.d.). Preventing disability dependence:
Return-to-work studies. Rehabilitation Brieﬁ 9(3), 1-4.

Ott, R. L. & Longnecker, M. (2001). An Introduction to Statistical Methods and Data
Analysis, Fifth Edition. Paciﬁc Grove, CA: Duxbury.

Pransky, G., Benjamin, K., Hill-Fotouhi, C., Himmelstein, J., Fletcher, K. E., Katz, J. N.
& Johnson, W. G. (2000). Outcomes in work-related upper extremity and low

back injuries: Results of a retrospective study. American Journal of Industrial
Medicine, 37, 400-409.

Pransky, G., Snyder, T., Dembe, A. E & Himmelstein, J. S. (1999). Under-reporting of
work-related disorders in the workplace: a case study and review of the literature.
Ergonomics, 42(1), 171-182.

Reid J., Ewan, C. & Lowy, E. (1991). Pilgrimage of pain: The illness experiences of
women with repetition strain injury and the search for credibility. Social Science
Medicine, 32, 601-612.

Reville, R. T. (1999). The impact of a disabling workplace injury on earnings and labor
force participation. In Haltiwanger, J., Lane, J., Spletzer, J. R., Theeuwes, J. &
Troske, K. (Eds), The creation and analysis of employer-employee matched data,
Amsterdam: Elsevier Science.

88

Rodgers, J. L. & Nicewander, W. A. (1988). Thirteen ways to look at the correlation
coefﬁcient. The American Statistician, 42, 59—66.

Rosenman, K. D., Gardiner ,J. C., Wang, J., Biddle, J., Hogan, A., Reilly, M. J., Roberts
K. & Welch, E. (2000). Why most workers with occupational repetitive trauma do
not ﬁle for workers' compensation. Journal of Occupational and Environmental
Medicine, 42(1), 25-34.

Schultz, I. Z. (2003). The relationship between psychological impairment and
occupational disability. In Schultz I. Z. & Gatchel, R. J. (2006). Handbook of
complex occupational disability claims: Early risk identiﬁcation, intervention,
and prevention. New York, NY: Springer Science Business Media, Inc.

Schultz, I. Z., Crook, J., Meloche, G. R., Berkowitz, J., Milner, R., Zuberbier, O. A., &
Meloche, W. (2003). Psychosocial factors predictive of occupational low back
disability: Towards development of a return-to-work model. Pain, 107, 77-85.

Schultz I. Z. & Gatchel, R. J. (2006). Handbook of complex occupational disability
claims: Early risk identification, intervention, and prevention. New York, NY:
Springer Science Business Media, Inc

Silverstein, B. A., Stetson, D. S, Keyserling, W. M. & Fine, L. J. (1997). Work-related
musculoskeletal disorders: comparison of data sources for surveillance. American
Journal of Industrial Medicine, 31(5), 600-608.

Stone, D. (1984). The disabled state. Philadelphia, PA: Temple University Press.

Strunin, L. & Boden, L. I. (2004). The workers’ compensation system: Worker friend of
foe? American Journal of Industrial Medicine, 45(3), 338-345.

Sullivan, M. J., Feuerstein, M., Gatchel, R., Linton, S. J, & Pransky, G. (2005).

Integrating psychosocial and behavioral interventions to achieve optimal
rehabilitation outcomes. Journal of Occupational Rehabilitation, 15, 475—489.

Suter, P. B. (2002). Employment and litigation: improved by work, assisted by verdict,
Pain, 100(3), 249-257.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Boston,
MA: Allyn and Bacon.

Tate, R. B., Yassi, A. & Cooper, J. (1999). Predictors of time loss after back injury in
nurses. Spine 24, 1930—1935.

Tate, D. G. (1992). Workers’ disability and return to work American Journal of
Physical Medicine and Rehabilitation, 71(1), 92-96.

89

Texas Workers’ Compensation Research Center. (1995). Economic outcomes of injured
workers with permanent impairments. Res. 27(3), 1-4.

Thomason T. L. (1989). The compensation of permanent partial disability in New York
State: An examination of wage-loss and ex ante workers’ compensation beneﬁts.
Dissertation Abstracts International, 50 (07), AAT 894535.

Thomason, T. L, & Burton, J. F. (2001). Econorrric effects of workers’ compensation in
the United States: Private insurance and the administration of compensation
claims. Journal of Labor Economics, 11(1), Sl-S37.

Turner, J. A., Franklin, G., Fulton-Kehoe, D. Sheppard, L., Wickizer, T. M., Wu, R.,
Gluck, J. V., Egan, K., & Stover, B. (2007). Early predictors of chronic work
disability associated with carpal tunnel syndrome: a longitudinal Workers'
Compensation cohort study. American Journal of Industrial Medicine, 50(7), 489-

500.

Volinn, E., Van Koevering, D., & Loeser, J. D. (1991). Back sprain in industry: the role
of socioeconomic factors in chronicity. Spine, 16, 542-548.

Wolkinson, B. W. & Block, R. N. (1996). Employment Law: The Workplace Rights of
Employees and Employers. New York, NY: Blackwell Publishing.

90

A

”'iillilliiilliiiliiiiii“