.‘ . ”Manx: .,. .11; um: «$.11 1 ~. .. u‘fw M7.L _. . 1.1 1‘ Cad-'1. x 1 . \ $17511. "1'- '1‘“! . ‘4 ' 2W: 1:1 4114.12: Evita-u In 11.1. r2. r .-... --_. .15“. ‘i. .1- ‘1 1 1 1 . 1. .1 .51112111‘ L (31.11.1er 11 -- » -1- ’15. . -.1u 1u-4um1u.n..1. wm -. ~13: 1.11.11}: 1 1 1 1.1.1 'wmuL-rrw . . . 1 ..,. - ..:,;‘J.;., ‘1 . '1'” .1 .‘111'11 1-. 1. «1:11.. - 'Aeu' ‘ng 154'. _ ~fr-1-4-‘U4 3: 'b 4- .4- , ' '4 . . ‘ , .f 1‘1!" u “71.1w"1‘.l.1-’.;4"f 1 I f 1 ' u r' 1 ._. . fl ‘ -' v' “a 51':- F7; 11351519" .111 -I-- wumv.’ nu I‘un' - 1.4 ~ 1- '1 ' :1 I1 1 1 71 J .. ;."' fuel-~71 3'31141 r. w- . n -- 11-1 - ( . .1 - i-v-r 1 1 [rug—14.. 7: ".111 . 4' , .-1- .. » u 1: J1 ,{’ 92$ . 1“ :11 . 1 11 » 1 ‘ . .1. ”1-. “1.43- ,1 «4 . ‘U'T'l‘ ' ' A I ' ‘1“"f‘ ft " "3;". . l ["1" I I' l 1. 11mm: 1 1 .— 4 1f 1 1 .11. 1 -—1 1- npfizwfnn’ 1. «- .s-Imu “11-1.... -. . . 1' ..- >1 1, '1 . r .11 u...” 4- any row-114.12.. 1:". filil‘luk Fl 'I-vu . _ . J » .3, .J "x' ‘ ' 1- v ‘-1‘/‘1LU' fll' | ' N." x I.- vuwm {guen‘sru - (1 2- » v 1.4".1w1"l|n1‘r .1- . . . . .. '1‘!- Wm?" 1.3.3.1 ,“J, 1’1;'1‘"- 1 1 '1'"; L .. 1.1... . n '... . 1 3 . - w ' 1'. 3-5th! -1.- 1U.’:.::. .. . :1 KEN...” , . 1 1 4 J , r. .... ‘2'... N H 3.1"... 1.17:1...“ ,.. . :‘l.- m 1. ' . ,.. ., ,. l .. 1.1,... .51.. “‘1 u».v‘.~."{.fiu;. ~_1- . ,.. .1«.11:.1- 1.1 I' ..,..,.1. n ‘ vatavuiig z': .-""‘ . . r 1.1 u... ,. . .. .11. . . I"? . 10m. ”1 -\1v 1.1 .1. . , ’ .. .,.._.',,', . 24.3 1.1-3“; wxpfuu “51' <1 ,.. 1 . .1 .11.... 1.1.1.1.... 21,. , in} T1 ‘13.. 1 1:" 1 , 2'1" ‘15-'1- '1'” '1' 1-.- 1‘ 3., Apt»:- Maw.“ 1- . . nr‘r ' ’V ' - ~ ' ‘ - ~ 7 ' a {rm-J14 ' ' ‘..IL:I ' " '*“ I -r I 1- M u— 4i 4-. .-- 1 1| 1 1 "4rl'~-'rII-I: . \. mm". h1:1‘*‘!‘.>g; _ L 11 . .- . . ._ m}. “'~ " ' f . 1 I. '1 -.1 1 - 1- - 1'... 1 . ”(rummauo. mum-N n .-.. . 1 . 1-1.1 1. . ,. , . “‘2“ .“ .... T2.“ 1 31,. .. . 1 k . 1. . , , .’ .1 1.1.3.16 119-54 :1 (""ve‘ 1 "1.111.: .1 . 1 «I . 1 1‘11142 .. an 4 ,1. . .x—‘u'pcan-uu» run!" '1 -¢- 1 ' ' w . J 1 . u... 4-: un . 3 p 1. .gfl. 1. 1 . 1- , - u, I 1.. ,.. ’.‘i ' '. , ”0.11.1. ., M..K....u....,..,.1m. . , _ .-. ~11. .» w 1 -., mun-m 1111? 1 . . 1 r . .. 1.“, . 33:11.“. 1. "xv-11w. 1,11,... ... 1. u 1521'?” - ,-I. 4 r Y. 1 1 _ 1 .1- 1 1 w: "as? ’. u v. '1' 1w... an. .1111..,'.,..., L... ,., v . “~ ' a 4- .. nu ~11. . k ~-:v~ .. . , . . . 1 a, _" ’ “ 1. .. .. .1 z" ' '1' ’ E". .'.'=‘:;- a, \ m." " I..t'£‘:'-r”'.r"-~; . ~‘ - -~'f_'r-'-'"'-~ ' -'-~ "'1“ {5.151322 '2'" .. -1.-. r1 . . . . '.‘j......~...,...,.....u ”~14. - 1. 1 4.: '.‘..' "mu-.31:- '17.- . ._ 1 .11 1 up: 11 x .1 r4 4 «1.». .1... my- ‘n-o 1 "w x 11". .1457"..- ... 1 111-; r... ....,.r,.:.-',..-::1'.‘. '1. 1 - .4 1..u.1.4=:‘ .. ... .,. 114' A‘u-m- 1.5.1::Z'n- .. 5.. "‘ '12.?" 41;».w1u11‘... ,,. .;,. " év‘e'vr‘nr. .. n. . . .. 2‘ 1v - -’vrx 41.14335.“ 1 — . vv-311'y.m”4r41~..... , - .. .1: - .'.. rl_ unis-w,m¢.:, . vy- 1'12-11—1-43: _. 4-2. ' ”1-1114 1 :9r'” r, r‘ :v'n'r run-34m... 4 :I‘rurrufr' .a r_ ., 1 4~v1.v'rr. ni- . 12..."... 'Ju'v .. .. h. . . r “'1 n m 4 t £32.:«vv' 1 . m... . 11...; unnumm. "22.12% «2'z'v‘t‘3-fly --~.'1‘~Xe~.- 4 2 cu —.r-1' 1- '- ”:2. NS STAT llllllllllllll ll lllllllllllllllIlllllllllllllllllll 31293 00882 6327 This is to certify that the dissertation entitled THE RELATIONSHIP OF AFFECT TO TREATMENT AND OUTCOME IN CHRONIC LOW BACK PAIN presented by PAUL W . DELMAR has been accepted towards fulfillment of the requirements for , ” ,' 7 // / 'Z // degree in g’ xxxflf'g" 5‘67 ; « 4, ’ /’. 3‘” . / / /'/::’:'; [/5 4 / - \- f - '"V 5. V Major professor Date J L'C' 4-7,, MSU is an Affirmative Action/Equal Opportunity Insn'ruu'on 0-12771 MBRARY Miehigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE “1' I ..’.,~ rm; 9 my] up; /75 ,1 @9150: AUG (1809M \\\ MSU Is An Affirmative Action/Equal Opportunity Institution c:\civc\datedtnom3—p.1 — THE RELATIONSHIP OF AFFECT T0 TREATMENT AND OUTCOME IN CHRONIC LON BACK PAIN By Paul w. Delmar A DISSERTATION Submitted to Michigan State University in partiai fulfiilment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseiing, EducationaI Psychoiogy, and Special Education 1992 ((7/5 5 CL” 17/ ABSTRACT THE RELATIONSHIP OF AFFECT T0 TREATMENT AND OUTCOME IN CHRONIC Low BACK PAIN By Paul N. Delmar This two-part study of patients with chronic low back pain examined the relationship between measured affect and rehabilitation outcomes. The first part reviewed medical records of subjects who completed a multidisciplinary pain program at Mary Free Bed Hospital and Rehabilitation Center (MFB). The second part consisted Of a follow—up of the current status of these past patients. Medical records provided demographic and testing information concerning the patients. Scores on the Affective dimension of the McGill Pain Questionnaire (MPQ) were obtained at intake and at weekly intervals during treatment from 75 Chronic low back pain patients. Follow-up information was obtained through the use of MFB’s 6- and 12—month follow—up questionnaire. An additional request was made for the client tO complete the Productivity Of Life Questionnaire voluntarily. The purpose was to determine the existence and nature of any relationship between scores on the MPQ’s Affective dimension and outcome results of a multidisciplinary pain clinic (MPC) treatment Paul N. Delmar program. Second, the researcher attempted to determine whether different populations existed within the sample and, if so, how the various populations’ levels on the MP0 Affective dimension differed. The outcome data did not substantiate the hypothesis that chronic pain patients’ scores on the MPQ’s Affective dimension would correlate positively with successful outcome. Correlation and regression analyses were used to evaluate these issues. To evaluate affect over time, two procedures were used. The analysis of variance (ANOVA) demonstrated a significant quadratic effect and a significant cubic effect. The hierarchical linear method was used to evaluate the level of affect at intake, 4 weeks, 8 weeks, and 12 weeks. The variabilities in both the slope and intercept were significant at all four time points, indicating that the slope and intercept differed significantly among subjects. The variables that were significant in this analysis were reviewed and discussed. The chi-square test of association was not significant in assessing the relationship between membership in the worker’s compensation, auto no-fault, or noninsurance subsample group and outcome. An ANOVA was done to determine whether group differences existed on the productivity variable. No differences were found by group membership. menu. ACKNOWLEDGMENTS I would like to thank the various individuals who have assisted me in the development of my research proposal through the completion of the dissertation. First, I must thank my wife, JOAnn, for her support and understanding throughout the completion of this project. My sons, Eric and Christopher, have been very patient with me. Their faith and confidence in me provided the motivation to gO on at various times in this effort. I believe my family will have benefited from my efforts; I intend to be very attentive to their specific needs. The members of my committee must be singled out and thanked. I thank Nancy Crewe, my advisor, for her assistance, direction, and support throughout this entire project. Her willingness to share her time and skills with me from the original concept through repeated and demanding revisions attest to her patience and skill, keeping me focused on the task at hand. Richard Johnson has been supportive and very helpful in his assistance in refining the purpose and direction of the research project. His support is greatly appreciated. Betsy Becker has been tremendously supportive and helpful in overall editing and providing direction for the statistical analysis procedures. Her friendship and humor assisted in completing this task. Donald Stanton has been a source of support and medical information. I appreciate the time and information he so freely shared with me. His contributions were many, the least of which was the broadening of my compassion for other individuals. The staff and management Of Mary Free Bed Hospital and Rehabilitation Center must be thanked. I specifically thank Daneen Caro, administrator of the Chronic Pain Program; Sue Couturier, secretary; and Ed Kremer, director of the Chronic Pain Program. Ed provided me with the topic of pain measurement and helped refine the study, acted as a resource person, and supported my effort through— out this process. I also thank the patients of Mary Free Bed’s Chronic Pain Program for their participation in the project. Ken Austin provided the data entry. Christine Schram was my statistical consultant for this project. Steve Raudenbush and Mike Seltzer assisted in the more technical areas of HLM. Sue Cooley provided editorial and final word—processing services. Bob Winborn and James Engelkes provided support, edito— rial assistance, and direction at various stages of writing the dissertation. Personal thanks are extended to Rollie Davis for his computer assistance, and to him and Wayne Workman for providing basic shelter and accommodations. TABLE OF CONTENTS LIST OF TABLES .............. LIST OF FIGURES .............. . ....... Chapter I. THE PROBLEM . . . . . ........ Introduction ............ . . . . Identification Of the Problem ....... . Purpose of the Study ........ . . . Importance of the Study ...... . Generalizability .......... Research Questions ...... Research Hypotheses ...... Assumptions . . . ...... Definition of Terms . . . . Overview . . . . . . ...... II. LITERATURE REVIEW . . . ........... Introduction . ..... . . . Overview Of the Concept of Pain . . Pain Receptors ....... The Transmission of Pain Signals A— Delta Pain Fibers ..... Type—C Pain Fibers . . Double Pain Signals ..... Specific Pain Pathways Theories of Pain . The Specificity Theory of Pain Pattern Theory Summary of the SOecifiCity and Pattern TheOries . of Pai In . The Gate ContrOl TheOry . Behavioral Methods for Treating Chronic Pain The Respondent Pain Model . . . The Operant Pain Model . Cognitive-Behavioral Therapy vi Page xi Review of Treatment-Outcome and Follow-Up Studies Of the Multidisciplinary Treatment Of Low Back Pain ...................... 6( Introduction .................. 6( The Operant Approach .............. 62 The Relaxation Approach ............. 61 The Cognitive Approach ............. 7E The Multimodal Approach ............. 7] Conclusions ................... 8l Predicting the Treatment Outcome of Multidisci- plinary Pain Clinics .............. 92 Introduction .................. 92 Prediction Studies ............... 95 Pain Measurement ................. IOE The McGill Pain Questionnaire ..... . . . . Ill The Affective Dimension of the McGill Pain. Questionnaire ............ . ..... l3l Affect and Worker’s Compensation . . . . ..... 132 Summary ...................... 131 III. METHODOLOGY . . . ........... . . ..... I35 Introduction ................... T35 Research Questions .............. . . l3€ Research Hypotheses ................ T36 Description Of the Study Site and the Treatment Process ..................... l3} Selection and Description of the Sample . . . . . . l3£ Instrumentation .................. l4( The McGill Pain Questionnaire .......... l4l The Productivity Of Life Questionnaire ..... 141 The Follow-Up Questionnaire ........... l4l Data-Collection Procedures ............ l4! Data-Analysis Procedures ........... ' . . l4( Observational Research .............. 152 Summary ...................... 15¢ IV. RESULTS . . . . . . . . . . . . . . . . . . . . . . . TSE Introduction ............. . . . . . 155 Demographic Characteristics of the Sample . . . . . lSE McGill Pain Questionnaire Results . . . . . . . . . l5£ Correlation Data .......... . . ..... lSS Regression Analysis . .......... . . . . . l6l Missing Data . . ........ . . . . . . . . l6! Affect by Time . . ........ . . . . . . . . l6! The Hierarchical Linear Model . . . . . . . . . . . l6£ HLM Analyses . . . ..... . . . . . . . . . . . I72 vii Intake Analysis ............. . . . . Week 4 Analysis ......... . . . ..... Week 8 Analysis .............. . . . Week l2 Analysis ........... . . . . . Summary of the Four Time—Point Analyses ..... Success—by-Compensation Analysis ......... V. SUMMARY, DISCUSSION, AND IMPLICATIONS . . ...... Summary ...................... Review of Literature .............. Purpose ..................... The Study Sample ................ Method ................ . . . . . Correlation and Regression Data . . . . . . . . Affect by Time ................. Success-by-Compensation Analysis . . ...... Discussion ............. . . . . . . . Limitations of the Study ............. Implications for Future Research . . . . ..... APPENDICES A. THE MCGILL PAIN QUESTIONNAIRE ........ B. LETTER TO SUBJECTS ............... THE PRODUCTIVITY OF LIFE QUESTIONNAIRE RANKINGS AND WEIGHTINGS OF PRODUCTIVITY OUTCOMES OF) E. MARY FREE BED FOLLOW—UP QUESTIONNAIRE ........ F. APPROVAL LETTER FROM THE UNIVERSITY COMMITTEE ON RESEARCH INVOLVING HUMAN SUBJECTS . . . . . G. MARY FREE BED PROGRAM-—ORIGINAL INFORMATION SHEET . H. DATA SUMMARY FORM ........... REFERENCES .................. viii Page T80 T8T l82 l83 T83 T84 T86 T86 T87 T90 T9T T92 T93 T94 l95 T95 205 208 209 210 211 212 215 216 224 225 h-h-D-h 4’) LIST OF TABLES Outcome Studies Using the Operant Approach . . Outcome Studies Using the Relaxation Approach Outcome Studies Using the Cognitive Approach . Outcome Studies Using a Multimodal Approach Outcome Studies Using a Prediction Approach Distribution of Dichotomous Variables Used in Analysis . . . . . . . ...... . . . . . Distribution of Continuous Variables Used in Analysis ......... . . . . . . . . McGill Pain Questionnaire Results Correlation Coefficients . Regression Analysis for Outcome Variables in the Equation for Outcome Regression Analysis for Productivity . Variables in the Equation for Productivity . Mean Scores on the Affective Dimension of the MPQ for Each Time Period . . . . . . . . . . . . Analysis of Variance: Affect Over Time Average Slope Coefficients Across Individual Growth Models . . . . . . . ...... . . . . . . . . Statistical Table for HLM Coefficients at Intake (L = 0) . . . . . . . . . . . . . . . . . . . Variance Components Table at L = 0 . Page 65 70 78 82 107 156 156 159 160 161 161 162 163 166 168 173 176 176 Table Page 4.T4 Statistical Table for HLM Coefficients at L = 4 Weeks ....................... T77 4.T5 Variance Components Table at L = 4 .......... T77 4.16 Statistical Table for HLM Coefficients at L = 8 Weeks ....................... T78 4.T7 Variance Components Table at L = 8 .......... T78 4.T8 Statistical Table for HLM Coefficients at L = T2 Weeks ....................... T79 4.T9 Variance Components Table at L = T2 . . . ...... T79 4.20 Outcome for Patients in Three Compensation Groups . . T84 4.2T Productivity by Worker’s Compensation . ....... T85 Figure 2.1 LIST OF FIGURES The Anterolateral System of Spinothalamic, Spinoreticular, and Spinotectal Fibers Conveys Information About Pain to Several Regions of the Brain Stem and Diencephalon Schematic Drawing of the Dorsal Horn of the Spinal Cord, Illustrating That the Nociceptive Neurons, Whose Axons Form the Ascending Antero— lateral System, Are Found in Lamina I and Lamina V of the Dorsal Horn . . . . . Effects of Myelinated Afferents on the Pain- Transmission Cell . . ..... . The Gate Control Theory of Pain . . Schematic Diagram for the Components of Pain (Only Pain Behavior is Measurable and Observable) . Weeks by Average Affect (All Subjects) Average Affect by Success/Failure . Average Affect Intercepts ...... xi Page 22 23 39 41 45 T67 T67 T73 CHAPTER I THE PROBLEM Introduction Persistent complaints of pain inflict significant economic loss on society. Bonica (1980) estimated that the cost of acute and chronic pain to the United States economy is more than $90 billion annually. Casey (T979) estimated that five million people in the United States are at least partially disabled by persistent low back pain and that 93 million workdays are lost each year due to this affliction. For as many as 78% of the individuals who are severely disabled by chronic low back pain, nO pathophysiological basis can be found for their complaint (Loeser, T980). Traditional medical care Often has failed to relieve chronic pain because such treatment is based on an acute-care model. The acute-care model assumes there is Objective evidence of a pain gen- erator. Acute care involves evaluating the symptoms and attempting to treat the immediate cause. Because chronic pain patients’ com— plaints do not respond to acute-care treatment, these individuals often become problems in the health care system. Patients with chronic pain have become recognized as a specific treatment popula— tion. To reduce the effects Of pain, behavioral treatments have been developed over the past 20 years. The multidisciplinary pain clinic (MPC) was developed to meet the challenge presented I patients with persistent complaints of low back pain. Research conducted during the past 20 years has providi considerable support for the efficacy of behavioral techniques fI reducing pain perception and pain-related behaviors. Outcor studies have documented positive results of such techniques increasing levels of productivity and a return to normalcy chronic pain patients (Linton, T982, T986). Outcome and fOTlow-I studies have demonstrated that substantially and statistical' significant improvements in reducing pain-related behavior can I made through the use Of MPCs. Initially, MPCs emphasized in-patieI treatment programs. Fordyce, Fowler, Lehman, and de Lateur’s (T961 pioneering effort was an in-patient-based, operant program that 11 the way for others to follow (Cairns & Pasino, T977; Cairns, Thoma: Mooney, & Pace, T976; Fowler, T975; Greenhoot & Sternbach, T97l Sternbach, l974; Swanson, Maruta, & Swenson, l979; Swanson, Swensou Maruta, & McPhee, T976). After MPCs’ results were documented, movement toward out-patient programs became apparent (Arkawa, l98' Chapman, Brena, & Bradford, T98l; Herman & Baptiste, T98l; Wang Illstrup, Nauss, Nelson, & Wilson, T980). The shift to out-patieI programs has substantially reduced the cost Of these programs making them more accessible to chronic pain patients (Linton, T982 T988). Several researchers have attempted to identify variables 1 predict the treatment outcome of patients in MPCs (Block, Kremer, Gaylor, T980; Dworkin, Richlin, Handlin, & Brand, T984; Keefe Block, Williams, & Surwit, l98l; Kleinke & Spangler, l988; Maruta, Swanson, & Swenson, T979). Prediction studies traditionally have avoided including recipients Of worker’s compensation because of the implications of "compensation neurosis." The prevailing view has been that patients on compensation or awaiting litigation are neurotics or malingerers--people trying to live Off the efforts of others by feigning their pain (Melzack, T985). Researchers have reported that patients receiving compensation had a significantly poorer outcome than those who were not receiving compensation (Block et al., T980; Finneson, l977; Hammonds, Brena, & Unikel, T978; Herman & Baptiste, 1958). In other pain—clinic outcome studies, no relationship was found between compensation and treatment response (Arnoff & Evans, T982; Brena, Chapman, & Bradford, T980; Brena, Chapman, & Decker, 1981; Chapman & Brena, T982; Maruta et al., l979; Painter, Seres, & Newman, T980; Rosomoff, Greene, Silbert, & Steele, T98l; Seres, Painter, & Newman, l98l; Swanson, Swenson, Maruta, & Floreen, T978). Further study is needed to be able to predict which individuals are best served by MPCs. Measuring the level of affective distress is one way of determining differences among groups of patients with chronic pain. It is also a way of determining which individuals’ suffering is greater than that of others. Identification of the Problem Scientific measurement procedures have focused on pain as an individual sensory quality that varies only in intensity. Although intensity is a noteworthy dimension Of pain, to describe pain solely in terms of intensity is like describing the visual world only in terms of light, failing to consider color, pattern, texture, and other dimensions of the visual experience. Pain is a complex perception, rather than a simple sensation. This complexity extends beyond the multiple dimensions of sensation; affective and motivational aspects must be recognized. Failure to consider the motivational-affective dimension of pain has seriously limited the total picture of the pain experience. The motivational dimension is crucial to the concept of pain as a perception. The motivational dimension comprises an individual’s perception Of past experience, attention to the perceived stimuli, sensory stimuli, and intensity of the experience. If researchers consider only the sensory features Of pain, they are ignoring the most important aspect of pain-—its motivational and affective properties (Melzack & Wall, T983). In an effort to develop an instrument to measure the sensory, intensity, and nmtivationaT—affective properties Of pain, Melzack and Torgerson (l97l) constructed the McGill Pain Questionnaire (MPQ), a verbal pain questionnaire that would allow for the quantification of the concept Of pain. A verbal questionnaire was chosen because of the importance Of the use of language. Individuals suffering from pain use language as a medium to relate the perception of pain to significant others in their lives. The MPQ is a paper—and—pencil instrument designed to quantify the dimensions of the pain experience. The dimensions are sensory, affective, evaluative, and miscellaneous. Subjects are shown 20 sets of word descriptors and asked to select those words that are relevant. The most appropriate word in each word set is circled. Each set contains up to six words. Ten word sets describe sensory qualities, five are affective descriptor sets, a single set describes the evaluative dimension, and the rest are classified as miscellaneous. The researcher scores the total number of words that apply to the pain. The words within each word set have been assigned rank orders, so one is able to compute the total rank Of the words chosen. Researchers have found that elevated scores on the Depression and Hysteria scales of the Minnesota Multiphasic Personality Inventory (MMPI) were associated with higher scores on the Affective dimension of the MPQ (Aronoff & Evans, T982; McCreary, l98l). In addition, elevated scores on the Depression and Hysteria scales of the MMPI have been associated with positive outcome in studies of multidisciplinary clinics for the treatment of chronic low back pain (Kleinke & Spangler, l988l Painter, Seres, & Newman, T980). Thus, there might be a relationship between elevated scores on the Affective dimension Of the MPQ and positive outcome for patients receiving treatment from an MPC. However, the nature of that relationship has not yet been explored. This research was undertaken to address that problem. The researcher also determined how to better understand the characteristics of individuals receiving treatment through an MPC and to assess the outcome using the affective dimension of the MPQ. Purpose of the Study The purpose of this study was to determine the existence and nature of the relationship between patients’ scores on the Affective dimension on the MPQ and the outcome of an MPC treatment program for chronic pain. The researcher also attempted to determine whether the three treatment groups within the study population (patients receiving worker’s compensation, patients covered by no—fault automobile insurance, and those receiving no financial support) differed significantly in their scores on the Affective dimension of the MPQ over the course of treatment: at intake, upon completion of the program, through follow-up. Importance of the Study This study is important for many reasons. The researcher sought to determine whether a relationship exists between patients’ scores on the Affective dimension of the MPQ and the results of an MPC treatment program for chronic low back pain, and to measure the extent of that relationship. This information will help distinguish between individuals with greater potential for success in treatment programs and those with less potential for success. Such knowledge will allow practitioners to use resources more effectively in determining the appropriateness of client selection for particular treatment programs. It also will allow clinicians to use their time more efficiently to investigate barriers to success in individua' clients. This study is the first to use a Hierarchical Linear Mode' (HLM) to analyze the chronic pain population. Previous research OI individual change has been plagued by conceptualization and design inadequacies (Bryk & Raudenbush, T988). Pretest and posttesi designs are generally inadequate for the study Of individual change (Bryk & Weisberg, T977; Rogosa, Brand, & Zimowski, T982). ThI': means pretest and posttest designs are inadequate for studying effects on learning because learning is a process of individua' change involving the acquisition of knowledge, in this study the knowledge of pain-reducing techniques, and time skiTl Of employing these techniques over time. Bryk and Raudenbush believed "researcl on learning requires multi-time point data and a statistical mode' that permits an explicit representation of individual growth" (p 67). The results of this study will include statistics derived using new advances in analyzing multilevel data by employing the HLI procedure. This will allow for the enhancement of scientifi< understanding Of the MPQ by using an appropriate form Of statistica' analysis. This means that the researcher evaluated the Affectiw dimension of the MPQ in a manner that has not been used to date More important, the researcher used an advanced and more appropriati method of statistical analysis to analyze the data. As a consequence, the results of the study will be morI statistically sound, Tending themselves to a greater understanding of the relationship of the affective dimension Of the MPQ, historical and demographic variables, and outcome results following treatment in an MPC. The study should answer whether or not a relationship exists between the level of affect of the MPQ and outcome. If there is a relationship, does it exist for the various groups involved? The individual patterns Of affect over time were analyzed to determine whether patterns of growth exist. Generalizabilitv Kleinke and Spangler (T988) addressed methodological issues in their overview Of research involving MPCs. They found five methodological issues inherent 'hI previous research investigating variables relating ix> treatment outcomes. 'These issues are as follows. First was the need for a unified definition of successful treatment outcome. Kleinke and Spangler objected to indices Of treatment outcome based on combinations of patient and therapist ratings. The second issue was that most previous studies predicting outcome used univariate statistics subject to Type I error. The third objection was based on the fact that variables used for prediction of outcome were Often intercorrelated. The fourth objection related tO whether studies predicting outcome should look at patients’ discharge scores or whether they should look at intake scores. The fifth issue stemmed from the need to define the kinds of pain patients for whom the researchers were trying to predict outcomes. The present researcher addressed all of the above-mentioned objections in an effort to enhance replicability and to facilitate the generalizability to other patients with chronic low back pain. First, two measures of outcome were included in this study. One was continuous, whereas the other was dichotomous. These both presented a unified definition of successful treatment outcome. Second, appropriate multivariate statistics were used, minimizing Type I error. Multiple regression and HLM were used to investigate systematically the relationship between the predictor variables and outcome, as well as individual growth. Fourth, this researcher used the affective dimension of the MPQ at intake, through treatment, and through follow—up. The fifth issue was addressed by confining the study to those individuals suffering from chronic low back pain. Research Questions The researcher’s primary purpose in this study was to explore the nature and extent of the relationship between patients’ scores on the Affective dimension of the MPQ and the outcome of an MPC treatment program. A secondary purpose was to determine what combinations Of MPQ Affective score, demographic variables (age, educational level, spouse’s employment, marital status, employment status at intake, involvement in litigation, involvement in rehabilitation, and sources of income), and historical variables (length of chronicity, number of hospitalizations, number of surgeries, and number of past employers) best predict the outcome of treatment. 10 A successful outcome was defined as the individual’s returnh to work, participating in :1 vocational rehabilitation program, I achieving a high level Of productivity on the Productivity Of Li' Questionnaire. An unsuccessful outcome was defined as the patient not being employed, not being involved in a vocation; rehabilitation program, or maintaining a low level Of productivi' on the Productivity Of Life Questionnaire. The following research questions were posed to guide tl collection Of data with which tO achieve the purposes Of tl research. l. Will subjects’ scores on the Affective dimension Of the MI correlate positively with successful outcome following treatment ' an MPC? Individuals who acknowledge psychological distress aI dissatisfied with their present situation or are distressed becau: of prolonged chronic pain. Thus, these individuals may have greatI motivation to succeed in an MPC. 2. Will the MPQ Affective dimension scores Of subjects in tl worker’s compensation group, the group covered by nO-fau' automobile insurance, and the group receiving no ffinancial suppOI differ over the course Of treatment: at intake, upon completion I the program, through follow-up? The researcher examined the pattern of patients’ scores on tl Affective dimension Of the MPQ from intake, through the course I treatment, and at follow-up. The typical level Of affect for tl three treatment groups was then determined. 11 3. Is there a difference in the treatment—success rates of t three groups (those receiving worker’s compensation, those cover by no—fault automobile insurance, and those receiving no financi support)? Research Hypotheses The following hypotheses were formulated to guide the analys of data gathered in this study: Hypothesis 1: Chronic pain patients’ scores on the Affecti dimension Of the MPQ at intake will correlate positively wi successful outcome following treatment in an MPC. Hypothesis 2: Subjects. with high scores on the Affecti dimension of the MP0 at intake will have a nonlinear pattern Affective dimension scores over the course of treatment, up completion of the program, and at follow-up. Hypothesis 3: There will be no difference in the treatmen success rates Of the three study groups: those receivi worker’s. compensation, those covered by no-fault automobi insurance, and those receiving no financial support. Assumptions l. The researcher assumed that both the administration and t scoring of the MPQ were consistent with standardized procedures. 2. It was assumed that line subjects responded truthfully the research instruments. Definition of Terms The following terms are defined in the context in which th are used in this dissertation. Chronic pa_i_n_: Pain that is perceived beyond the expect period Of healing Of an injury (in the absence Of any Observab 12 physical abnormality) 9: that extends for a lengthy period in association with a chronic condition (e.g., arthritic degenerative changes in the spine). Chronic pain syndrome: A condition characterized by deteriora- tion in physical, social, and psychological functioning in people experiencing prolonged pain. Disability: In the context Of health experience, any restric- tion or lack of ability (resulting from an impairment) to perform an activity in the manner or within the range considered normal for a human being. Handicap: In the context Of health experience, a disadvantage resulting from an impairment or a disability that limits or prevents a person’s fulfillment of a {pie that is normal (depending on age, gender, and social and cultural factors) for that individual. Impairment: In the context of health experience, any loss of or abnormality in psychological, physiological, or anatomical struc- ture or function. Malingering: Conscious and willful feigning and/or willful exaggeration of a disease or injury that is the basis for reports of alleged pain in pursuit Of a goal that is usually socioeconomic. Medical rehabilitation: Restoration of an individual to optimal physical and mental functioning in light of his/her impairment. Muscular relaxation: A type of therapy in which patients use instructional tapes or techniques to relax the skeletal muscles. Training may be facilitated by electromyographic biofeedback. Pain behavior: Verbal or nonverbal actions understood observers to indicate that an individual might be experiencing pa and suffering. The actions include audible complaints, faci expressions, abnormal posture and gait, use of prosthetic device avoidance of activities, medication—seeking behavior, and seeking medical assistance. Residual functioning capacity. The ability to perform specif social and work—related physical and mental activities followh medical rehabilitation related to an impairment or when condition(s) has reached a point Of maximum improvement Limitations imposed by an impairment are included in assessment I residual functioning capacity. Work—capacity evaluation: An Objective assessment Of a peI son’s relevant social and functional capacities in order to mal reasonable predictions about his/her ability to learn and perfm discrete job—related duties. Work-hardening Droqrams: Programs designed to help a patim develop enough physical endurance and self—confidence to carry out regular workday (either full time or part time) with acceptab' levels of discomfort and pain, taking into account limitations bag on the assessment of residual functioning capacity. Overview Chapter I contained an introduction to the study; an identif' cation of the problem, purpose, and importance of the study; aI generalizability Of the findings. The research questions hypotheses, and assumptions were stated, and definitions of key terms were given. Chapter II is a review of literature on topics related to the study. Writings and research on the following subjects are included: the physiology of chronic pain, how pain signals are transmitted, theories regarding pain, the results of multidisci- plinary treatment programs for low back pain, prediction of treat- ment outcomes of MPCs, and the Affective dimension of the MPQ and the relationship of affect to receipt of worker’s compensation. The design and methodology of the study are explained in Chapter III. The results of the statistical analyses performed in the study are presented in Chapter IV. Chapter V contains a summary Of the findings, conclusions drawn from those findings, implications for research and practice, and the researcher’s reflections. CHAPTER II LITERATURE REVIEW Introduction This chapter contains a review of literature and research related to the physiology of chronic pain, how pain signals are transmitted, and theories regarding pain. Also discussed are the results of multidisciplinary treatment programs for low back pain, prediction of treatment outcomes of multidisciplinary pain clinics (MPCs), An overview of pain measurement with an in-depth review of the McGill Pain Questionnaire is provided, followed by a section on the Affective dimension Of the McGill Pain Questionnaire (MPQ) and the relationship of affect to receipt of worker’s compensation. Overview of the Concept of Pain This section lays a foundation for understanding the concept of pain. Pain is classified in four categories from the medical perspective. The first category is acute. Acute pain is pain that lasts less than 4 weeks. There is Objective evidence of nociception or a pain generator. Recurrent pain is the second category. Recurrent pain is less than 4 weeks’ duration; thus, it is acute in nature, but this acute condition returns at least every 6 months. The interlude is pain free and lasts fewer than 6 months. 16 Potentially chronic is the third category. Pain in thi: category lasts more than 4 weeks and less than 6 months. The fourth category of pain is chronic pain. This pain is 01 at least 6 months’ duration. There are three types of chronic pain. The first is from a known cause. There is Objective evidence tI serve as the basis for the pain sensation. A cure is not possible. The second type is from an unknown cause. The diagnosis may be inadequate and/or the treatment is inadequate. If the physician has objective evidence for the basis Of the pain, it is curable i1 diagnosed and treated properly. The third type of chronic pain is where no objective evidence for the basis of pain exists. The next section contains a description Of how pain signals are processed. This section is included to describe the physiology 01 pain. The Physiology of Chronic Pain The physiology of the transmission Of pain signals is discussec in this section. The way in which pain is induced, how pain is received, and how pain signals are transmitted to the brain are examined. Tissue Damage Tissue damage is one cause Of pain. A person generally begin: to feel pain at an average critical temperature Of 42.5 degree: Centigrade; 45.5 degrees Centigrade is the temperature at whicl tissue begins to be damaged if it is exposed to radiant heat. I‘ the temperature remains at this level indefinitely, the tissues arI eventually destroyed. Pain resulting from heat is closely correlated with tissue damage. In a study of soldiers who had been severely wounded during battle, Beecher (T956) found that most of them felt little or no pain except a short time after sustaining the wound. This was a function of stress. Pain was felt later. Pain Receptors Pain receptors in the skin and other tissues are all free nerve endings; they are widespread in the superficial layers of the skin. Pain receptors also are found in certain internal tissues, such as the periosteum, the arterial walls, and the joint surfaces. Most of the deep tissues are not extensively supplied with nerve endings. However, any widespread tissue damage will cause the aching type of pain found in these areas. This is discussed later in the section entitled "The Pattern Theory of Pain." In contrast to most other sensory receptors in the body, pain receptors adapt either slowly or not at all (Guyton, T97l). Under some conditions, the threshold for excitation of the pain fibers becomes progressively lower as the pain stimulus continues, causing the receptors to become more activated with time. This increased sensitivity of pain receptors is called hyperalgesia. One can readily understand the importance of the failure of pain receptors to adapt in chronic pain syndromes. These receptors continue to keep the patient apprised of possible damaging stimuli long after that information has served a biologically useful purpose. 18 Individuals with chronic pain do not become accustomed to it; Instead, they become more "sensitive" and suffer more as time passe (Bonica, T973). This sensitivity tO pain eventually is reflected a anxiety, before gradually giving way to various degrees O depression as patients evaluate their pain and related medica problems. As patients focus more attention on their medica problems, the reasons for them, and possible treatments, they begi‘ to think and talk more about the pain and become morn hypochondriacal on examination. 'The constant attention to an selected sensitivity Of the painful area eventually lead to hyperalertness to all symptomatic cues (somatization). The Transmission of Pain Siqnals A series. Of' complex electrical and chemical events occur between the stimulus of tissue injury and the subjective experieno of pain. The anterolateral system is a major ascending system frOI the spinal cord to the brain that transmits pain impulses. ThI anterolateral system contains three major pathways. The pathway are named for their site of termination: the spinothalamic (O' neospinothalamic) tract, the spinoreticular (or paleospinothalamic tract, and the spinotectal tract. Within the anterolateral system, two pathways carry paiI impulses within the central nervous system (Haugen, T956). Keat and Lane (T963) stated that pain impulses are transmitted over th sets Of nerve fibers, both of which are present in all periphera nerves. One system, known as type-C fibers, is small anI 19 unmyelinated. The other system is A-Delta fibers, which are larger and thinly myelinated. They are discussed in greater detail below. A—Delta Pain Fibers A-Delta pain fibers are nociceptive neurons with small, thinly myelinated axons. Pain signals are transmitted by large A—Delta fibers at velocities Of 5 to 30 meters per second. They are activated most efficiently by strong mechanical pressure and extreme heat. Activation of these nociceptors is associated with the sensation of sharp, pricking pain. When A—Delta fibers are blocked by moderate compression of the nerve trunk, the pricking quality of pain disappears. The pricking-pain pathway terminates in the caudal portion of the ventral—basal complex, an area that is the human analog of the posterior nuclear group in lower animals. From this point, signals are transmitted into other areas of the thalamus and to the somatosensory cortex. Most of the fibers from this area go to somatic sensory area II, which is concerned with pain localization. Type—C Pain Fibers The terminal regions of C—fiber axons are unmyelinated and are smaller than A-Delta pain fibers. Type—C fibers transmit pain signals at velocities Of .5 to 2 meters per second. When type—C fibers are blocked by low concentrations of local anesthetic, the burning and aching quality Of pain disappears (Guyton, l97l). C-fibers are believed to be activated by a chemical released into the extracellular fluid as a result Of tissue damage (Kelly, l985). 20 The nerve endings are activated by high-intensity mechanical, chem“ cal, and thermal (greater than 45 degrees Centigrade) stimulatiOI Type-C fibers are called polymodal nociceptors. They are wide' distributed in deep tissues as well as skin. The burning/achiI pain fibers terminate in the reticular area Of the brain stem and ° the interlaminer nuclei Of the thalamus, the entire cortex, Timb' system, the frontal lobes, and the hypothalamus. Both the reticulz area of the brain stem and the interlaminer nuclei are part of ti reticular activating system, whose function is tO transmit activii signals to all parts of the brain. Dppble Pain Siqnals The sudden onset Of a pain stimulus gives a "double" pai sensation--a physiological acute pain sensation followed a second c so later by a physiological chronic pain sensation. The prickir pain rapidly alerts the person to the possibility of tissue damag and facilitates the person’s reaction to withdraw from the harme stimulus. The burning/aching sensation tends tO become increasingl distressful over time. It is the slowly rising C-fiber stimulatior combined with the patient’s affective and evaluative responses, tha gives rise to chronic pain. In addition to local tissue reactions, reactions at th segmental levels of integration are manifested by a spasm c skeletal and smooth muscles and by glandular hyperactivity (Bonica T974). Reactions at the suprasegmental levels are manifested I: more highly integrated, but still automatic, protective and adaptiv 21 patterns involving primarily respiration and circulation Ultimately, reactions at the highest integrative levels involve th cerebral cortex. These slowly rising pain patterns are susceptibl to central control (Melzack & Wall, T970). Specific Pain Pathway; Two types Of second-order neurons transmit information abou pain in the spinal cord. One type is called the relay cell, whic projects to the brain stem or thalamus. The other is called th interneuron, which transfers information about pain to othe interneurons or to relay cells (see Figure 2.l). The second-orde neurons receive A-Delta and C-fiber inputs from interneurons in th substantia gelatinosa (lamina II). Relay cells for pain are located in two regions of the dorsa horn. Their axons ascend in the anterolateral quadrant Of the whit matter. The axons Of relay cells in Rexed’s lamina I and II projec to the neospinothalamic tract. The pain-projection pathways ascend in the anterolatera portion of the lateral column; hence, they are collectively calle the anterolateral system. The anterolateral system is primaril crossed in humans; this is especially true for the neospinothalami component. A small but significant ipsilateral component exists These uncrossed fibers may be the reason pain returns in som patients despite an initially successful surgical section 0 anterolateral fibers (Kelly, T985). 22 Cerebral cortex: Superior Irontal Somatic sensory areas (S-l. S-Ill Posterior parietal cortex Thalamus: Intralaminar nuclei .. L . . Posterior nuclear group / Tectum Periaqueductal gray matter Mesencephalon Pons Reticular Iormation Medulla lpsilateral tract A5 and C “be" Anterolateral system: Neospinothalamic Spinal cord Paleospinothalamic Figure 2.1: The anterolateral system of spinothalamic, spino- reticular, and Spinotectal fibers conveys informa— tion about pain to several regions of the brain stem and diencephalon. (From Kandel & Schwartz, T985.) 23 There are two general populations of spinal cord nociceptive neurons (See Figure 2.2). They are neurons in lamina I and II and neurons in lamina V. Figure 2.2: Neurons afferents. tive neurons. Dorsal column I 1' [I’E'Smau (Ab or C) libers /‘o '1 x ,- - -‘ ‘Dorsal root ganglion ~ rMargInal zone } SubstantIa gelaIInosa T t \ ~47 Spinal cord l I f / Paleosmnothalannc tract Neospgnothalarnuc tract Schematic drawing of the dorsal horn of the spinal cord, illustrating that the nociceptive neurons, whose axons form the ascending anterolateral system, are found in lamina I and lamina V of the dorsal horn. (From Kandel & Schwartz, T985 ) in lamina I are selectively activated by pain This group is made up Of ascending dorsal horn nocicep- Their axons form the neospinothalamic component of the anterolateral system. This group may be responsible for the localization of sharp or acute pain on the body surface (Kelly, 24 1985). This type of pain is most Often studied in the laboratory but does not normally prompt patients to seek medical attention. Neurons in lamina V receive input from mechanoreceptors, thermoreceptors, and pain receptors. The second group Of ascending dorsal horn nociceptive neurons are located primarily in lamina V. These large cells respond to both nonnoxious and noxious stimuli. They are called multireceptive or: wide—dynamic-range nociceptors. The responses tO each of the fiber groups differ. Threshold touch stimuli produce a brief burst Of firing in lamina V cells, followed by a brief inhibition. When large-fiber touch input is suppressed, there is a build-up in the activity of the same cell following C-fiber stimulation, leading to prolonged afterdischarges and facilitation. The cell axons contribute to the paleospino— thalamic and spinoreticular components Of the anterolateral system. This seems to promote the more diffuse, chronic types Of pain that are of primary importance in the practice of medicine (Kelly, l985). In summary, pain receptors are free nerve endings and are stimulated while tissue damage is occurring. Pain receptors adapt very slowly, and hypersensitivity to pain develops. Slow-moving C-fibers transmit chronic pain to essentially all areas of the cerebral cortex and reticular activating system. This causes a general activation of the entire nervous system, which leads to fatigue, anxiety, and hyperactivity of the skeletal muscles. 25 Theories of Pain The physiology of the transmission of pain signals was reviewed in the preceding section. This section contains a review Of theories on pain and how they have changed over the years. This progress will demonstrate how ‘theorists went from a traditional theory of pain to the more recent and accepted behavioral and cognitIVe-behavioral perspective (if pain perception. It is important to recognize the foundation on which current research on pain perception is based. The specificity theory of pain and the pattern theory of pain are discussed and evaluated in this section. The gate control theory Of pain by Melzack and Wall also is considered. Fordyce’s respondent and Operant pain model, a behavioral method for treating chronic pain, is reviewed, as well. The Specificity Theory Of Pain Specificity theory is the traditional theory Of pain and was widely accepted during the first half (Hi the twentieth century. Specificity theory originated irI a classical description by Descartes in T664 (Melzack, T982). He conceived the pain system as a straight—through channel from the skin in) the brain. Descartes suggested that the system is like the bell—ringing mechanism in a church: Someone pulls a rope at the bottom of the tower, and the bell rings in the belfry. He proposed that if a person’s foot came too close to a fire, the sensation would be transmitted up the leg and back, into the head, where presumably something like an alarm 26 would be set Off. The person would feel pain and respond to it. Descartes’ concept of the specificity theory proposed that the brain is aware of the outside world only through messages conveyed to it by sensory nerves. In l842, Johannes Muller contributed to an understanding of the sensory process, theorizing that the brain receives information about external objects only through the sensory nerves (Melzack, T982). Activities in nerves represent coded or symbolic data concerning the stimulus Object. Muller’s theory led to the search for a terminal center in the brain for each of the sensory nerves. Muller conceived a straight-through system from the sensory organ to the brain center responsible for sensation. He believed that the cortex was at the "top" Of the nervous system, and a search was undertaken for the cordal centers. Visual and auditory projections to the cortex were found very early. Muller assumed that these cortical areas were the seat of sight and hearing. Between T894 and T895, Max von Frey published a series of articles in which he presented a theory of cutaneous senses (Melzack, T982). His theory was expanded during the next 50 years, forming the basis of modern-day specificity theory. Von Frey’s theory designated free nerve endings as pain receptors. He combined three kinds Of information to form this theory. The first was Muller’s doctrine of specific nerve energy. Von Frey believed that Muller’s notion of a single sense of touch or feeling was inadequate, so he expanded this concept to include four major cutaneous modalities: touch, warmth, cold, and pain. Each .1. 27 presumably was projected through its own special system to a brain center responsible for the appropriate sensation. The second kind of information von Frey used was the distribution of sensitivity to warmth and cold at the skin. He believed the skin comprised a mosaic of four types of sensory spots: touch, cold, warmth, and pain. A third kind Of information von Frey used pertained to the fine structure Of body tissues. In the nineteenth century, anatomists used particular chemicals to stain thin slices of tissue from various parts of the body and then observed the tissues through a microscope. They found two types Of specialized structures: (a) the free nerve endings that branch out into the upper layers Of the skin and (b) nerve fibers wrapped around hair follicles. Von Frey reasoned that, because the free nerve endings are most commonly found and the pain spots are found almost everywhere, free nerve endings are pain receptors. Meissner corpuscles are found in the fingers and palm of the hand, where touch spots are most abundant and most sensitive. These and the fibers surrounding hair follicles are touch receptors. Von Frey noted that the conjunctivum of the eye and the tip of the penis are both sensitive to cold, but the conjunctivum is not sensitive to warmth and the penis is not sensitive to pressure. Because Kraus end-bulbs are found in both places, Von Frey concluded that Kraus end-bulbs are cold receptors. Hence, the major sensation——warmth--and one major receptor, Ruffini 28 end-organs-—were left over. Thus, von Frey concluded that Ruffini end-organs were warmth receptors. Von Frey’s theory dealt only with receptors. Later researchers considered specific fibers from the receptors to the spinal cord and then specific pathways in the spinal cord itself. Bishop (T946), Rose and Mountcastle (T959), and Sinclair (T967) demonstrated that there is a one-to-one relationship among receptor type, fiber size, and quality Of experience. The fiber—diameter groups are held to be modality specific. The theory indicates "specific nerve energy" on the basis of fiber size, so specificity theorists speak of A—Delta fiber pain and C—fiber pain Of touch fibers and cold fibers as though each fiber group had a straight-through transmission path to a specific brain center. The location of the pain center has been a source Of debate among specificity theorists. Head (T920) proposed that it is located in the thalamus because cortical lesions or excisions rarely abolish pain. Therefore, theorists believed that the thalamus contained the pain center and that the cortex exerted inhibitory control over it. Today, specificity theory defines pain as a primary sensation with special peripheral receptors, neuronal transmitters, and receivers in the central nervous system. The pain pathway is seen as an uninterrupted transmission system. The intensity of pain a person perceives is in direct proportion to the intensity Of the stimulus that is applied. Peripheral pain receptors are believed to be specific receptors that are distinct from the other main group Of 29 sensory receptors——mechanoreceptors and thermoreceptors. Specifu ity theory maintains that free nerve endings are the pain receptor: These nerve endings generate impulses carried by specific ner\ fibers (pain fibers), A-Delta and C-fibers, in peripheral nerves These fibers carry the impulse of pain primarily to the later; spinal thalamic track in the anterolateral part of the spinal corc The spinal thalamic track transmits the impulses to the pain cents in the thalamus. Specificity theory does not account for the discrepancy the exists between the intensity Of noxious stimuli and the subjectit perception of pain. Some researchers have believed that tt receptors excited by intense, noxious stimulation are specialized t respond to those particular kinds of stimuli. It is questionabl whether stimulation Of the receptor must always elicit only th sensation of pain (Melzack & Wall, T970). Possibly, a fe specialized fibers respond only to intense stimulation. This doe not mean they exist solely to produce pain when stimulated. Certai spinal cord tracks and central nervous system pathways carry mos impulses related to pain. It does not follow that they wiT constitute a specific pain system. Melzack and Wall (T965) place the specificity theory in the following perspective: Physiologic specialization is a fact that can be recognize without acceptance of the psychologic assumption that pain i determined entirely by impulses in a straight-throug transmission system from the skin to a pain center in th brain. (p. 972) 30 Pattern Theor Pattern theory is a general heading for the theories that were proposed as a reaction to specificity theory. Goldschneider, once a strong supporter Of von Frey’s theory, proposed in l894 that stimulus intensity and central summation are the critical determinants of pain (Melzack, T983). Studies of pathological pain influenced Goldschneider in developing his pattern, or summation, theory of pain. Goldschneider believed that mechanisms of central summation, most likely in the dorsal horns of the spinal column, were essential to consider in understanding pain mechanisms. He proposed that particular patterns of nerve impulses that evoke pain are produced by the summation Of the skin’s sensory input at dorsal horn cells. Pain results when the total output of cells exceeds a critical level. This occurs as a result of the excessive stimulation Of receptors of pathological conditions that enhance the summation of impulses produced by normally nonnoxious stimuli. Goldschneider assumed that a spinal summation path transmits pain signals to the brain. This path consists Of slowly conducting, multisynaptic fiber chains. He presumed that the large fibers projecting up the dorsal column pathways carry specific information about the tactile discriminant properties of cutaneous sensation. Three theories emerged from Goldschneider’s concept. The peripheral pattern theory deals primarily with peripheral as opposed to central patterning. Pain results from excessive peripheral 31 stimulation, which produces a pattern of impulses interpreted as pain. Waddell (T955) and Sinclair (T955) proposed a pattern theory suggesting that cutaneous qualities emanate from spatial and temporal patterns of nerve impulses as opposed to separate modalities’ specific transmission. The peripheral pattern theory proposes that all fiber endings are alike, so the pattern for pain is produced by intense stimulation of nonspecific receptors. Whereas this pattern ignores the possibility of physiological specialization, it does reveal a high degree of receptor-fiber specialization. The central summation theory proposed by Livingston (l943) suggests that a specific central neural mechanism accounts for the summation phenomena in phantom limb pain, causalgia, and neuralgias. Livingston proposed that pathological stimulation of sensory nerves initiates activity in closed, self-exciting loops Of neurons in the gray matter Of the spinal cord. Normally nonnoxious inputs trigger this activity to generate volleys Of nerve impulses interpreted as pain. Livingston’s theory is useful in explaining phantom limb pain. He suggested that the trauma associated with removal of a limb initiates abnormal firing patterns in reverberatory circuits in the dorsal horns of the spinal cord. This action sends volleys of nerve impulses to the brain, causing pain. Livingston proposed that the reverberatory activity may spread to adjacent neurons in the lateral and ventral horns to produce autonomic and muscular manifestations 32 in the limb. These actions then produce further sensory inp creating a vicious circle between the central and periphe processes to maintain abnormal spinal cord activity. Mh irritations near the site of the trauma then feed into active p0 of neurons to maintain an abnormal disturbed state. Impu patterns normally interpreted as touch are then perceived to trig neuron pools into a significantly higher level of activity, wh' sends volleys or impulses to the brain to produce pain. Brz activity registering emotional disturbance feeds into the abnorn neuron pool by evoking neural activity. Gerard (T95l) proposed that a peripheral neuron lesion n bring about temporary loss of sensory control of firing in t spinal cord neurons, which may fire in synchrony due to the spre of electrical fields. This combined firing of neuron pools cou recruit additional units and move along in the gray matter. I firing is maintained by impulses different from and Of Te intensity than those needed to initiate it. The firing of t neurons discharges excessive and abnormally patterned volleys to t higher central cord activity. Livingston’s and Gerard’s concepts are powerful in explaini phantom limb pain. However, they fail to account for the fact th surgical lesions of the spinal cord do not always abolish pain. The sensory interaction theory proposes that a specializ input-controlling system normally prevents summation from occurrin The destruction of this system allows pathological pain states 33 develop. According to this theory, there is a rapidly conducti fiber system that inhibits synaptic transmission in a more slow conducting system that carries the signals for pain. The t systems are described as myelinated and unmyelinated fiber system Noordenbos (T959) proposed that the small fibers carry nerve-impul patterns that produce pain, whereas large fibers (A-Beta) inhib the transmission of pain. A proportional increase in the number small fibers in relation to large fibers could increase neur transmission summation and pathological pain. There are diffus extensive connections within the ascending multisynaptic system, contrast to the idea of a straight-through system. A change in the fibers in favor of the small fibers consistent with the Observed relative loss of large fibers aft peripheral nerve injury. This may explain delays, temporal a spatial summation, and other properties of pathological pain. Pattern theories of pain were developed in response to t deficits Of the specificity theory. Pattern theories claim th large cutaneous fibers comprise a specific touch system. T smaller fibers converge on dorsal horn cells, which summate the input and transmit a pattern to the brain, where it is perceived pain. Pattern theories suggest that intense noxious stimuli at t periphery set up a vicious circle that maintains abnormal spin cord activity to generate abnormal volleys Of stimulath consciously interpreted as pain. A specialized input—controllh system normally prevents stimulus summation from occurring Destruction Of this system leads to chronic pain states. L—— ‘l-r: \r‘ 34 Pattern theories propose the existence of a rapidly conducting fiber system that inhibits synaptic transmission in the more slowly conducting system carrying the signal for pain (Zotterman, T939). The two systems Of pain transmission have been identified as epicritic and protopathic (Head, T920), fast and slow (Lewis, T942), and myelinated and unmyelinated (Noordenbos, T959) fiber systems. Proponents of the pattern theories believe that, under pathological conditions, the slow system establishes dominance over the fast; the result is diffuse, burning, chronic pain. The various pattern theories have failed to provide a satisfactory general theory Of pain. They Tack unity and have not been integrated into a single theory using the various theoretical mechanisms. Summary of the Specificity and Pattern Theories Of Pain The specificity and pattern theories of pain both include valuable concepts that supplement one another. The first concept is recognition of receptor specialization for the transmission Of particular kinds and ranges Of cutaneous stimulation. The information generated by peripheral receptors is coded in the form of patterns of nerve impulses. The specific physiological assumption that attributes the painful experience to activity of one type Of receptor, fiber or spinal pathway, is the basis for both theories. l—————“— 35 There is evidence that a higher central nervous system activity or cognitive function influences pain. Anticipation of pain, anxiety and attention (Hill, T952), cultural background (Chapman, Finesinger, Jones, & Cobb, l947), early experience (Melzack, T973), and prior conditioning (Pavlov, T927) all have a profound effect on pain experience and response. The previous assumption that pain was a primary sensation relegated affective and evaluative processes to the role Of "reactions to pain" and made them secondary considera- tions in the whole pain process. Both the specificity and pattern theories assume that the affective and cognitive process must follow the primary noxious sensation. However, this assumption fails to account for basic data. Beecher’s (T946, T959) observation that American soldiers wounded during World War II "either denied pain from their extensive wounds, or had so little that they did not want any medication to relieve it" implies that central control of painful stimuli can be elicited. The previous traditional views of the pain mechanism failed to account for two psychophysiological processes that result in pain. They were exploring the physiological basis for pain. The first is a sensory discriminative process, in which stimuli are localized in space and time along an intensity continuum. The second is the cognitive affective component, which motivates the organism to stop the pain as quickly as possible. Sherrington (T906) noted that pain comprises both sensory and affective dimensions. He proposed that the "mind rarely, probably never, perceives any object with absolute 36 indifference, that is, without ’feelings.’ . . . Affective tone is an attribute of all sensation, and among the attribute tones of skin sensation is skin pain." The Gate Control Theory The gate control theory Of pain proposed by Melzack and Wall (T965) provides the basis for considering the cognitive and affective dimensions of pain, in addition to the more Obvious sensory dimensions described by the specificity and pattern theories. In T965, Ronald Melzack, a psychologist, and Patrick Wall, a physiologist, propounded a theory incorporating substantial portions Of the specificity theory and the pattern theory Of pain in an attempt to account more fully for evidence regarding pain mechanisms. The researchers attempted to integrate what they believed were the requirements Of accountability into a comprehensive theory Of pain. The differences between Melzack and Wall’s (T965) theory and previous theories were significant. They proposed that there is an explicit mechanism for the inhibition of the slowly conducting nociceptive system by the fast-conducting one. They also proposed that the descending controls from the brain could moderate the passage of nociceptive signals. In T982, Melzack and Wall revised their theory to address the explicit mechanism on physiological grounds. The theory may be an oversimplification regarding the physiological mechanism (Britton & Skevington, T989), but it provides the most useful starting point to date. 1 37 According to the gate control theory, the perception Of or reaction to a pain-producing stimulus applied to the skin is the result of an interplay among three systems within the spinal cord: (a) cells of 'the substantia gelatinosa (SG) (lamina II) in the dorsal horn of 1the spinal cord, (b) central transmission cells (T cells) in the dorsal horn, and (c) afferent fibers in the dorsal column of the cord. The hypothesis of the neural mechanisms (Melzack & Wall, T965) underlying pain sensation states that the interaction between myelinated and unmyelinated inputs to the spinal cord occurs at two sites. These sites are the inhibitory interneurons in the SG (lamina II) cells and dorsal horn pain-transmission neurons. Both myelinated and unmyelinated primary afferents were proposed to have a direct excitatory action on the T cells. The SG neurons inhibit transmitter release from both classes Of primary afferents, in this way presynaptically inhibiting all afferent input to the pain- transmission cells. The myelinated afferents excite the inhibitory SG neuron, reducing input to the T cell and thus inhibiting pain. Fields (T987) supported this hypothesis by making the clinical observation that selective stimulation of large-diameter myelinated fibers produces analgesia. Melzack and Wall also proposed that activity in unmyelinated nociceptors inhibits the inhibitory 50 cells, resulting in the enhancement Of transmission from primary afferents to the T cells, consequently increasing pain intensity. This allows the 38 unmyelinated afferents to have two distinct excitatory effects on dorsal horn pain-transmission cells: first, a direct synaptic excitation and second, an indirect excitation resulting from the inhibition of the inhibitory SG cell. The gate control theory emphasizes that the perceived intensity of pain is the result Of a balance Of input from myelinated and unmyelinated fibers. Normally, stimuli activate both types of afferents, and the sensation produced by brief noxious stimuli is of rapid onset and short duration. Here, both the intensity and duration Of the pain are reduced by the inhibitory input from the myelinated afferents. Myelinated afferents have a direct excitatory and an indirect inhibitory effect on the pain-transmission cell (see Figure 2.3). Myelinated afferents have the potential either to produce or to inhibit pain. Activity in non—nociceptive myelinated afferents produces pain when the inhibitory SG neuron is maximally inhibited by activity in unmyelinated nociceptors. The indirect inhibitory effect of the myelinated fibers is blocked, and their direct excita- tory action on the T cells would predominate. These three categories Of activity are believed to interact with one another to provide perceptual and cognitive information regarding the pain and the motivational tendency toward the pain. The afferent fibers in the dorsal column of the spinal cord are involved in a system of interplay within the spinal cord. Signals from the pain-producing stimuli reaching the spinal cord apparently are transmitted to both the SG cells and the afferent fibers. Pain 39 signals reaching these afferent fibers bypass the gating mechanism (the SG cells) and are transmitted directly to the central controls. Here they activate selective brain processes, resulting in the transmission Of descending signals to the spinal gate (the SG cells). These return impulses influence the modulating properties of the spinal gate (SG cells). They may have either an excitatory or an inhibitory effect. These descending messages account for the influence of numerous variables affecting the pain experience. These variables are attention, memory, conditioning, the meaning of pain, and anxiety. M U Afferent I Cell T Cell T-Cell Output Input Effect Effect (T4) M + + O U — + ++ M + U 0 + + Figure 2.3: Effects of myelinated afferents on the pain— transmission cell. (From Fields, 1987.) 40 The gate control theory accounts for the effect Of cognitiv and affective factors in the reaction to pain by incorporating the into the sensory perception Of pain. Melzack and Casey (T968 developed a Iconceptual model Of the sensory, motivational, an central control determinants Of pain. They proposed that the outpu of the T cells of the gate control system projects to the sensory discriminate system and the motivational-affective system. 'Th central control trigger is represented by a large-fiber system b the central control processes and then back to the gate contro system, as well as to the sensory-discriminate and motivationa affective system. This way, all three systems would interact wit each other, as well as project to the motor system (see Figure 2.4) The motivational-affective dimension is based on th phylogenetically Older spinO-reticular and paleO-spino-thalami components Of the anterolateral somatosensory pathway, sendin projections to the brain stem reticular formation and the Timbi system. These tracts respond to stimuli, fear and anxiety. The cognitive-evaluative dimension of pain depends on th cortical process. There is a central control system that act rapidly in identifying, evaluating, and selectively modifying th sensory input. This accounts for the influence Of past experience sex, attention, and evaluation Of the input as threatening or not 0 the sensory-discriminate aunt motivational-affective systems. Thi rapidly works to allow analysis of the input, comparing it wit other input in memory, and bringing it into action-reSpons strategies. 41 Descending COQ‘T‘l‘Ye Inhibitory control COnIrOi 0 7 V X X4 if SG InthIIOry v ‘/ / SG excitatory Figure 2.4: The gate control theory of pain. Melzack & Wall, T982.) Action system (Adapted from 42 Melzack (T973) theorized that the cognitive-evaluative system is served by the dorsal column and the dorsal—lateral pathways. This system has an inhibitory influence on the dorsal horn cells in the spinal cord. It is believed to have the capacity to modulate sensory input before it is transmitted to the sensory—discriminative and motivational—affective systems. This component of the gate control theory is important. The cognitive-evaluative system directly influences the motivational-affective system, whereby the frontal cortex plays an important role, interacting with both the cortical area and the reticular and limbic structures. Cognitive, affective, and evaluative input can alter pain perceptions. Stimulation of the brain activates descending afferent fibers to influence afferent conduction at the earliest synaptic levels of the somethetic system. It is possible for conscious central nervous system activity such as attention and distraction, anxiety and conditioning, tO exert control over sensory input. These central influences are thought to be mediated through the gate control system (Wall, T967). It is important to recognize the role of higher central nervous system activity, such as attention, distraction, anxiety, and conditioning, in the pain process. The degree of conscious control Of pain is determined by the extent and duration of the input pattern. Some Of the most unbearable pain rises so rapidly in intensity that the patient is unable to achieve any control over it. Conversely, more slowly rising chronic pain is susceptible to central control; thus, the patient may be able to keep his/her pain under control. 43 Melzack; and Casey (T968) pointed out ‘that it might be appropriate to direct treatment efforts toward the neglected contribution of affective and cognitive processes as Opposed to surgical remedies for pain. They believed that pain can be treated, not only by trying to cut down on sensory input by anesthetic block or surgical intervention, but also by influencing the affective and cognitive dimensional factors. The strength of the gate control theory is that it offers an explanation for puzzling and paradoxical clinical findings (Fields, T987). Britton and Skevington (l989) used a mathematical model of the gate control theory Of pain tO account for acute pain in humans. They developed three possible explanations for chronic pain, based on an evaluation of the gate control theory. It might be that chronic pain is associated with plastic changes in the nervous system. It inight also be that psychological factors result in cognitive control being more excitatory (or less inhibitory) than would otherwise be the case. Finally, one input into the control system-~one value for the firing frequencies in each Of the large and small fibers--could result in more than one possible output from the 'T cells, depending (HT the history of ‘the injury (Britton & Skevington, T989). Behavioral Methods for Treatinq Chronic Pain Fordyce played a pioneering role in investigating the psychology Of chronic pain. He helped identify the importance of behavioral factors in patients with chronic pain. He then developed 44 strategies for diagnosing and treating chronic pain. Fordyce (1968) described the use Of behavior—management techniques for problems associated with treating chronic pain. Following the publication Of his papers (Fordyce, Fowler, & de Lateur, T968; Fordyce, Fowler, Lehmann, & de Lateur, T968b), the use of behavioral methods to treat chronic pain increased rapidly. Fordyce and his colleagues (Fordyce, T976; Fordyce, Fowler, & de Lateur, T968; Fordyce, Fowler, Lehmann, & de Lateur, T968a, l968b; Fordyce & Steter, T978) refined behavioral learning theory (operant and classical conditioning) for the chronic pain population. Behavioral methods, broadly defined, are now used in virtually every legitimate treatment program in the United States. Traditional medical care responds to the perception Of chronic pain mainly from an acute-care model. This model, however, fails to address special characteristics Of chronic pain. A discipline known as behavioral medicine has developed over the past TS to 20 years, during which time there has been a significant increase in psychologically or behaviorally based treatment of pain. Loeser (T982) categorized the phenomena Of pain in the human being into four domains: nociception, pain, suffering, and pain behavior) (see Figure 2.5). 45 Figure 2.5: Schematic diagram for the components Of pain (only pain behavior is measurable and observable). (From Loeser, 1982.) The first three domains—-nociception, pain, and suffering-—are personal, private, internal conditions that can only be inferred to exist. They are unmeasurable in a clinical setting. Nociception connects pain with thermal, mechanical, or chemical stimuli that damage or threaten tissue. Nociception is a common symptom of disease, warning the individual that something is wrong and signaling the need for protection. Nociception is the detection of tissue damage by transducers in skin (Loeser & Egan, T989); deeper structures receive this information by means of A—Delta and C—fibers in the peripheral nerves. Pain is the recognition of nociceptive stimulation by the central nervous system. Suffering is the negative affective response to pain or other emotionally latent events, such as fear, anxiety, isolation, and depression. Pain behavior is what a person does or does not do or 46 say that leads the Observer to believe the individual is suffering from noxious stimuli. All pain behaviors are measurable and observable. Excessive disability-related behavior may characterize chronic pain--pain that has been in existence for 6 months or more. Fordyce recognized that there is a difference between chronic and acute pain. In many instances of chronic pain, the noxious stimulus is not observed or known. Previous medical examinations may have implied, without direct confirmation, that there 'was an organic basis for the complaints. Chronic pain also exists when the indi- vidual has little basis for estimating when the pain and associated disruption of life will cease. Chronicity exists when learning or conditioning effects become established and add to the complication of understanding the problem. In this circumstance, chronic pain represents a number of problems involving a complex interaction Of organic/physical, environmental, and psychological factors, princi- pally those Of learning or conditioning. Fordyce described pain problems by making a functional distinction between respondent pain and Operant pain. Respondents are actions of an organism that are reactive in nature (Fordyce, 1981). In the presence Of an adequate stimulus, the response will automatically follow. Antecedent stimuli control respondents and involve the body’s glandular or smooth-muscle functions. Operants are responses that are under voluntary control; they also may be elicited by antecedent stimuli but are influenced by the consequences Of the response. In this manner, Operants come under 47 the control of those consequences. Operants involve primarily striated or voluntary muscles controlling such functions as speech, bodily movements, and facial expressions. Fordyce distinguished between problems of respondent pain and operant pain in determining the appropriate approach to the problem of chronic pain. The expression Of pain is influenced or Controlled to a significant degree by environmental consequences; therefore, it should be viewed as operant pain. The role Of these environmental consequences must be considered when seeking a solution to the problem. In contrast, expressions of pain that are influenced cw“ controlled by an antecedent environmental stimulus reflect respondent pain. The behaviorist can classify ii patient’s pain as primarily respondent or primarily operant in nature. This classification is made only after carefully assessing information relating to body damage and the environmental consequences of the patient’s expression of pain. Fordyce’s use Of the terms "respondent pain" and "Operant pain" does not address the issue of real or imaginary pain. From this perspective, the proper question is not whether the pain is real, but rather what factors influence it. The pain is perceived as real. The Respondent Pain Model Chronic pain traditionally has been viewed from the perspective Of a disease or medical model. In the diagnostic process, it has been assumed that the indications (Hi pain occur' as a result of 48 underlying body damage. The medical model also assumes that, if body-damage factors were eliminated, the pain would cease. This conceptual model works quite effectively with most problems Of acute pain, but its efficacy diminishes as time passes and the pain becomes more chronic. Chronic pain frequently is described as at deep, dull aching that persists for hours, days, weeks, or even months and increases with activity. Treatment from it respondent—pain perspective has limited Options. Treatment Of chronic pain with the medical/disease model Often fails. This is especially true when plastic changes in the central nervous system take place. The physician who is committed to the medical/disease model has few Options. Traditionally, most physicians begin treatment with a course Of conservative management techniques, which may include heat, massage, traction, or a variation of these modalities. Further diagnostic work may be the next step for patients who do not Obtain relief. If the diagnostic evaluations demonstrate positive results, surgical intervention may be the next step. If the low back pain continues, repeated diagnostic procedures may' be the next step. If' the diagnostic evaluations demonstrate positive results, surgical intervention may be repeated. If the low back pain continues, further treatment from a conservative perspective may occur. If the patient continues to complain Of pain, the treating physician may refer him or her to another physician, who is likely to apply the same conceptual model. 49 Unless the original physician erred or the second one has special skills or methods unknown to the first, it is unlikely the referral will lead to a cure. The second physician may follow the course Of treatment begun by the original treating physician. After other therapies have proven unsuccessful, the physician’s next Option is to tell the patient the pain is not real. Physicians often respond to the implication that the pain is "all in the patient’s mind" by labeling the pain psychogenic. This implies that a personality problem is responsible for the pain symptoms. At this point, the patient typically rejects the referral for psychotherapy or reluctantly accepts it. Even if psychotherapy is accepted and the patient participates fully, such therapy rarely is successful in resolving the pain problem (Fordyce, 1981). The most common response by physicians who perceive chronic pain from the medical/disease model and whose traditional treatment methods have failed is to tell the patient that nothing more can be done. The chronic pain patient then is given .a renewable prescription for medication, told his or her physical limitations, given recommendations for rest and exercise, and finally advised to learn to live with the pain. Although the treating physicians’ intention is to curtail patients’ endless pursuit of further diagnostic and treatment interventions, this response often has the opposite effect. The patient thinks the authenticity of his or her pain has been challenged. TO prove that the pain is real, the patient may actually increase rather than decrease involvement with health care providers. 50 The traditional or respondent approach to chronic pain occurs as a result of' physicians’ inability to recommend alternatives. Another way to deal with chronic pain is to address pain behaviors rather than simply treating the underlying pain. This approach has been called the Operant pain model, in contrast to the respondent pain model that traditionally has been used. The Operant Pain Model Operant behaviors are similar to respondent behaviors in that specific stimuli can produce them. However, operant behaviors are different because they are sensitive to the influence of conditioning. Conditioning relates to factors occurring during as well as after presentation Of the stimulus. Operant behaviors are sensitive to their consequences. Accordingly, Operants will increase or decrease in relationship to the likelihood Of recurrence when followed by a reward or punishment. Operant pain may increase in frequency and magnitude following positive consequences IN“ a reduction iri neutral or negative consequences. Respondent pain may quickly develop into operant pain if the pain behaviors consistently occur in an environment that promotes conditioning effects. The use Of respondent pain avoids the issue of whether a patient’s pain is "real" or "imaginary" (Fordyce, 1981). Society perceives this pain-related behavior as pain. ‘This operant pain, or the pain-related behavior, allows a previously healthy person to adopt the image Of infirmity (Fey & Fordyce, T983; Fordyce, 1981). Pain signifies distress and 51 suffering, and "sick" individuals frequentlyi receive sympathy and attention. Because they are sick, these individuals are not expected to carry out the duties they normally would be required to perform. Being insulated from stress and important decisions is one of the benefits sick people receive. The person with chronic pain is not expected to attempt to reverse the sickness or illness by pursuing some form of health care. Fordyce (T981, 1983) used the marital situation as an example Of how pain behavior can bring about subtle yet powerful changes in the partner’s actions. The chronic pain patient receives special sympathetic deference, and uncomfortable discussions CH“ arguments may be discouraged. The chronically ill person may avoid unpleasant tasks and use his or her pain to influence how the couple will spend their time or money or plan for the future. In most respondent pain problems, the above-mentioned conditions or reinforcers are not present long enough or strongly enough to distract the person from a normal life. However, in some cases the nature and extent of the injury, combined with the psychological atmosphere, are such that repeated pain behavior is consistently reinforced while l'well'l behaviors are ignored or extinguished. Pain behaviors slowly become Operant in nature (Fey & Fordyce, 1983); they may increase or decrease in response to the positive and negative consequences tO which they lead. The pain behaviors may become increasingly independent of the nociceptive stimuli that are 52 present. Early medical-management strategies, social contingencies, or psychological problems may allow 'the conditioning process to occur. This conditioning allows the chronic pain patient to be sick, reinforcing the insulation it provides from responsibilities. Pain behaviors may appear to occur independently Of physiological stimuli and may be under the contrOl of environmental stimuli. Treatment for this type (Hi pain problem involves examination and Immdification of the environmental contingencies. Continued medical-model-oriented treatment (”1 the underlying pathology' will fail to interrupt the pain behavior' or halt the pattern of Chronicity; in fact, it may systematically intensify the problem. The most Obvious and, at times, potent reinforcer of pain behavior is direct positive reinforcement (Fey & Fordyce, 1983). This reinforcement may be in the form Of attention, solicitousness, affection, and comforting behavior' frONI a spouse or significant other, as well as prescribed medications or rest that is contingent on the pain. The spouse’s pain-contingent behaviors are in response to overt displays Of pain and are evident only when pain behavior is present. Direct reinforcement also may come from the health care system itself. In treating the chronic pain patient, a physician may demonstrate professional concern, which may serve to reinforce the pain behavior, especially in patients who have few outlets for their pain complaints. 53 Chronic pain patients may have multiple physical and emotional complaints and Often receive a wide variety of medications from sympathetic physicians. Although some patients medicate themselves with alcohol, most. problems arise through the use of prescribed medications that are self-administered on a pain-contingency basis. Patients may dose themselves repeatedly during a 24-hour period, increasing their dosage as the narcotic or barbiturate habituation and tolerance increase. Notwithstanding the Obvious problems of drug dependency, addiction serves as a potent reinforcer of pain behavior. The addiction interferes with work, family, and social activities. In addition, addiction to prescription medications may exacerbate the patient’s preoccupation with his or her pain. Monetary compensation is another source Of direct reinforcement for pain behavior and disability. Insurance benefits in the form Of disability payments frequently equal or exceed what a person was earning while he or she was employed. These payments provide a significant disincentive for reducing sick or pain—related behaviors. Those pain patients with the least interesting jobs, the fewest employment Options, and the fewest personal and occupational skills are most likely to resist returning to work and to continue demonstrating pain behaviors. A disabled pain patient becomes trapped in a vicious circle, dealing with the potential termination of benefits or being forced tO return to work. The chronic pain patient continually! must prove that. the pain prObTeNI is real by 54 demonstrating appropriate pain behavior or reinstituting diagnostic workups and/or medical treatment. Rest is another direct positive reinforcement for continued pain behavior. Patients with low back pain and acute respondent pain may find that movement produces increased levels of pain and that rest reduces pain. Although rest frequently is prescribed in the acute phase of any injury, as pain persists in chronicity, rest may become a potent contingent reinforcer of pain, serving to promote the disability. Indirect reinforcement for pain may be subtle. Avoidance conditioning plays a role in the evolution Of the problem of operant pain and is particularly resistant to extinction. Pain serves to remove the person from a stressful or difficult situation. An lindividual faced with an aversive or noxious event may act to avoid it if at all possible. Pain behavior is a convenient way to avoid other direct, effective solutions for burdensome or physically demanding employment situations, jobs with low pay, or marriages that are psychologically stressful. Chronic pain patients learn that temporary solutions based on pain behaviors are easier to achieve than the difficult long-range solutions to many Of the problems they face. Avoidance responses are seen in chronic pain patients as compensatory body positions or devices such as crutches or braces. In the acute phase of the injury, walking or moving about produces pain and encourages the patient to favor the affected limb by limping. At times, the immediate reduction in sharp pain strongly 55 reinforces the limp or use Of the device. The avoidance response becomes a habit that is not easily broken, Because he or she anticipates the pain elicited by the stimuli, the patient may continue to limp or use the device. Bandura (1969) showed that modeling is a strong influence in learning complex behaviors and attitudes. Such learning is accomplished by observing the actions of others. Pain behavior may be learned through modeling or imitation. Children growing up with a chronic pain patient in the home will learn to imitate the pain responses they Observe. When the children see these pain responses reinforced at home or at school, they quickly learn tO display similar behavior when they are ill or injured. The purpose of treating chronic pain with the operant method in behavioral-medicine programs is to change the patient’s treatment and home environment, so as to develop and sustain alternative responses that are incompatible with sick behavior. The Operant method is based on the assumption that pain behaviors persisting past the predicted or anticipated healing time are under the control of conditioning effects. Minimizing social reinforcement encourages the extinction Of pain behaviors. The role Of the treatment staff with whom the chronic pain patient comes into contact is to minimize the reinforcement Of pain behaviors. Treatment staff and family encourage "well" behavior and systematically reinforce it. The Operant pain treatment program must address the individual needs of the chronic pain patient. Core aspects Of a program .I. 56 usually include medication reduction, physical reactivation through physical and occupational therapy, family counseling, treatment for depression, and vocational counseling. Any additional problems that limit functioning are addressed individually. After 17 years of achieving positive results through a behaviorally oriented treatment program for chronic pain, Fordyce (T985) reconfirmed the goals Of behavioral treatment. He believed that his critics have confused the hypothesis that pain behaviors can be learned and unlearned with the notion that pain itself can be learned and unlearned. Behaviorally oriented treatment programs for chronic pain reduce the disability associated with chronic pain problems. Although decreased pain is not a primary goal of rehabilitation programs using behavioral treatment, many patients have reported it lessening (If pain following application Of behavioral methods. Fordyce conceptualized behavioral treatment as therapy for excess disability instead of treatment for pain. His concept of excess disability highlights the major problem Of people with chronic pain. Those people who are more functionally disabled than necessary pose the problem. An Operant pain treatment program is appropriate for patients who are capable of functioning at higher levels with less disability, discomfort, and distress. The chronic pain patient’s potential for increased functioning may be overlooked or underestimated when the physician sees him or her only fer the presumed problem, which might be hysteria, poor motivation, degenerative disc disease, spinal stenosis, or pain Of 57 unknown origin. If the behavioral—medicine practitioner attends to what the patient does or does not do, learning technology can be used to increase, decrease, or maintain selected behaviors to bring about improvements in functioning. These improvements in function- ing may occur without changes in the underlying medical or psycho- logical problem (Fordyce, 1985). The reduction of excess disability is the primary purpose of behavioral pain-management programs. Although most patients increase their functioning and quality of life without reporting a decrease in pain, this is not a shortcoming of the treatment approach. The behavioral approach to the management Of chronic pain recognizes psychological factors that are not derived from psychopathology, but instead come from learning theory. Learning theory hypothesizes that pain behaviors can result from several things, one of which is the contingent reinforcement of pain behavior by factors in the environment. The issue of whether the pain is organic or psychogenic is not relevant. The behavioral approach assumes that pain behaviors may occur as a result of nociception, the adverse effects Of disuse and/or overguarding Of involved body parts, and contingent reinforcement from the environment. It is further assumed that pain behaviors are modifiable; they may diminish and activity level increase without directly and specifically dealing with the source Of nociception. Viewing pain in behavioral terms is ii radical shift in perspective. In an answer to the critics Of the behavioral management Of chronic pain, Fordyce (1985) stated: 58 There is perhaps no more difficult intellectual task than to learn and assimilate that what we know to be true is not. The traditional concept of pain as a kind of sensory system and sensory experience linked to nociception is probably essentially valid in acute, but not necessarily in chronic, pain. We must recognize that "pain" is an open system. Once pain behaviors begin to occur, for whatever reason, they are subject to influence by factors outside the person. In the extreme case, those pain behaviors may come under virtual total control of environmental factors and thereby persist in the absence of nociception. (p. 123) Cognitive-Behavioral Therapy Cognitive-behavioral therapies reflect a recent development in the field of pain management. In recent years, cognitive—behavioral theory and techniques have been applied more frequently to the treatment and management of pain in both acute and chronic nonmalignant pain problems (Goldfried, T977; Meichenbaum & Turk, 1976; Tan, 1982). This approach is an outgrowth of Melzack and Wall’s (1965) gate control theory Of pain and the growing interest among psychologists in the application of cognitive and cognitive- behavioral therapy to health-related problems. Ellis (T962) and Beck (1976) pioneered the use of "cognitive restructuring" techniques, which are used in identifying and modifying cognitions associated with negative emotions and maladaptive behavior. The cognitive—behavioral approach is based on the assumption that the beliefs, expectations, and attitudes people maintain in certain situations determine their emotional and behavioral reactions to the situation. Cognitive variables such as distraction and the meaning of the pain for the individual, as well as such emotional variables as anxiety and depression, influence the 59 experience of pain. Logically, the modification of cognitions can be used to alter the pain experience. The cognitive-behavioral approach does not ignore the subjective experience of pain; rather, it views suffering as one of several aspects of a complex pain problem. The goal of cognitive—behavioral therapy is to correct faulty cognitions underlying emotional and behavioral disturbance. When this concept is applied to pain problems, the approach may be considered "psychoeducational" (Turner & Romano, T989). Interven— tions consist of cognitive restructuring, coping statements, relaxa— tion, visual imagery, and assertiveness and communication training. These interventions increase the patient’s awareness of events that reduce pain. The patient may effectively avoid or adaptively cope with pain-increasing events by using pain—relieving actions. Awareness and knowledge give the chronic pain patient a sense of control over pain, replacing feelings of anxiety and helplessness. Individual adaptive coping plans consist of two components: (a) ways to prepare for pain flare-ups and (b) coping strategies to use at times of increased pain. Advance preparation involves continued practice in stress- management, communication, relaxation, exercise, and stretching skills. If these skills have not been practiced, they are unlikely to work effectively in a crisis. These coping strategies allow the patient control over pain perception. The more the strategies are practiced, the more effective they become. 60 These interventions are valuable components of an interdisci- plinary treatment program for pain. Although cognitive-behavioral methods for pain control appear promising, the evidence from con— trolled studies is still limited (Tan, 1982). Review of Treatmant-Outcome and Follow-Up Studies Of the Multidisciplinaryalreatment of Low Back Pain Introduction Traditional medical treatment aims at removing or relieving the pathogenic process, but such treatment is not possible or effective for many patients with chronic benign low back pain. Chronic low back pain presents a major challenge tO traditional medical science. Casey (1979) estimated that five million people in the United States are at least partially disabled by persistent low back pain, and 93 million workdays are lost each year due to this complaint. Yet for as many as 78% Of the individuals who are severely disabled by chronic low back pain, no pathophysiological basis can be found to support their pain complaints (Loeser, 1980). Various statistics have been cited about the vast economic loss chronic pain inflicts on society. Bonica (1980) estimated that acute and chronic pain costs the United States economy more than $90 billion annually. Chronic pain is a problem of considerable proportions. Traditional medical management Often has failed to relieve the pain. Chronic pain patients usually become "problems" in the health care system and are recognized as a specific treatment population. To 61 meet the demand for relief from pain, several behavioral treatments for chronic pain have been developed in the past 20 years. Chronic pain usually refers to the persistent complaint of pain for at least 6 Inonths. The multidisciplinary pain clinic (MPC) constitutes a major response to the challenge presented by patients with protracted complaints of low back pain. Patients for whom traditional medical management has proved inadequate, as well as those with no identifiable pathophysiology, are treated at MPCs. Criteria for admission to MPCs include the provision that patients have pain problems Of a minimum duration--usually 6 months, although most patients have suffered much longer. The highest mean duration Of chronic low back pain-~8.7 years—-was reported for a treatment group at Rancho Los Amigos (Cairns, Thomas, Mooney, & Pace, 1976). MPC patients share a history Of failed surgical interventions (a mean of 2.5 reported surgeries), accompanied by multiple previous hospitalizations (an average of about six) (Fordyce et al., 1973; Swanson, Swenson, Maruta, & McPhee, 1976). The ensuing review Of MPCs describes their response to the Challenge presented by patients with protracted complaints of low back pain. The studies reviewed in this section were categorized according to four general-types of behavioral treatment for pain: Operant, relaxation, cognitive, and multimodal. The treatment focus in each report was used in categorizing the study. Studies encompassing two or more treatment methods were included in the section on the multimodal approach, but research that focused on one technique and 62 used another secondary treatment were classified according to the primary treatment approach. The Operant Approach According to proponents of the operant viewpoint, pain behavior is a set of overt responses (i.e , medication taking, limping, pain reports), which can be controlled by reinforcers (i.e., attention, medications, and so on) if the reinforcers are given contingent on the pain behaviors. Fordyce and his associates (Fordyce et al., 1973) developed a treatment program based on operant principles. This program concentrated on decreasing medication use, pain levels, and pain behaviors. Fordyce (1976) focused instead on increasing activity and other constructive behaviors. In Fordyce’s treatment program, activity levels were increased gradually by systematically reinforcing each increase with social praise and the chance to rest. Medications were systematically decreased as the patient was provided with progressively smaller doses in a "pain cocktail," which was given on a fixed-time rather than pain basis. Trained not to reinforce pain behaviors, family and staff avoided providing sympathy or a reduction of work responsibilities contingent on pain. The experience of pain was dealt with only in the sense that pain behavior was ignored and well behavior reinforced. Of the 15 studies conducted from 1968 to 1982, most used a one— group pre/posttest design. Cook and Campbell (1979) pointed out that there are several problems (i e., controlling the history, 63 maturation, regression, and reactivity) to» measurement with this type of design, which makes causal inference difficult. In this type of research, factors extraneous to actual treatment such as placebo, attention, and demand characteristics may not be accounted for in the one-group pre/posttest design. It is difficult tO attribute significant change to the treatment unless there is convincing control over the entire experiment. In these studies, subjective pain complaints ordinarily decreased, but only by a moderate amount. In all studies that employed inferential statistics, statistical significance was Obtained, but clinically' the reductions in subjective pain complaints were not impressive. Swanson et al. (1976) reported a 14% decrease in pain ratings. However, in none of the studies was pain reduced below a rating Of 4 on a O to 10 scale. Although subjective pain ratings were reduced significantly, patients still complained of having considerable pain. In all studies except that Of Ignelzi, Sternbach, and Timmermans (1977), reported reductions were maintained. The operant program resulted in clinically significant increases in activity levels and reduction in analgesic intake. The researchers usually reported modest reductions in subjective pain ratings. The follow-up studies indicated maintenance of treatment gains. Although a high percentage Of subjects completing the program improved, only a nfinority Of the original pain patient population completed the program. It was difficult to organize, administer, 64 and gain control over important reinforcers in the operant treatment programs (Vinic, 1981; White & Donovan, 1980). The results from these studies support the usefulness of the Operant approach. There is no longer a question about whether the operant program is beneficial in increasing activity levels and decreasing the use Of medication. The question is no longer whether it works, but how well it works, for whom, and why. Experiments dealing with the specific components Of the program need tO isolate and improve on the effective ingredients. Defining which patients are likely to respond to the program is also important. In comparative studies, the cost effectiveness Of the operant program needs to be determined. Table 2.1 contains summary information on outcome studies using ' the Operant approach. The Relaxation Approach The basis Of the relaxation approach to pain is the assumption that organic processes are relevant and are influenced by learning. The basic idea in treating chronic pain is to break the vicious pain cycle that exists. Whenever injury occurs, typically causing pain, the tendency is to tense the muscles in the affected body area, thereby immobilizing the site to further trauma. In acute pain, this response has Obvious value. However, when the muscles are chronically tensed, the tension produces more pain, which in turn causes more tensing; thus, a pain-tension cycle begins. Eventually, other problems such as lack Of sleep, depression, overuse of 65 .ucoeo>0caem canyon co auwcouos o>oz ou noeooo ego: Nam .aa.:oddo» u< .oma cowuou -_uoe cmosu oomoocooo muco_uma ecu Aonon *0 nos uce .xcp concocooo c_ma .xo— “moa\oca cacao: a .auacmx mzucoe o nomoococ_ >u_>muoo .omcmzom_o u< ancem — Amnowv .dm um muxncou cu cm._emm xomn u~n--om .c0mcm3m .c0mcmzm .xum>_uoa commococ_ co ammo oommocuoc can no» new AQAOPV .mdo>o. co_umu_vos oommocooo on; umoaxoca «own a .xocoox mcucoe o— muco_uma any wo xmm .a3-zo..ow u< aaocm w Amquv ..a we ouxocou c_ m< xomn :o.--om .mmeozp .mcc_mu .>um>_uua vomoococ_ “woaxoca ooumum uoz .oOmmocooo oxoucm comuao_ooz queen — Ankowv .da 30 ooxocom c_ m< goon :o.--om Amkopv Lo.xoa .n:.:odd0w um poc_ouc_me mc_mo .c_ao new mco_umo umoa\oLa xcomcam po>_mooc om.a muco_u~a mN xomn 30d >.umoe mzucoe o -_noe commocooo .>u_>_uoa oomaococ_ anecm P “Accopv .de um ooxpLOE ou cm._e_m .owco>wo--mn AQNo—v zomncLon umoa\oca Amoo—V .do no ooxpaou Aqnorv compacoum ocoz .c_ma powwocooc .>u_>_uom nemmococu anocm — cu cad_s_m seaweed uco_uma-c_ omco>_p--vm w uOOECoOLU “MKOPV .noc_muc_ms mcmmm Emma new >u_>_uom cascummoc» a .qum .a:.:o..0w u< .c_ma new mcomueo_ooe umoa\oca xumn so. .camump on .ccmEcoo mcucoe NN pawmocooo .>uw>muom commocucm anocm — Amoo—v ..m up ou>vcou ma osmm >u_co_me--om .codzou .ooxocou momaocuoo comumo_ooe Amoowv .oma comumu_ooe vammocooo .m.o>od m< new mowmococm >u_>_uom o_uanum>m Laoump up a .ccmecop ocoz >um>muom pee mcmxdmz cowoocuca uuo_n3m odmc_m gum: Eecmoco acacoao aco_uma-c. xomp :o.--m .co.30m .ooxoLOE a:-:o.d0u madamom cummoo mucoEuoocp cmaaxz >u3um .goaOLQQm ucmcooo ecu mc_m3 mo_u3um oeouuzo--.w.~ o.noh 665 memox m-F sauces ~F 38: v mascoe an ocoz mzucoe o .CFmo uco .om: mace .xcoz .oeFu-a: o .n A a .Qalxoddo» um xddmeLoc mchOFchaF muco_uea ucocoao Fo xxx .socm0ca mace cu nococoo xo~ .mc_xcoz NoN .ooCFaucFmE mm: 33.8% 58 53-3030» u< .53 commocooo ..o>o. co_uao_uoe oo>ocae_ mucoFumo Fo xom “oomoocuc_ >um>Fuuo .omaazummu u< .uoac uaoaocu me .a:-:o..oF um voc_auc_a: .uo.n:oo 3o>3 5330c 6:29.33 «mega co .uao uoaaocu non aaoa< .ucmu_w_cm_m uoc mm: ommocooo Ewan .>comc:w mc_>_oooc uoc amazu co» pan ucFma oomeocuoo .co_uau -_ome vommocuoo .>u_>_uoa commococfi .xamcocu .oummxca cw wo_um>muum ooFFFanm co. m x n u u .>u_>_uoo new om: mace Fo mdo>od Lactoc on; Emcm0ca on» mc_uodoEou omocu Fo ans upcoeumocu co So 688% as us .883. can .oouooFoc xvm .mucomuma amCFcho Fo umoaxocq anocm m umoo\oca ancm F “moaxoca macaw F a:-30..o» macaw F umoaxoca Queen m umooxoca ancm F ucoeuoocu mcwmawoc mucoFueo AoV .ucoeumocu Law poquFoc mucomuma any “Amnon ..e «o ouxpc0u cu ca.FE_m any “exodv ..6 uo comcmzm mx_. eecmoca acomquIc~ do>o. >u_>_uoa xdxoo: CF ommococ_ dameoooam co ucom -c_ucoo wxoodn o>coc o_momdmc< Acnon noon -chum a Fooccooco we a:-3o..ou xumnvoov omsaacm + acoeoocoF -cmoc .mnco> Auv .ucoEoocoFcFoc Lmnco> any .mc_c_acu da_ooam 0: “my umc_c_ocu >u_>muum acomuma-c_ Amnodv .Lo uo ooxoLOE o» cm._e_m mwco>_o--wm goo: .xuma--oo~ scan so. xoa--sa owco>_p--vm homo: upon; -c_om a muconox Aomon camcozm w .muscm: .cowcmzm Amnon .oxFCD a .955 .moCQEm: “to: mam—EEC w .comncLoum .wndocm_ AFFQFV oc_mma w mac_mu £33 3593. w .mcomoax .comxomddao .odou .c0mcooc< aa-xod.oL madamox cm_moo mucoeumocF xuafl--0 puma x0e--qm c_aa\z >n3um .naac_ucou--...~ 4.9a» (57 ocoz ocoz ocoz .mxcac vac: cF new new >u_mcmuc_ Emma an: c_ ka Fo omaocooo "EFma nommoEQCF ou ud:o_FF_o "om_uLoxo “vacuum mCFLJU oocFauc_mE uoc wan .ucoeu06cu umcmw mc_c:o ommoLucF «\m ">u_>_uom can: mucoeumocu mc_cao o cu ommocooo «\N .LOLucou umco_>acon c_ma uxmm u ammoEooo cams ..0cucou ooom "COFueoFoo: ommoono dau_m>ca cm omaococm .couoeooon CF omaocuc_ xmw .cFmQ Fo acoaoc.F.ow c_ omaocooo NNm .co_ucommo cm ammoLoc_ x- .mucmflgo cm «3303 EN .33 ->mcon c_mo danco>coc c_ omeocooc Xmo .xdmu c_ma c_ omaocooo xmw .ooLF co_umu_ooe mucoFHma Fo xoo .ucowoca ucoeumoLu cog: xdco commocuc_ >u_>_uum new .mco_>mcon :ddox: we ucoeoucoFc_oc co ucomcmucoo commocuoo co_>mzon c_ma vo>comno moc .mmcmuuom mmoLoa OEF. -Ommn oda_udax xaococu dwowm>ca mchsn mmuono wade =m: Any "co_>ezon ammo ommmcooo cu new someway dao_m>za pea 6cm: co >u_>_uow ammococF cu .uxoxucmeoucowcmoc .a_oom Amy mcmcmocu >._Ew» .mcmcou_cos-Fdom .mdmom uoacucoo .xaacozu anecm co_>m;on .co_u -oxaLoc new xumnuoowo_n .xamcozu “momeca .COFuosuoc co_uoumooz mommoLUCF >u_>_uoa oFquoumxm UCw Lo_>a;on gadoz: poocoFc_mm booFDDw-o.mc_m oooF .xooc .cmma xomn 30. o_cocco--F~F Amwo_c ac>:a_sos a Loud_x ANQOFV compose a _c_a_c_uc_u Aomon onsmumu a .ommaa .cmemmom ._ch> a:-:o.doa mudawox cmFmoo mucoeumocF Aacanv c_ma umcocco-.F c_aa\z >naum .voac_ucou--.F.~ o.naF 68 .ucaomchmchoc xoom “a a n A o noo.m “a a n A o co_umo >u_>_uua ucmcoao .co_umxo.oc -_voe “ucoeo>0caem o >.co momu_>muum po_.aaa uxoacocu Lom>mcon + venom: :_ma no A o mcm>_. >._mo Fo momuw>muom uwoa\oca we :9: Auv “acmeuaocu u_c_do La. xomn so. Aqwon wochoFHm occ: “aceoFFFCmchoc on: “a a n A o c_ma aaocm m -amoc now “dacucou ummd-mcmu_m3 Aav UFCOEEo-.m~ a .c_.oz .c0uc_p .u A n mdmom uucmu_F_cmmm -coc woos new >uomxcm “a A o a n >um>muum .cmmo co_mmocaoo uucoo_F_cm_mcoc QeFu .mcomueu_vos co» Emcm0ca ocm3->mp -czop “a A o a n mc_>_. >._mo Co “OOFnam odmc_m acmcoao + comumxodmc oOFHQOm Auv c_ma mo_u_>_uou “a A o ow_ucoxo um A n + awoaxoca “mucomuma-uao mo co_umxm.oL oo_daaa xumn :0. Ammon ocoz c_aa .ucaoFFchmmcoc >uceuuoaxw aaocm m Anv “dacucoo um_.-mc_u_m3 “my o_coLco--mF Emumouou a cauc_4 Emcee coF .uooFFo Looco oz .oLucoo gum: ammococ_ >UF>Fuom cow ecosooLowcmoc .uquFo .msmcme on; on uca “my mm mmocoo oc_. nguom on .mc_c_acu co_ucommm Auv c_ma .Auv >9 oozoddoF .ucoeo>oco5_ .ommn o.a_u_:E .comuaxwdoc o>_mmocmoLa any .m_m xomn :0. ocoz .dwco>o cu umoe oouanmcucou any uCOcLaocooc02 ->.mce co_>acon-cmma .mcomuucau any o_cocco--q AmwoFV mconcom 3-13 28 3 38¢ 5:80 seesaw: c :5; >93 .ooacmucou--.F.~ odan 69 analgesics, and lowered activity exacerbate this cycle. Treatment through the relaxation approach focuses on reducing muscle tension and psychological stress to control the pain. Table 2.2 contains summary information on studies in which electromylograpyhic (EMG) biofeedback or other relaxation techniques like progressive muscle relaxation were used TX) break the vicious pain cycle. EMG and pain levels are the principal measures reviewed here because many researchers did not employ other measures. Most of the studies on relaxation treatment conducted between 1977 and 1981 used a one-group pre/posttest design. The limitations of this type of design were discussed in the previous section. Another limitation of these studies is that they lacked a broad range Of objective outcome measures. The typical investigation included EMG and pain ratings, but few researchers reported on other pain-related behaviors. Follow—up periods also were quite short. Overall, the data from these studies suggest that many patients may benefit from relaxation treatment. II lack of data from well- controlled studies makes this conclusion somewhat tentative. There was a general indication that chronic pain patients can significantly lower their EMG levels and experience it decrease in pain. It is not clear from these studies how much pain levels decreased, whether other pain behaviors were affected, (H‘ why the pain decreased. One advantage Of relaxation therapies is that they are inexpensive and easy to administer. The use Of relaxation as a part of the coping-skills approach to pain appears prudent. 7() .voCFauCFQE ago: mcomuoaooc CFeQ Fan .o anocu coF m.o>o. ucoeumocu -oca cu poacauoc oxm .a3-3o..oF umoaxocn .0cucoo um_.-mc_u_e3 any Aonon mzucoe m u< .mcomuoauoc ammo pea arm a A a aaocm ~ .xomnoooFo_n ozm odomae-xumm any xomn--o~ Lomcmdom a cozaoz .oocmouc_ms mcme >u_ ->Fuoa uca co_uoaooc cmma .omcmzo -m_o «a >um>muua oommocuc_ m3.a m< flakon mcucoe o xocoaoocw cFma vca ozm oomaocooo Foo_nam odmCFm somncooFoFQ uzm odomee-xomm xomn--F cocoa a caflmm 85-3033 um oocmouEFme mc_mm omocF .mcomuusooc ammo on; c *0 ~ new chFFuaooc arm on: muco_uaa are odes: .co_a -oapoc ammo a no; a mo F van mco_u umoa\oca wa o Amquv wcucoe m -oaooc ozm acmmdm on; mucoFFma xomm anocm F AcmuommmE\xomnv xomnoooFan uzw .xomn w--qF uchx w xuoa AFNQFV .Fo_.oe «Sam on; mo_naum ammo ate; a .w.do>< cocoa F muco_uma mF on» Co 0 .a:.:oddoc u< paddocucouc: xomnpooFo_n ozm w__muc0cm omco>_o.-mF .mFummocoo .Lo.ncw: .Q:-3o.doF um mCme uocmauc_me c any Fo ouch .omcmcom_o an 6005 oo>oLaE_ oco wo_uaum ammo mc_m:oo* memo> ~-F ammo commocuoo no; muco_uma dd< cod~0cucouca new cowuwxmdoc o>_mmocm0ca oc_am--q Amnon xm_mm~cu .Q3-:od.oF um oocmauc_me ace: mucoe -o>0cae_ CFaa .momwoLooo c_ma on; >ozu umzu vo_:u=_ mm: um umdo>od mo_naum ammo Auquv mxoo: 0 arm oooaooc om; muco_umo cuom no..0cucooc: xumnoooFo_n uzw m_.mucocu xomn-.~ dmcLom w >cucmo a:.:o..0m madamom cm_mmo mucosueocF cmma\z xbaum .zumocaam co_umxmdoc any mcFm: mowuauw escapee...~.~ a.gac 71 .m mCFLDU >.co £3c_8838_835F% é mcmcao uoc uan .o~_dmauo o» oovcou oc_aw Fo mmoFm “we. can uszL co do>od uzw o~_.aauo cu Amwon >d_ox a see: as“ .maamca Ame “casuaoau mc_cao mu_._noe .mucoeo>ocaEF oc_.ommn o.a xcomme_ noo.m .mco_uu:ooE cFmo ucmu -_u.ae gum: m< poUFJm new .mommocoxo mc_;umocn mficoe «F 4357. 3385? 85-3033 u< 38.33 39.3 £033.39. Bowl: 6339595 mttficmim £ng _c._m> comuodaooc_ .uomcmzuc: who: mao>od c_ma m< mmocum pea .mcmcmacu co_umxo_oc Aowon ocoz new uzm wan .vo>ocae_ co_mmocaoo nooFDJw odmc_m .xoenpoowomn odomoe-xomn ozm xumn :o.--F cmdmm w pooF .oo>comno m< mc_c_acu co_umxmdoc Aomon ocoz wdo>ad CFma pea ozu commocooo HOOFQJm odmc_m new xomnuoo»o_n arm o.omae-xomm xo¢p--q cosmFE w .m_z .m.o>o. oxaucm-maco new >u3um ommo Aowon mxooz m .c_ma .uzm vomoocooo .a:-:o.doF u< paddocucOUCD xomnpoo~o_n ozm ogumae-xomm xomn--F v.03 w mecca .EE v2.80: 8.. o co m use .c_aa «was no; a co 3 .a.o>o. sowocc ozm c_ co_uoaooc Now a vocowoLaam umoa\oca cm.m: a .om_ma mcucoe m mucoFFma m wo m .a3.:o.dow u< anecm F xumnpoo»o_n arm xomn :o.--w .cxmdau .cmeoocm 8-33 .8 2 38¢ 333 35.58: c 3%: 9x... .8:£ES:.~.~ Be: 72 ocoz ocoz ocoz ocoz ocoz .mco_mmom coozuon new c_;u_3 omoocuoo "commcou umco_mmom c_;u_: omoocuoo uc_ma umco_mmom cooxuon uco c_;u_3 omoocooo "ozm .ucoo_w_cm_mcoc modnomcm> Locuo Eda u>um>muoo .co_uouvm_umm n A o co_uo3~m>o “mo_u_>muoa dm_oom n w a A 0 Home» >u_>_uom “Ao.nmo ._.oao uoc av “a A o "co_umu_noe “Aomeocooo wa u by a A o acme; .ommocuoo CF u u a norm “Anon .uonv coFFoucom -oca Fo mcooco oxw c_ um oommococ_ a .cme pounce; o a n "CFma ucmomwmcmchoc "ammo “acmumw_cm_w .8: n 90.5 689.0% m 90.6 norm 6033?: SB F .0398: -moso norm no A a "wooucou Ecuooo uucao_F_cmchoc "COFuao_poE no a n A a "ocwocco_umoav o>_u_cmoo “u a n A m mco_u_Cmou .u a n A a o>FuooFFm .ucmu_chm_wcoc >comcom “Fa: “co_u -acso oce >uchoucF m a N A o "cpoa ucflcogsmag meo:c_ucoo ;o_3 umoa\oaa Queen F umoaxoca ancm m um< 30.33 035m umoa\oca QJOLm N umoaxoca azocm n >mo Loo mco_u_uoaoc ou:c_e-F on umc_c_ocu coFchLu «ow .coFUaxodoc o>_m -mocmoaa “no .xumnnooco_n oz“ Aac co_umxmdoc oomdaam + n a: .ucguaob 823:5 -mcoc any .docucou um_.-mcmu_m3 “my comm CFma --xumnooowo_o uzm on .xomnooo» oc --.ocucoo any .vo_coa mcFUFmJ Amy decucoo umF.-mc_u_m: any .Aomc_aw Ecuoocov xomncooFo_n arm Aav .0cucoo um_. -mCFusz nov .xomnooow.ooaoma Anv .moau c_mo o» oodaaoo ou_u -oeca use; + o>_umcmoo + Amadomas acm>o.ocv xomnoooFan ozm “av c_ma some so. u_coL;u--mF CFmQ some so. u_coC;o--AF c_ma odomae o_coEzo--mF c_ma xomn 30. o_c0c;o--o~ m_u_czuco Bassisté AFcon umzcam a .mea_.._: .czoLm .ac_oozum .oFoox Ammon c_.az a coac_3 Amwon £3 a omcmp Ammo: c0130: Ammon Lodzoox w .xcaF .mao: .Lodm a:-:o..oa mu.:mox CmFmoo mucosuoocF cmaa\z >33. .uoac_acoo--.~.~ «.9oc '73 .mo_nm_co> age mQJOLm coozuon acmo_w_cm_mcoc “ca ou pocwaeoo >u_mcouc_ c_ma uco ow: memoIzudao; oomaocoov masoLm com>mcon-o>_u_cmoo new comuaxadoc o>mmmocmoLao .oaz new >u_mCOuc_ c_ma LoF mmmocox; A co_uoxa.oc unoc_muc_me >.dococoo n .modnomcm> Locuo co moucocoFFFo aaocm ucooFF_cm_mcoc “emcmoca ucmcooo + co_umxmdoc co_.oaa A co_umxm.mc oo_daam .c_ma "masocm coon .wucmeo>ocoe_ voc_m~c_mxm .mo_u_>_uuu .oUCoco.ou cFma .m.aom n A o "mmc_uac.wdom “CFma mucoeoumum-F.wm .c_ma ou momcoamoc o>_uoowFa + mco_u -chou .coo_ .xcommEF + co_umxa.oc .mdwom "Lo_>mzon o>_u_cmou AoF Emma omcmox .a_m m A o a a “Aa_wv >u_._nam_o umoaxoca ua>._mo\Fv co_umxmdoc o>_mwoLm xomn 30L o.~-m.F .COmeocaoo .cmma po>ocas_ u a a cacao m -oca any “Lecucoo um_d-mc_u_mz any o_coLco--om Amwon coacaF .aa: pea >u_mcouc_ c_ma a: -zoddoF an n A a "moococoFF_o mecca .MEFu-Q: wan mo.nm_cm> ddm co m.ocu c_ma Amwon cod.mu w -coo onooada nan ammuoca cu pecan xomp 30L .ocmd._3 .xcmcu nmcucoe m .Eoo xducmuFchmFm vo>oEaEF n a m w_moca>;-w.mm any .co_umxm.0c “av o_c0ccu--~F .codonF .xodamuox .o A n ”m.m0m uucmo_F_cm_m -coc "boos new >Uo_xcm mm A u a n >u_>_uum ocm .c_ma "COmeocaoc uucmo_w_cm_m50c "me_u .mcoFFaoFooE co» EmLmOLQ ocm:->mu -czoo no A u w n umc_>md >._mo Fo uoomnew admCFm ucmcoao + comumxmLoc po_.aam gov CFma mo_um>_uua “a A o "om_ucoxo “m A n + uwoa\OLQ uucoFuma-u:o um co_uaxm.oc oomdaam xomn 30. Aqwon mmzucoe o "c_ma uucmuFF_cm_mcoc ">oc6uooaxm oaocm n “no “doLucoo um_d-mc_u_mz any uwc0czu--mF Emumouou a coucmp a:-:odd0m mudsmoa cummoo mucoeumocF cmwa\z >v3um .uoac_acoU--.~.~ O.BAF 74 Between 1982 and T986, the relaxation approach was used in five studies employing biofeedback; these studies had mixed results. Flor, Haag, Turk, and Koehler (T983) found that biofeedback was better than no treatment or pseudo-therapy for pain, but the group difference with regard to EMG level was not significant. KEefe, Schapira, Brown, Williams, and Surwit (1981) reported a decrease in EMG, but they Observed decreases in pain only during feedback sessions. Nouwen (T983) reported significant EMG decreases in his biofeedback. group, but pain was not reduced significantly. TO confuse matters even further, Large and Lamb (1983) found that both biofeedback and pseudo-biofeedback could reduce EMG and pain under certain conditions. The data concerning biofeedback do run; furnish convincing or conclusive evidence as to its utility. Electrode placement and differences in training procedures might have accounted for some of the variance in results. In addition, the site of pain might have influenced the outcome. Various forms Of relaxation training have been used in remedial studies. Sanders (1983) found that progressive relaxation was the most important component Of his program. Turner (1982) compared progressive relaxation with no treatment and a coping strategy based on relaxation and cognitive strategies. Although both active treatments resulted iri significant improvements, Turner iRnnui few differences between subjects treated with progressive relaxation and progressive relaxation plus cognitive methods. 75 Linton and Melin (1983) added applied relaxation tO a day ward’s rehabilitation program and compared it to the ward’s regular rehabilitation treatment in El waiting-list control group. The results supported the idea that adding applied relaxation to this program led to improvement in a variety of variables. Linton and Gotestam (1984) continued the research and found that applied relaxation by itself led to improvements of approximately the same magnitude as did both applied relaxation and an Operant program. The operant-plus-relaxation group was somewhat better on variables of activity and medicine use, whereas the applied-relaxation group was superior on pain-intensity ratings. The results Of these studies indicated that relaxation training was effective in treating certain aspects of chronic pain. Although none of the researchers compared relaxation training with biofeedback, Linton (1986) stated that relaxation training was at least as effective as biofeedback in treating chronic pain. Turner and Chapman (1982) said that biofeedback has had mediocre results in reducing pain because it is; too simple an approach for such a complex problem. The Cognitive Approach Cognitive-behavioral therapy has become increasingly pOpular in recent years. The cognitive program includes several approaches to regulate pain. In this approach, pain is seen as an experience mediated by cognitions. If this approach successfully negotiates pain, there might be a concurrent change in other pain-related 76 behaviors, such as activity and medication intake. Regardless Of the origin of pain, modifying cognitions may reduce pain levels and provide the patient with a better method of dealing with it (Linton, 1982). There is little evidence that cognitive strategies are effective in treating chronic pain (Tan, 1982) or that they are the treatment Of choice (Linton, 1980). A lack Irf studies, in combination with poor designs and the use Of cognitive with noncognitive treatment, does not allow one to draw conclusions regarding the effectiveness (Hi the cognitive approach 'hi treating pain (Sanders, 1979). In addition, the clinical validity Of this approach has not been tested (Sanders, 1979). On the positive side, however, Rybstein-Blinchik’s (1979) four- group study suggested that cognitive strategies may be useful in a clinical setting. Whether the effects of cognitive treatment are strictly cognitive or rather relate to a general relaxation response is an important theoretical question that needs to be addressed. Between 1980 and T985, no additional studies were conducted on the use of purely cognitive strategies. However, three reviews related to this topic have been published. Pierce (1983) reviewed cognitive-behavioral approaches to treating chronic pain. The author classified them as either pain directed, requiring restructuring (H‘ distraction, versus stress directed. A common element Of stress—directed programs is relaxation. However, there is no empirical evidence that the addition Of the cognitive techniques is helpful (Linton, 1986). Pierce (1983) stated that 77 there are not enough substantial data, particularly with regard to pain, to draw firm conclusions about the utility of cognitive aspects of stress-directed techniques. Tan (l982).and Turk, Meichenbaum, and Genest (l983) evaluated literature including laboratory studies. Turk et al. concluded that laboratory studies, as a group, have not shown one cognitive strategy to be superior to strategies subjects bring with them to the laboratory. Tan and 'hn4( et al. acknowledged ‘that few researchers have investigated chronic pain other than headache. The review of cognitive approaches to date showed the need for more research. Linton (1986) acknowledged that not enough is known about cognitive approaches to controlling chronic pain. Turner and Romano (l989) acknowledged the beneficial effects of cognitive-behavioral therapies in the field of psychotherapy. However, no study was found in which a purely cognitive approach was used to treat chronic low back pain. Table 2.3 contains information on five outcome studies using the cognitive approach. The Multimodal Approach Because pain is a complex problem that is influenced by many variables, the multimodal approach is an attempt to increase improvement by using several techniques to control as many pain variables as possible. The multimodal approach includes operant, relaxation, and cognitive strategies, in addition to a wide range of other techniques. 388333 91:. .2350 5833 50:338.. .333 it .5023 88: 9.8: o .CEEm mFEumoEQm: £3 a A m 98m N -39: $33 $33wa mfidm A3 «235:9 583ch a cmEtm: 608:3 3E8 -CEmE no: 5.553.: 53 2:3 Btumwn 3 8.5: “an .mco._>mswn $8.wa 33$ VERSE a 9.8: m 53 £9.53 53 Bmmwcumo 30.3 F 503383353 witcmou wmcoZuim x_c8:m.c_3wn>m ._ .8358“. “:93“: .53 5.5 3965 wificmou AB 8 33:35 58 m A v 53:28 ucm>3PETEma 5:. 7 Ba u.m A a $203 $33.88.". umoahta xamcwfi wificmou A“: $3.55“ 33: ocoz van 32>ng Ema v36 x a 38m v -53 3V Jobcou 523.623 A3 mmcmlvlg xEBZmEEHmPAx .35033 S >35 «28 53 we 528.55 55:93:58 9:82 o 8.35 man Boga; vac out Ema 3:95:08: tw>ou 502332 wamEmoca fluffcmlp £33 3333 _..voumcv_uoE= >35 33 6::wade 3.5.2.500 .xcommc: “who: Numcxcma vcoz mm: 53 Em 330.5% em: 98 8303895 tw>oo 60.5983; 333.595 .mc..EoBm.; a 39833 94.3030“. 333x :2va 3553: fan; :53 .zomocaam wZZcmoo or: 9:8 353m 2638--.m.~ m3: 79 Weak designs characterize the investigations reviewed in this section. The studies either used single-subject AB designs or single-group pre/posttest designs. A broad range of dependent variables was used in the investigations. In these studies, considerable improvement in such variables as pain reports, drug use, mood, and activity was reported at discharge and at follow-up. The clinical significance of these improvements as a group is difficult to judge. The studies varied in methodology and treatment; in many of them, the actual data were not reported. The preliminary nature of some of the studies and weak designs in others prevent one from drawing conclusions about the utility of their approaches (Linton, l982). Because several techniques were employed, it is difficult to determine the efficiency of a particular component by itself. However, despite the research problems involved in nufltimodal treatments, their popularity seems to be increasing (Linton, I982). The promise (H: multimodal treatments lies in combining the most effective aspects of various approaches. Weak; designs of’ the one—group type continue to plague the multimodal approach. This problem makes it difficult to isolate the active ingredients of the approach. Each clinic or study has used somewhat different treatment components, giving the multimodal approach an "everything but the kitchen sink" character. In addition, comparing studies is difficult because the actual treatments usually have not been described in detail (Linton, l986). 80 Another difficulty with the nmltimodal approach is that, the more components the subject participates in, the lower the level of compliance and memory of infbrmation tend to be (Linton, Melin, & Gotestam, 1984). Although treatments should be broad based, they also should consist of as few and as potent components as possible (Linton, 1986). Follow-up reports exist, but there is no evidence that analyses of effective treatment packages are being conducted. This reflects the complexity of conducting such controlled investigations in the pain-clinic setting. The preceding criticism provides a foundation for reviewing two recent studies of multimodal treatment programs. Maruta, Swanson, and McHardy (1990) performed a 3-year follow—up study (H: patients who had been treated in a multidisciplinary pain-management center. At the end of the 3-year follow—up, 46.6% of the successfully treated patients had maintained their improvement. This was viewed as an indication of the long-term efficacy of the MPC treatment. Connally and Sanders (1991) included cognitive coping strategies as a treatment variable in the evaluation of pain treatment. They examined the ability of initial overt pain behavior and cognitive coping strategies to predict chronic 'an back pain patients’ subsequent responses to lumbar sympathetic nerve tflocks and general interdisciplinary pain rehabilitation. Cognitive coping strategies were assessed using the Coping Strategies Questionnaire. This instrument measured the cognitive coping techniques of diverting attention, reinterpreting pain sensations, making coping self-statements, ignoring pain sensations, praying or hoping, and 81 catastrophizing. ‘The results indicated that, time more overt the pretreatment behavior, the poorer the outcome. 'The researchers concluded that overt pain behavior might well be a significant predictor variable for specific and combined interdisciplinary pain rehabilitation techniques. Specific studies in which the multimodal approach was used are summarized in Table 2.4. Conclusions The information obtained over the past 20 years has provided considerable support for the efficacy of some behavioral pain- management 'techniques. Over' the .years, studies have had almost exclusively positive results. Appropriate methodology has allowed researchers to draw more definite conclusions. Studies of operant programs have provided strong evidence that the main activity and medication programs are effective. In at least 30 controlled studies, positive results have been shown for the behavioral approach to pain management. However, researchers have had difficulty demonstrating that overt pain—communication behaviors such as pain, talk, rubbing, grimacing, and the like, were reduced through the operant approach. Reductions in pain-intensity ratings also have resulted from treatment, although only a few of the researchers who found decreases in pain ratings used proper pain- rating methods. Controlled studies either combined relaxation with the operant treatment or did not show a clear reduction in pain 82 .co_mm_50a co Coca cmma omcoz co menu ogu v0; mucomuma we >u_co_as on» uan .vommocumv “who—V mummou comuao_uoe .vomaocucm >u_ umoaxoca coumc__cmo a .oamo> mxom: om -d_noe uco ammucwxm .a:-:o..o» u< aaocm F Reno—v cmsxwz w macaw ma «som gown :o.--om .mmcom .coezoz .a:-:o.HOm uo voc_au .cmme mcmmu .vmcazum_u um xducmo >aacmcu >deaw ucacoao nan -_»_cm_m vammocuou mmmCmmodoao; pea umoaxmca .xnacozu o>_u_cmou .mc_c_mcu xomn Amnorv cam» F .om: comumu_vos .co_mmocamu .c_ma anecm F -voouo_n Lo co_umxaduc uco_uma-uso omco>_v--m swam w _eaumcx 43-3033 no 638 mm: 33.65 x5. a $5532 823 xducmumwmcm_m vomawcuou cmoa a._53 .mco_umu.3mcoo .au_uoe new om: macn ANNOFV vowmocoov .vOmmucoc_ mm: oucacuHOu umoaxoca acumc_flcmo a .oamo» mxoo: om om_ucoxo ucm >u_g_noe .a:-:o..ow u< QJOLm P Aono—v cmezmz a mocmm mo osmm xomn Jo.--om .cmeamz .mocom Awsowv >aacmcu dwu_m>ca new .co_uac0ummc Loammz a .Cmeao.u .co_umuo> mc_xoom Lo mc_xco: 0cm: .mcomuwuo> .xumnbovmomn .co_uu3nmc .£u_Emcoxuo: Rpm ncm .vo>oLne_ >..mcomuuc:» new umoacha maco .comumusvo .mc_c_mcu wmwc .xxmconm: .cod.ou cucos _ >..mu_c_du ago: Noe .omcmcum_n u< ancm F -m>_ucomwm yo smLmOLQ uca_uaa-c_ xomn :6.--m~ .mu_cum .oo_duuou .a:-:o..0+ an voc_muc_me mew: van vommmcucm >u_._noe nca >u_>_uu< .a3-:odd0w um NNN new omcmzo xnacozuozu>ma nan .co_umo:no -mmv um Xm cu mucomuma ocu we um0a\oca .co_umxa.ac .xumnvoowo_n mada c_ma nonowv mzucoe m Rum Eccv vomcanu ow: comumu_vox naocm F mc_co_u_vcou ucacoao ucu_uma-c~ xumn 30.--oop c0510: w mucom a:-zo..ou madamox cm_moo mucoeumccp c_am\z xuaum .zomOcaam dmuoe_u.:E a mc_m3 momnaum decouno--.v.~ o.nmp E33 m.._xm co_u -au_c:EEou new .>aaco:u w>mumcmou .mmoco>_ucmmmm commocuc_ pew «moa\oca .co_ua.:moc-w.mm daummodommxca made Aomon Lod>mo a 0:02 .comwmwcaou uco c_ma vammmcooo anacm F mc_co_u_vcou ucacoao ucm_uma-c_ omcm>_u--om .Lmeocx .xoodm .:o_mmocamu qu .xuumxcm umoaxoca “onopv cm_cm.o a 0:02 .c_ma .mmmCmmm.mao; nmmmocomo QJOLm F Awnowv snag a _em~wcx mm mamm mwcw>_n--o .>booz ._Emum£x .a:-:oddo* um mc_mm vocmmu -c_mE muca_uma vo xmn .OcoN cu 3366.. 805356: 605.55 om: demo guano; van .gu>md >u_>_uom ._azz .mauaum xco: .mco_uadoc xamcmcu >d_5m. wcmm> N >._Ewu .ucmeumocu wo uao onQOLv umoa\oca mafia Amnorv ..m um oo>Uc0u mxm. xumn -mcucoe 0 Na van .vOmaeoc xo— .uauum_oc uh» asocm F mc_co_u_ncou acmcoao uco_uma-c_ :0. xon--q~ AoNo—v mcomoax “Ono—v .mdd_xm .oLucou-c_ma om: ou vo:c_u ;u_8mcoxuoz a -cou new co_uuuo> mc_xuom co mc_ awoa\oca .cwddox .COmcoa.< cm?» — -vto: 3:3qu we x3 85-3033 u< 90cm — 9R0: Jo am 903.38 mm 2.3 x08 307-: .8333 .a:-:o.d0w um vac_muc_me mucmeo>0cas_ .xmm vwmmwcuc_ w.m>m. >u_>_uom cave new .xnn vommmcuop mmacu ow: co_umo_poe cmvs .xmn vommmcumu awoa\mca o_a0cuo;u>ma new >amcozuoca>z mzda “chorv cmm> — mmcmumc cmma came .mmcmgowmu u< aaocm — mc_co_umucou «cacoao ucm_uwa-c_ mwcw>_v--m— mmuu_vocom mo a:-zo.d0u madamom cm_moo mucceumoch cmma\z xnsum .vmacmucou--.q.~ manmp 84 .nooanoL FHFCmu_F_cm_m uoc oxmuc_ co_umo_vae .mxoodn m>gmc .mc_c_cgu coFua~_u_mcomon oFmomdoc< .ou3u_uua vo>0Lae_ umoa\mga new co_uoxa.ug .xaogmsu u>_u_cmou FFmon vac: “smug uco co_mmogaov vommuguoo QuoLm F van anagm .co_uau:vm uco_uma-u:o omgo>_v--m~ mum_uamm a cmfigm: .vomaogucm >ducmu_FFCm_m moFuF>Fuum xdmmv mLo_>m;wo new .mwmou Luzo. c_ new mmDLu =.do:= mo ucoeauLoFcFmL van .co_u now: muchuma L030» .uowmogumn umoa\mga -muano .>aagogu amu_m>ca .mxoodn FFmon uLoFumgm a 92:8. Fm Eucmfitcgm 53 59-3033 u< 903 F 9.,me 5033:3833 £533-30 mmgoiviooF .305 69de .mdo>o. >u_>_uoo ummmoguc_ new wxaucm comumu_nme Lmocu voosvog mucomuma ogu Fo Rho “a: ogauocanaua vCa .sz -zoddoF uo xoq cu xmo EoLF vmuaoOg awoaxoga .mxuodn o>goc .mco_umu_nme vousvog mumox ~ ohm: c_ma ogo>om mcF>mc muca_uma anagm F .m_moca>c .co_uaxadog ucm_uma-u:o wmgm>_v--FmF AFmoFV mzmxaL< Aomon xmmocu .omgazummn um mco_u >amgucu aaogm vca m>_umcmou .comu a .ccmeuuwxzum -mu_ume mc_m3 mm: muco_uma ocu Fo umoa\aLQ -oxodog .comumuaco w:_a EmgmoLa .>o>go: .umdmz mcoz mcoc “ummogoc_ cu voeoom >u_>_uo< anoLm F mc_co_u_vcou ucmeao acoFFma-c_ mmgm>_v--oq .gwcxmu .mcmcpzh .mco_umu_voe Luaow mc_w: wgu: 0\m “mammogoc_ xum>muum uxw new “mucus >amgozu Fowon .m>oLaE_ woos o\~ .mommuguou c_ma m< o>FuLoaQ3m new .mc_c_agu co_um aged; a .>u.Lou mzucoe o acmumwmcmmm Um; o\o .a3-:od.oF u< poomnam wdmCFm -xmdvg .co_uoo_F_xoumv ucw_uma-c~ vmgm>_v--n .xu_c~:.~ .godme a:-:o.dou madamoa cm_moo mucoeumoLF c_ma\z >n3um .uoacmucou--.<.~ o.naF £35 .m.o>o. >um>muou vammogucF vac xov vac .mu_wom.uco Loscw now: new .ogoe go mgauocaaauo .mcomuquCF .mFCFOQ Lomm_uu .mxuo.n 8&on :03; w space F Law woF.og cFma meow was umoa\wLQ w>Lwc .Emgmoga >u_>_uum ~c3380.. .comduz .mmsmz mgmux m-F muchuma «cu Fo x~v .a:-zodd0+ u< anoLm F comuau_uwe .co_umxmdog uco_uma-u:o wmgm>_v.-noc .aaguwd._ .msz .a:-:od.oF um voc_auc_me mc_mu .mc_c_agw cm Lo mc_¥L03 ago: mF Fe m new .vommmguwu >amgmcu >._EmF uca .anLm .dmco_u vs cu voucau c_ma .o o» vooavog umoaxmLa -maauoo .dmu_m>ca “mc_c_agu comp “Fmon me> F mm: om: comumo_nae .omgmzummv u< aboLm F -mxado; “co_uwu_w_x0uwv ucomumac~ omgw>_u--mF comgmnc< w ong .kumogu >.d:meouu3m uaeoou NNM .d.ago>o .c_~a mmwd go mama mg» vac Row vca .mc_c_agu E .8 Exodus xmu 3.62:3ch 8:358 .9533? we» mcFaou Swot mgmmx m Lo .mucmxmdmg .mu_uougmc mcFma umoa\wLa m:.Q AnkoFV .dm um mu>bL0u cu gm. mchE_F a .>m>Lm: -wcucoe o no: mucmmuma Fo xnm .a:-:o..oF u< anoLm F -_Emm acmeummgu ucmgwao ucmmuma-c_ voumum yoc--~m .mexmu .uodmr .>u_>_uoa c_ mammogoc_ vougoaog xmo “oxmucm >ngmzu co_umo_vos gmwzu vogmzo. mucoFHoa dwu_m>;a van .mmagu o_a0guo;o>ma mo Noe van .xom vouauog swan .oxauc_ comumu_uos vooma-wdvm .mc_ FFmoFV .>.ucmowwmcmmm vammoguou w.u>a. um0a\mga -cmwgu co_umxm.og Fo co_umumdmgocom u_:L:m a .mEmmdd_3 mcoz ozm uca mmcmuag co_mcou m>Fuuomnam anoLm F .co_umxa.mL voum_mmo-ozm ucmmuma-c_ xumn :o.--FFF .xuo.m .mFoox a:-:o.d0u madammx cm_mmo mucoeummgp c_a¢\z >naum .uoac_ucou.-.q.~ «dam» €36 .ucmEuaoLu c_ma xgmc_da_um_ugmuc_ w>oc< .3 9.83 33:59:... .595. 3 «.3835 53 .do: ma ucoamog uoc own go_>mgwn vmuomamg mxoodn o_ua;umas>m gmngz. .co_umu xqu 30‘ “Foon mumnCmm mcoz c_ma ugw>o whoa mc_:ogm macaFFma n x N -md_nmzog cmma xngFHQFum_bLouc~ o_cog;u--oF w >.dmccou >aonguozu>ma m>_ugoaa:m .xuanvwwwomn .>amgmguocu>ma anoLm .n: .comuausvo .uCQEommcme co_uau_nme mou_m -30..o» gmo>-F mm 05mm .Q3-:o..o+ mCFuag .mmgammms co_umu_d_nmcmg .mu_m>ca m:o_um> .c_wa Aooon >ULmzuz w mgmm> m an acmEm>oLna_ vmc_muc_me xoc ucmvcwaouc_ .mc_co_u_ucou ucmgmao\m>_u_cmou u_coLcu--0q~ .comszm .muzgm: ucmeomQCmE mmogum .co_umxo.oL .mFmoca>z .xumn .mxmuCF comumuFuoe .md>umoFFd .CFmQ -vooFOFQ .xaogocuoguxma .co_uuaoog mLmvdzocm .xooc vo>0gaEF "modna_gm> odmcww “Amudnm a:-:o..oF co_uau_uoe .mcomuom_c_ .xaagwnu .c_ma xumn 30d Amwon Camdco -_Lm> ddmv ucoEw>oLaEF ddmgm>o aaogm F dmco_uma3uoo .xumgmcu .mu_w>;a u_coLcu--Nm a .ugmo._m .~u34 .voxoda -Ewc: no» xducmuwv_cm_m vommmgocm >amgmcu mFFmF> cudmmc .Fmeo.30ga gozmw 30F.me vcm .co_umELoFc_ .>amgocu um; mchLo: omeguv magnum xgoz cu daumm>ca .uco_uma gum: mocogavcou c_ma xomn vouadug .m>o. co_uau_voe new c_ma a:-:o.doF ammo .co_uu3vwg cowumUFvoE .mcF :od o_coLco Awwon Loddou w .Fxmev mCFxLo: Lvosdgr; ommoguc_ anagm F -dmmcaou doQFmodocu>ma .xomnuwowo_m xdcmme--- .COmLoQ.< .no_.uuoo a:-:o.dou madamom cmemo mucwsummgF c_ma\z xnaum .vo:c_ucou-..q.~ ~.nmp 87 ratings. Linton (l985) and Sanders (l983) agreed that operant treatment programs resulted in "moderate" decreases in pain ratings. Relaxation methods continue to be widely used, as demonstrated by the number of studies in which these techniques were employed. In controlled studies, biofeedback has had mixed results. Limited support exists for its efficiency when used alone. 0n the other hand, there has been more support for the utility of relaxation training. Relaxation training has resulted in significant reductions in pain reports, along with moderate improvements in medication intake, activity, and mood. Applying relaxation to everyday problem situations, such as using it as an active coping strategy, may be more effective than employing static training, but the evidence is still insufficient to draw any sound conclusion. The biofeedback programs with the best results have used some form of coping (Flor et al., l983; Keefe et al., l98l). Sanders (l983) and Turner (l982) used ordinary progressive relaxation in a program that required a good deal of practice. This suggests that a major difference between successful and unsuccessful programs may be the method of training and employing the relaxation. Linton (l985) believed that the operant and relaxation approaches are most effective in changing those behaviors toward which they are primarily oriented--activity levels and medicine use as opposed to subjective pain ratings, respectively. A generaliza— tion of treatment effect seemed to occur in both operant and relaxa- tion programs, but variations in the studies did not allow the reviewers to determine the extent of generalization. 88 Linton and Gotestam (1984) and Sanders (l983) examined the effects of relaxation in comparison to the results of some operant techniques. They found relaxation to be surprisingly effective in comparison to operant approaches. Linton (1985) questioned the need for in-patient operant ward programs. He believed that patients who had a moderate problem with chronic low back pain would require only an out-patient program, whereas in—patient treatment might be required for effective Inanagement of severe cases with narcotic addiction, very low .activity levels, and the like. Linton and Gotestam (1984) and Linton et a1. (1985) showed that out-patient behavioral treatments could be as effective as ordinary (nonbehav- ioral) in-patient rehabilitation programs but cost only a small fraction of what in—patient programs cost. Mellin, Jarvikoski, and Verkasalo (1984) found no substantial difference between rehabilitation center and out-patient treatment in terms of physical measurements or back pain indices. FOrdyce (1976) said that, historically, treatment for chronic pain had to occur in an in-patient setting to afford staff control of those factors that were thought to reinforce or maintain pain behavior. This treatment, although effective, was costly; the average cost of a 4- to 6-week program ranged from $20,000 to $36,000. Cicala and Wright (1989) compared 25 patients with work—related injuries who were treated as in—patients at the University of Tennessee Center for Pain Management with 25 matched patients who were treated in an out-patient program. hi both cases, the 89 treatment approach was multidisciplinary and involved the same staff; treatment consisted (Hi psychology, physical therapy, occupational therapy, and medical coverage. The single measure of successful treatment was whether the patients returned to gainful employment. Fifty-two percent of the in-patient subjects returned to work, as compared to 44% of the out-patients. The average cost per patient on an in-patient basis was $22,848, whereas the average cost per patient on an out-patient basis was $7,640--only about one- third the cost of an in-patient program. Guck, Skultetly, Meilman, and Dowd (1985) compared an in- patient multidisciplinary pain—management center treatment group of 20 patients to 20 no-treatment control patients. At 1 ix) 5 years follow-up, 60% of the treated patients met all the criteria established for success, whereas none of the untreated patients met those criteria. Treated patients reported less interference with activities, more up-time, lower pain levels, less depression, and fewer hospitalizations than untreated patients. At follow-up, more treated patients than untreated patients reported employment and fewer used narcotic or psychotropic medication. In evaluating relaxation treatment, researchers have tried to determine the role that EMG level plays in the perception of pain. Keefe et al. (1981) found that subjective ratings of pain and tension correlated significantly with EMG level (r = .61), but EMG was not related to pain reduction. Nouwen (1983) demonstrated that EMG levels could be decreased but did not find a concomitant decrease in pain. Large and Lamb (1983) found a correlation of .67 90 between present pain and pre-session EMG. Wolf, Nacht, and Kelly (1982) found that increases cu“ decreases “H1 EMG level beyond a "normal" range produced more pain for their patients. Thus, the evidence suggests that absolute EMG levels might be less important than providing the patient with an active method of coping with and controlling the pain. Turner and Clancy (1986) used 74 patients with chronic low back pain to assess the effectiveness of group out-patient cognitive- behavioral and operant-behavioral treatment. Both treatments resulted in significant changes in types of coping strategies used to deal with pain. In this context, coping refers to thoughts and behaviors people use to manage their pain or their emotional reactions to pain, in order to reduce emotional distress. Two .years later, Turner and Clancy (l988) used 81 mildly dysfunctional patients with chronic 'an back pain iri a study to compare the efficacy of behavioral approaches in treating chronic pain. Subjects were randomly assigned to _an operant—behavioral treatment, a cognitive-behavioral treatment, or a waiting-list control condition. The operant-behavioral and cognitive-behavioral treatments were conducted in eight out~patient group sessions; the results demonstrated decreased physical and psycho-social disability. The operant-behavioral patients showed greater pre- to post-treatment improvement, as rated by the patients and their spouses, than did the cognitive-behavioral patients. The operant behavioral patients leveled off in inmrovement at the (5- and 12- month follow-ups, whereas the cognitive-behavioral group continued 91 to improve over the 12 months following treatment. At lZ-month follow-up, patients in both the operant-behavioral and cognitive- behavioral groups remained significantly improved; 1“) significant difference was found between the two treatment groups. Compliance is an essential concern in fulfilling any treatment. In the pain programs reviewed, patients were instructed to continue to practice relaxation and other exercises at home. Compliance data were reported in several studies, although the method of obtaining such information was usually global estimates or follow-up rating questionnaires, which are considered rather weak. The general conclusion was that patients did continue to practice their assignments, but more infrequently than advised (Linton, l985). The long-term effects of nufltidisciplinary pain cflinic treatment are a significant issue. Long-term fOllow-up data have been given prominent attention iri the literature (Miller & Le Lieuvre, 1982). Since 1983, good ix) excellent nmintenance has been reported; even further improvements have been noted in some follow—up studies. The methodology, in general, has suffered from the usual problems seen in this type of research: low return rates, the use of global ratings and questionnaires, nonblind collection of data, and the like. More work is needed in this area, but there is no compelling reason to believe that the gains documented in the case studies were not generally maintained. Outcome and follow-up studies have stressed iflun; substantial and statistically significant improvements can be obtained through 92 the use of MPCs. Nevertheless, it is extremely difficult to rehabilitate subjects to their prepain level of functioning. Complete medical rehabilitation of zfll patients is Efll unrealistic goal; the alternative is to accept many types of chronic problems. Because a return to the preproblem state usually is not possible, treatment should be oriented to helping the patient live as normally and productively as possible. An alternative is to prevent the development of chronic pain. Predicting the Treatment Outcome of Multidisciplinary Pain Clinics Introduction Diagnostic procedures have become increasingly sophisticated, making it possible to determine the organic basis for many patients’ chronic: pain and in) choose appropriate medical or surgical treatment. However, some patients’ pain does not respond to surgical or medical treatment, and other individuals are not satisfied even after repeated negative medical/surgical work-ups. Therefore, the management of disability/suffering, rather than relief, is a reasonable therapeutic goal. The management of chronic pain allows individuals to learn to lead useful and satisfying lives despite the pain. Patients are able to increase their activity and become more efficient in caring for themselves. They gradually become able in) reduce pain medication and to improve family relationships (Fordyce, l974; Fordyce et al., 1973; Sternbach, 1974). Although many patients improve, others fail. Effort and expense could be spared, and more 93 success could be achieved in the treatment of pain, if referring and admitting physicians as well as team members could distinguish beforehand which patients are most likely to benefit from treatment and which are not. It is in this regard that prediction studies have been undertaken by individuals involved with MPCs. Representa- tive studies of this nature are discussed in the following section. Prediction Studies Maruta et al. (1979) attempted to ascertain which chronic pain patients are likely to benefit from a pain-management program and whether these individuals could be identified before treatment. They studied differences that were discernible at the beginning of treatment; the sample comprised a group who succeeded and did well at l-year follow—up and a group who failed. The researchers established the following admission criteria for patients with pain that resisted treatment: (a) a complete medical and psychiatric evaluation; (b) a pain problem of 6 months’ duration or longer; (c) no related malignant disease; (d) no specific medical, surgical, or psychiatric approach applicable; (e) no litigation; and (f) the patient’s acceptance of the treatment program. Maruta et al. considered selected data for the patients in the study. They looked at personal and clinical history, medical/ surgical diagnosis, subjective pain level at start of program, and score on the MMPI. The researchers analyzed these data by using the two-tailed t-test and the chi-square procedure. 94 The results of Maruta et al.’s study indicated that the success and failure groups did not differ significantly in terms of age, gender, or marital status. No significant difference was found in how many pain-related drugs the patients were taking, or “mether they were receiving disability compensation. The two groups differed significantly on prior duration of the pain, work time lost because of pain, number of prior surgical procedures related to the pain, dependence on medication, and pain level at the beginning of the program. Maruta et al. concluded that the likelihood of success in pain management declined with an increase in prior duration of pain, work time lost, number of prior operations, and level of pain at the beginning of the program. The researchers noted that composite MMPI profiles for women in each group on the Hypochondriasis, Depression, and Hysteria scales were elevated, but even though the elevations for the failure group were greater than those for the success group, the differences were not significant. The composite profiles for men in each group on the Hypochondriasis and Hysteria scales were slightly more elevated than those of women, but the differences in elevations between the success and failure groups were not significant. As a result of the experiment, the researchers concluded that duration of pain and lost work time were significantly longer for the failure group than the success group. Pain of less than 3 years’ duration and lost work time of less than 1 year were favorable indicators. Pain of 5 years’ duration or more and lost 95 work time of 1.5 years or more were unfavorable indicators. The number of operations related to pain was significantly higher in the failure group than the success group. The researchers concluded that three or more surgeries for pain were unfavorable predictors. Zero or one operation was a favorable indicator. Pain level was significantly higher in the failure group than in the success group. Pain levels of 7 or more were deemed unfavorable, whereas levels of 5 or less were determined to be favorable. The higher drug dependency found in the failure group was seen as an unfavorable predictor. The preceding items, along with elevations on the Hypochondria- sis and Hysteria scales of the MMPI, differentiated the two groups. Admission criteria using the above-mentioned information would assist in quick practical assessment and selection (If good candi- dates for multidisciplinary pain-management programs. Block et al. (1980) compared 36 consecutive admissions to a behavioral-medicine program for chronic pain management. They segregated subjects according to source of referral: medical- subspecialties referrals (treating physicians) and disability referrals. Disability referrals came from the state worker’s compensation board, independent rehabilitation consultants, and rehabilitation nurses working for disability insurance companies. The disability-referred patients reported having work-related injuries that prevented their return in) work; they received maintenance settlements and full medical insurance. 96 Block et al. found that both groups reported comparable severity of pain on admission. Both groups indicated a much lower mean level of pain at discharge. The medical-subspecialties- referred group reported El lower severity of pain than the disability-referred group. Analysis of covariance indicated a reliable difference between the groups with regard ix) severity of pain complaints at discharge. Both groups showed a decrease in pain severity, but the disability-referred patients (Thi less well than those referred by medical subspecialties. The researchers concluded that the apparent failure of the disability patients to benefit from the behavioral-medicine intervention was not due to the duration of their presenting complaint; their complaints were of a shorter duration than those of patients referred by medical subspecialties. In contrast, Maruta et a1. (1979) found that failure to benefit from treatment correlated positively with duration of presenting complaint. The major result of Block et al.’s study was that, although all patients experienced a1 decrease ‘Hl pain severityg 'the disability- referred group did not benefit to the same degree as patients referred by medical subspecialties. The authors concluded that the behavior of disability-referred patients in an operant-model pain program could be maintained if positive consequences, such as money, were contingent on the complaint. They contended that this operant notion of pain complaint can provide guidance in making clinical decisions. Patients involved in litigation centering on pain complaints require 97 careful assessment for motivation to engage in treatment. If patients cannot afford to reduce or extinguish the pain complaint, there is little purpose in establishing high expectations regarding treatment; hence, limited treatment goals should be adopted. Painter et al. (1980) stated that MPCs help individuals suffering from chronic pain. Follow-up studies have demonstrated that treatment populations maintained gains for an extended period or experienced only a modest decline. After reviewing individual patients’ records, Painter et al. found that one-fourth to one-third of the sample continued to progress, whereas a similar proportion declined rapidly to preadmission levels of function, despite demonstrable improvement during the MPC program. The authors attempted to discover the reason for this decline because the regression of these patients represented 21 failure of treatment” 'The authors questioned whether’ certain factors were peculiar to individuals who experienced a: decline following treatment because they believed that these patients should not be offered expensive treatment. Painter et a1. grouped causal factors conceptually into four areas: incentives, attitudes, operant factors, auwi psychological variables. Questionnaires were sent to 500 participants in the 1977 program at Northwest Pain Center; 145 of them responded. In addition, telephone interviews were conducted with 10 randomly selected nonrespondents. The researchers compared the 25 most successful patients (success group) with an equal number of patients 98 who had experienced initial success and subsequently regressed (failure group). Painter et al. found that substantial change took place as a result of treatment at 'the IWPC. The total group maintained a significant reduction in pain severity over time. A smaller portion of the sample proceeded to improve, in part because they continued to follow through with the active principles of the MPC approach. For about one-fourth of the patients, substantial gains declined after a few months. Men were somewhat less likely than women to maintain gains. Divorced people were more likely than married patients to continue improving after discharge. The researchers thought the most significant observation concerning whether to treat was the relationship between age and long-term success. There appeared to be a curvilinear relationship in that patients 'hi their twenties were more likely to regress than were those who were over 50. The researchers found that the failure group was more likely than the success group to receive compensation at the time of follow—up. The implication of this finding is uncertain. The authors said the presence of a real disability was itself the cause of continued compensation. However, they also said their findings suggested that a lack of incentive favoring a healthy, pain-free way of life might have caused patients to regress. The authors could not explain their finding by asserting that patients who were most severely disabled were those who did not work, because this was not the case. Success or failure of treatment was not related in the 99 expected way to duration of disability, nor to subjective levels of pain or disability at admission or discharge. Subjects’ educational level did not correlate with overall success, but it was associated with regression. Individuals with less education and correspondingly less opportunity for nonphysical work were more prone to regress after making initial gains. Painter et al. found that indices of depression were strongly associated with success and failure. Their findings indicated that the success group reported having significantly more depression at the time of admission. The authors summarized their findings by stating that the most striking difference between groups was the change they reported in their 'lifestyles following treatment. Individuals who regressed reported very little change in patterns of communication or reinforcement after they left the program, whereas the success group showed considerable change. This suggests that pain—treatment work with families can help reduce operant maintenance of pain behaviors and can improve communication. Keefe, Block, Williams, and Surwit (1981) had as their subjects 111 patients with chronic low back pain who were involved in a comprehensive behavioral-treatment program emphasizing relaxation procedures. The 28 patients who had the greatest decrease in pain were compared to 28 patients who had the least decrease in pain. The findings were discussed in terms of implications for behavioral assessment and treatment. 100 Keefe et al. found that patients who reported the most pain relief had had fewer surgeries, were not on disability payments, and had had continuous pain for a shorter length of time; they also had higher initial pain ratings than the group with the least decrease in pain. MMPI scores were not available (N1 all the patients who were compared. However, the scores that were available (15 for the best group and 10 for the worst group) indicated that MMPI scores were not associated with outcome. In general, the patients tended to have elevated scores on the Hypochondriasis and Hysteria scales. This study as well as that of Painter et al. (1980) suggested that elevated scores on these two scales were not associated with good or bad outcome of behavioral treatment. The authors summarized their study findings by saying that patients with chronic low back pain are not a group of "crocks, losers, or malingerers" for whom treatment is doomed to end in failure. They believed that a growing body of research is defining variables that are predictive of treatment outcome. The results of this study suggested that the likelihood of treatment success may vary, depending on patients’ individual characteristics. Aronoff and Evans (1982) sought to replicate line results of Maruta et al.’s (1979) study. Aronoff and Evans studied 104 chronic pain patients to predict treatment outcome from their MPC. The researchers used 'fifln‘ outcome measures: staff judgment, patient judgment, change in patient scores, and change 'Hi mood. 0f the variables examined in the study, only age was found to be a 101 predictor of outcome. The significant negative correlation of age with outcome indicated that older patients had lower rates of success. Aronoff and Evans believed this finding reinforced the need to screen older patients more carefully. They believed it is important for MPCs to develop a commonly accepted set of instruments such as the MP0 and the Profile of Mood States because they do not depend on the evaluator’s judgment. Use of these instruments would allow a transfer of information regarding the outcome of treatment from group to group. Turner, Robinson, and McCleary (1983) used 135 patients with chronic low back pain in an MPC setting as subjects for their experiment. They evaluated the effectiveness of several measures in predicting response to conservative treatment for these patients. The measures used iri the study included pretreatment demographic variables, scores on the MMPI, patients’ ratings of pain caused by daily activity as measured by the Activities Discomfort Scale, and physicians’ ratings. Turner et al. analyzed their results to determine which pretreatment variables were most highly associated with outcome. They calculated correlation coefficients between pretreatment demographic anul medical variables, physicians’ ratings, patient- completed measures, and patients’ follow-up ratings of pain relief, pain intensity, and return-to—work activities. The only statistically significant predictor of pain relief at follow-up was patients’ scores (N1 the Hypochondriasis scale of time MMPI. The 102 higher the patients’ scores on this scale, the greater their level of dysfunction. The findings of TUrner et al.’s study indicated that patients with high scores on the Hypochondriasis scale were less likely to improve than were patients with low scores on this scale. In her study, Carlsson (1984) divided 58 patients suffering from pain of a nonmalignant origin into three groups. Group 1 consisted of patients whose long-term outcome of treatment could be classified as good, Group 2 included patients who reported at least "good" pain relief at short-term assessment, and Group 3 included patients with ru) favorable effect of treatment. Carlsson used a pain questionnaire designed by the Department of Neurosurgery at the Karolinska Hospital ‘hi Stockholm, Sweden, in) gather data ‘finc the study. Significant differences were found among groups with regard to employment status and location of pain, as documented by Fischer’s Probability Test and the Pearson chi-square procedure. These findings indicated that employment status and location of pain could be used to predict treatment outcome. Carlsson admitted the questionnaire had poor validity; there was a lack of agreement between the questionnaire data and patients’ interview responses. Melzack, Katz, and Jeans (1985) included 145 patients with low back and muscular skeletal pain in their experiment. Using a weighted-rank method in a one-way repeated—measures multiple analysis of variance design, the researchers compared patients who were receiving compensation with those who were not receiving 103 compensation. Melzack et al. found that the two groups of patients had identical pain scores and pain-descriptor patterns. They had similar scores on the MMPI pain triad of Depression, Hysteria, and Hypochondriasis, and were similar on several other personal variables, as well. Patients who were receiving compensation had significantly lower scores (H1 the .Affective and Evaluative dimensions of the MP0 and made fewer visits to health professionals, as compared to patients who were not receiving compensation. Dworkin, Handlin, Richlin, Brand, and Vannucci (1985) studied 454 patients with chronic pain to evaluate the relationship between compensation benefits, litigation, and employment status, and short- and long-term treatment outcome. Univariate analysis indicated that compensation benefits and employment status both predicted poor short-term outcome. When employment status and compensation benefits were analyzed jointly using multiple regression analysis, employment was found in) be statistically significant. Employment significantly predicted long—term outcome in a third analysis; compensation and litigation did not. Dworkin et al. (1984) compared 79 clinically depressed patients with chronic pain to 375 nondepressed pain patients with respect to medical and social history, physical examination, treatment, and treatment response. The diagnosis of depression was based on an assessment by physicians who were pain specialists. The data were analyzed using a two-tailed t-test and the chi- square test of statistical significance. For nondepressed patients, “J. 4.: -;V-r‘~wr”. an .. : . 104 beneficial response to treatment was found to be related to a greater' number (If treatment visits, not receiving worker’s compensation, fewer previous types of treatment, and low back pain. For depressed patients, beneficial response to treatment was found to be related to being employed at the beginning of the treatment program and pain of shorter duration. The authors concluded that depressed chronic pain patients were not very different from those not suffering from depression. Guck et a1. (1986) used 77 chronic pain patients iri a l- to 5-year follow-up study in an attempt to determine whether long—term outcome can be predicted. The pretreatment variables used in the analysis included age; gender; marital status; diagnosis; number of months since onset of pain; receipt of compensation; pending litigation; educational and occupational levels; use of non- narcotic, narcotic, or psychotropicmedications for pain; number of pain—related hospitalizations; and number of pain-related surgeries. Using stepwise (iiscriminant analysis, the researchers found that successfully treated patients were less likely to be receiving compensation, were .younger, were less likely to be taking psychotropic medications, and had undergone fewer pretreatment pain- related surgeries than unsuccessfully treated patients. They also found that, in combination with each other, these five pretreatment variables could be used to classify a significant proportion of the patients (70%). Hurri (1989) undertook a study to compare treatment-group versus control-group subjects on sociodemographic factors, variables 105 related to work, severity of low back pain, and eight clinical measurements. The sample comprised 177 patients with chronic low back pain. The most important predictor (H: the outcome of treatment, as well as of spontaneous recovery, was patients’ scores on the Work Satisfaction Index. The results of the t-test and chi- square analysis suggested that work satisfaction was strongly associated with changes in the subjective functional capacity of patients with chronic low back pain. Hurri concluded that a variety of psychosocial factors might have influenced the treatment outcome for these patients, necessitating particular attention to the social network of the patients’ occupational environment. Bigos et a1. (1991) evaluated factors associated with work- related back injuries. They found that subjects who hardly ever enjoyed their jobs were 2.5 times more likely to report a back injury than were subjects who almost always enjoyed their job tasks. Kleinke and Spangler (1988) used as the subjects of their study 72 patients with chronic back pain from a nmltidisciplinary treatment. center’ in a: 28-day in-patient progranr Subjects were divided into two groups: patients receiving worker’s compensation and those not receiving worker’s compensation. Treatment outcome measures included the MP0, an audio-visual taxonomy, the Beck Depression Inventory, the Profile of Mood States, behavioral ratings by a primary nurse, demographics, and the MMPI. Multiple regression analyses, along with a two-tailed t-test, were used in analyzing the data. 106 The results showed that patients who were receiving worker’s compensation had less favorable scores on treatment-outcome measures at admission and upon discharge from the pain program. However, no significant differences were found between patients receiving and those not receiving worker’s compensation benefits with regard to the amount of improvement on treatment-outcome measures. Although patients receiving worker’s compensation benefited from treatment at ani MPC, they presented a challenge to professionals to develop individualized programs. The researchers believed that individualized programs would bring these patients’ range of scores on performance and treatment outcome closer to that of patients who were not receiving worker’s compensation. The researchers found a significant relationship between MMPI scores and treatment results. High scores (N1 the MMPI Depression and lbsteria scales predicted treatment success, based on measures of improvement from admission to discharge. Table 2.5 contains 21 compilation of information on outcome studies in which a prediction approach was used. Pain Measurement Regardless of individual discipline or scientific focus, pain researchers hold the goal of achieving scientific understanding and clinical control of pain. The development of a valid, reliable, and flexible measure is required to meet the researchers’ goal. Progress in pain measurement has been slow due to the complex 107 mcucoe NF-o ocoz mo» ocoz ozoLm wcadmmw co cmm> F .o>ocaE_ cu >.ox_. mm». «co: _az: ecu co mcomuu>uda o.aua mFmaFcucocoonxz zuF: muco_uma .oma ;u_: co_uadoccou o>_ummoz .mmoouam amouaocm ;u_: voua_00mmo new commmmpflm co .o>o. CFaa 805mm: .me_u Fo sumac. concocm coF cmma macac_ucou ;u_: mucmmumq coF FoFdoc cFma umuumocu .mmuuuaw cu_: voquUOmmw >.mcoLum co_mmocaov Fo moumnc_ md>uwow_. .333; 53 83359.8 93:35 .m.mom “casuaocu vou_e_. .co_um>_uoe co» vommmmmm >d.:wocmo comumm_u_. cm muco_uma .mcomumcwao F-o .Lmo> Fv x803 “meao> vn :Fma o.naco>mu ocaJUm-_;u .Aom_:aoumv co_mmocmoc oda_u.az memaom-_cu .m ocmzcm-_;u .umuu-u ocmDUm-_cu .umou-u m>oc< mcmsvm-_zu .umou-u £3.53 .mcflflmxfi 59 .7? .38 38958.8 woos c_ omcocu .ocoum £3 E 3:20 .ucgmBF “523 .ucuEmt:_ wwaum .mocammoe venouao anacm umcoz ow mamcm> uwon mm 8593 83.:qu Ba 3303 3.69.8 mdmccowoc >ugo_ooamn:m dmu_uoe can 39:30.. 53335 Beacon” £505 95:3 van 3303 8.59.8 c_ma ”.83 :3 occoEcu--mmF c_ma o_c0c;u--qu m_mocma_v .oLaoc--- .m_mocma_u .cu>ma--vF .cmma xomn 30.--qv cmma u_c0c;u--om mou_m oda_udae .cFmQ UFCOczu--om mmu_w o.a_u.:E .c_ma u_coLcu---F meon >camcuux w .comc_nox .coCLJF Amwon mem>m a FFOCOc< FFmon :23 a 83:5 .xuo.m .owvou Fowon cmezmz a .mocom .quc_ma FowoFV cod>au w .eoEocx .xoo.m Acnon cowCoJm w .comszm .auacm: a:-:o.doL madammm cmmmoo mucoeuaacF cmma\z >uaum .comocaam co_uumvoca a mc_m: wo_n3uw oEOUuao--.m.~ odan 108 ”C02 ocoz ocoz ocoz .ucoeo>oca5_ umoe v0305m ucoeuaocu Fo mc_cc_mon uo uoxodaeo mucm_uma oommocaoo .comuacau concocm Fo mm: c_oa can: uca .mma>u ucoeumocu mac -_>oca 8030* .co_uwmcoasoo m.coxcoz 0: .3558ch F0 .305: 833.6 cuw: vo>ocae_ muco_uma nommocaoncoz .uoc u_u co_u -mm_u_. new co_uamcoasou uan .meou -uao Ecou-mcod nouo_uoca acmexodaea “oeoouao Ecou-uco;m coca vauu_uoca 8:28.95 8.... 8:33 85869.8 .oEouuao Ecou-ucozm couaocm >ducou_w_cmmm oouo_voca acoeumocu we mc_cc_mon um ucmE>O.QEo .mucc_uma commocaoo cam .mdmcomm -movoca cudmo; ou mu_m_> 8830* van wmcoum on: o>mumadm>u co o>_uuo*F< .333 was 3533 52859.8 .CFma mo comumuo. .mcmum._n .m> dmcmumdmc: .cmFm u_umoco08Q mdnmco>mv a mo: ucmexodan dawcmmo oLmDUm-_;u .umou-u uo._au-ozu COmeocmoc adamudae .m_m>dmCm oua_ca>_c: vacuum uoz wQCm_cm> Fo mFm>dmcm cum -_Lm>_u.:E woe: -mmms Umuaoaoc 63.30 6059: xcac-vouzm_m3 ocmacm -on w.cowcmma .umou .._nmnoLa uomxo cozum_u 33th was mnm 3 “6.59.8 mucomuma commocawn >.dmo_c_du on acoE>0.QEa .co_umm_umd .comummcoasou commucawn uoc mkm cu oocmaEOu mucomuoa commocawv >ddaumcmdo on :0 3359.88,. .m> 52359.8 uquFo xcouomeFummc: "m aneco .uuowwo Ecou-ucocw u~ asoco .oeoouso Ecou.mco. coca "F anoLu c_ma o_coLcu--qmc c_ma o_coccu--cmq c_ma o_coLcu--qme CFma .muo‘oxm -odaomae new xumn so.-.qu cmm_co ucmc -mmdmecoc--wm Fowon ucaam a .c_.ucm: .c_.;o,¢ .c_xaozo Ammon _uQJCCm> a .ncacm .c_.;u_a .cmdqu: .cmxcozo AchFV acaam a .cmcncmx .c,.;u_¢ .cmxcoxo Amwon mCmon a .~uay .xum~.oz Aqwon cemmdcmu a:-3o..0u 3 38¢ cmmmoo mucoeumocF c_ma\z snaum .vo:c_ucou-..m.~ odan 53'." 0 1(39 .mmoooam ucoeumogu noom .modaom uazz .mu_cnmgmoeou .omgsc >Lme_ga m_ma_guco;uoa>= new co_mmogaoo _azz away-“ >3 mc_ua. .ago_>m;on .moumum uooz co mogoum cm_: .co.a am:v_>muc_ Uo._au-o:u wo o._w0La .ngucm>c_ nommogamo nodo>wn cu mdmco_mmo*oLa vomcad.a;u .commwagmog xuom .Lo_>czon c_ma we >Eoco cmma xumn Ammo—v Lodmcmam wcoz muco_a_oog comuamcoaeou m.LoxLoz o.a_udaz -xwu do:w_>-o_uaa _an uca ._aa .oaz o_coLzo--- a uxc_odx mugamame .mo,c_.u u;m_o .c_ma xomn zed #0 .>u3um >uwgo>om .xgoz cu vanadmg mmdnm cmma we wEOUuao *o gouu_uoga pamu mgmavm-_;u -_Lm> .wLOuum» o_caagm05muomuom gown sod wcoz -Looem umoe mm: comuumwm_uam xgoz .umou-u anoLm dogucoo .w> ucoeummgp o_c0Lzu--~n— Aowowv _LL3: co_u .moucvavwm kuodmg-c_ma “coeummgu -m~w.aumamoc new .mmmgmmgam c_ma -95 .533 089.35 .5. ufioboco *0 .895: .5. .6 330.59.ng ->ma w>mz cu >.oxmd mm». .mec:o> mmm>dmcm *0 mm: .c_ma wo gumcm. ..o>md “emu—V .co_ummcoaeou o>moumg ou xdmxmd ucmcmemgum_n dwcomumoano .m_mocmm_v .co_uwmcwa c_ma nice w .cmedmmx mgmmx m-F wmo. mucmmuma kumwgu >ddavmmouu3m mm_zaoum -soo .msuoum daumgme .gmucom .mm< umcoLco--- .>duau.:xm .xuao a:-:odd0m mudawom cmmmmo mucoeumogp c_aa\z >naum .uoacmucou--.m.~ 0.9m? 110 perceptual experience involved in pain. Pain can be quantified only indirectly. Pain researchers have adopted four major approaches to operationalize and quantify pain. The major approaches can be classified into four groups. They are (3) animal laboratory research, (b) human subjects laboratory research, (c) human physiological correlates, and (d) clinical pain assessment. Clinical pain assessment is involved with not only the evaluation (H: pain relief following pharmacological intervention but, more recently, with the assessment of the human pain experience. Beecher (l959) influenced the field, arguing that quantifying the intensity and duration of pain was not enough. He proposed that the reaction component of pain, its emotional dimension, was the basic issue in clinical pain control. Since Beecher, greater attention has been given to the scaling of pain as opposed to the measurement of pain relief. Measurement of behavior has been used as 21 basis for inference about clinical pain states. Pencil-and-paper test instruments have been developed to quantify multiple dimensions of the pain experience from subjective reports. Clinical pain measurement can be divided into three categories: behavioral measurements, subjective pain reports, and scaling based on word descriptors. Behavioral measurements include observational data and self-reported behaviors. These measures can be reduced to frequency of rate of occurrence. Subjective pain reports include 111 category judgments and visual analog scales. In category judgments, subjects are given a structured categorized scale and asked to rate each stimulus on that scale, which usually indicates pain intensity. The visual analog scale (VAS) requests that the subject indicate the intensity of his or her pain by marking a 10 cm line labeled no pain" at one end and "the worst pain possible" at the other. Investigators equate the length of the line produced with the estimate of pain to produce a number from T to TO to indicate pain intensity. A variation of this concept is for the subject to categorize the level of pain numerically. Usually "no pain" is at one end (l) and "the worst pain possible" is at the other (TO). The subject records the level of pain perceived. Scaling based on word descriptors (adjectival) was pioneered tux Melzack and Torgerson (l97l) and refined by Melzack (l975) when he introduced the McGill Pain Questionnaire (MPQ). This instrument is reviewed in depth in the following section. I_e McGill Pain Questionnaire Introduction. Achieving scientific understanding and clinical control of pain is the common goal of pain researchers. This goal holds true regardless of the researcher’s scientific focus or individual discipline. To attain this objective, a valid, reliable, and flexible pain-measurement instrument ‘h; required. Measuring pain in this way is an attempt to quantify an intricate perceptual experience. 112 Pain is a complex perception as opposed to a simple sensation. Individuals suffering from pain use language as a medium to relate their perception of pain to significant others. Scientific measurement procedures have focused on pain as an individual sensory quality that varies only in intensity. However, although intensity is a noteworthy dimension of pain, to describe pain only in terms of intensity is like describing the visual world only in terms of light, failing to consider color, pattern, texture, and other dimensions of the visual experience. The complexity of pain extends beyond the Inultiplee dimensions. of sensation; affective and motivational aspects must also be recognized. Failure in) consider the motivational-affective dimension of pain seriously restricts the total picture of the pain experience. The motivational dimension is crucial to the concept of pain as a perception, recognizing past experience, attention, and sensory quality and intensity. If researchers consider (NH)! the sensory features of pain, they are ignoring the most important part of the pain process-—the motivational and affective properties. Melzack (l983) summarized the role of motivation iri the pain process as accounting for the multidimensional properties of pain experience and behavior. He wrote: (a) The sensory-discriminative dimension of pain is determined primarily by activity in the rapidly conducting spinal systems; (b) the powerful motivational drive and unpleasant affective characteristics of pain are subserved by activities in reticular and limbic structures that are influenced predominantly by the slowly conducting spinal systems; and (c) neocortical or higher central nervous system processes, such as evaluation of the input in terms of past experience, exert 113 control over activity in both the discriminative and motivational systems. Melzack (l983) assumed that the three categories of activity interact to provide perceptual information concerning location, magnitude, and spatiotemporal properties of in“; noxious stimulus; motivational tendency toward escape or attack; and cognitive information based on analysis of multimodal information, past experience, and the probability of outcomes of different alternative responses. The three forms of activity may influence motor mechanisms that account for the complex pattern the patient (Melzack, l975). The patient is to select only those words that describe his or her pain at the time the questionnaire is being administered. It. is important that the subject understand the meaning of the words in the questionnaire; in turn, the person administerjrm; the instrument should be patient and understanding (Melzack, l975). Melzack found that two things became apparent in the administration of the MPQ. First, subjects were highly selective in choosing and rejecting words. Second, subjects were grateful to be provided with words to describe their pain. Klep, Dowling, Rokke, and Schafer (l98l) evaluated MPQ profiles according to administration format: researcher-administered versus self-administered. No significant difference was found between the two modes of administration. Reliability. Reliability means that an instrument measures the same thing consistently. Pain instruments must be shown to yield reproducible data. Reliability considerations become problematic when the variable under study, such as pain, is subject to variation across time. Assessing the reliability of pain scales is confounded by the subject’s memory capacity; that is, the subject might recall the pattern of responses of an earlier occashwi. Also, the pain experience has an inherent fluctuating quality (Reading, l983). Because the MPQ requires a number of judgments for each dimension under study, the reliability (H: the scores obtained (N1 the instrument is likely to be increased over those obtained through the 120 administration of an instrument with a single rating scale (Reading, l983). Grahman, Bond, Gerkovich, and Cook (T980) found the MP0 to be a major step toward quantifying the subjective aspects of pain. Their repeated administrations of the MPQ to cancer patients resulted in a consistency index of 75%, within a range of 35% to 90%, between the ‘first two administrations. The results came fWTMT both single and multiple administrations of the MPQ in two samples, each composed of TB cancer outpatients who were in pain. These results are comparable to those reported by Melzack (T975), in which the consistency of word-choices by T0 cancer patients over 3 days ranged from 50% to l00%, with a mean of 70.3%. These findings provide support for the use of the MPQ as a reliable, multidimensional measure of pain. Hunter, Phillips, and Rachman (T979) studied patients’ ability to remember and report their pain consistently. They administered the MPQ to T6 patients who were experiencing pain as a result of a neurosurgical procedure, such as ea lumbar puncture (n: myelogram. None of the subjects had an organic disorder that would have impaired their ability to recall. The researchers assessed pain recall by administering the MPQ after an interval of l 'U) 5 days. The results support the reliability of patients’ retrospective pain reports, as indicated by a high consistency in score profiles from the three administrations of the MPQ. 121 Validity. Validity is the extent to which a test measures what it is supposed to measure. A number of studies have been performed in a variety of clinical settings including the hHH) as a dependent measure. This testifies to its acceptability in clinical assessment. Face validity. Kremer, Atkinson, and Ignelzi (l98l) evaluated chronic pain patients’ reports of pain intensity on the Visual Analog, Numerical, and Adjectival scales of the THKL The results indicated that all patients were able to complete the Adjectival scale, but lT% were unable to complete the Visual Analog scale. There was a 2% failure rate on the Numerical scale. The patients who completed all three scales indicated a significant preference for the Adjectival scale. Results of a chi-square test indicated that this preference was reliable (X2 = 7.56, p < .025). The basis for this preference did not appear to be related to gender, etiology of pain, affective variables, or selected psychological variables. The result of the study testifies to the MPQ’s acceptability. This evidence serves to support the face validity of the MPQ. Construct validity. Melzack and Torgerson (l97l) addressed construct validity through the organization of words describing the different aspects (H: the pain experience into classes and subclasses. They first presented the major classes and subclasses of words 11) 20 subjects with college educations. Subjects were asked whether each word belonged in the subclass. Words with less than 65% agreement were used in a forced-choice test with 20 additional subjects, who were also asked to assign each word to a 122 category. A second study was done with three groups of subjects: T40 introductory psychology students, 20 physicians, and 20 patients. Melzack and Torgerson were then able to place words in each category in rank order on the basis of the wean ratings or scale values of each set of words for each of the three groups. There was substantial agreement that the words fell into particular categories or classes. Melzack (T975) obtained data from 29 subjects who participated in a study of the effects of brief, intense electrical pain. Highly significant intercorrelations were found between PPI percentage changes and the percentage changes for each of the PRI indices, as follows: Sensory = .9l, Affective = .82, Evaluative = .96, Miscel- laneous = .92, and Total = .94. Melzack believed the MP0 provides valid indices of some of the dimensions of pain and that the ques- tionnaire can be used to determine the effects of different thera- peutic manipulations. Crockett, Prkachin, and Craig (T977) attempted to determine empirically the nature and minimum number of dimensions necessary to describe responses to the MPQ. They compared two groups; one group received experimentally induced pain, and the other was composed of referrals to a back pain diagnostic clinic at a general hospital. Using 2”] unweighted least-squares solution, the researchers found five factors to be significant: (a) immediate anxiety, (b) perception (rf harm, (c0 somesthetic pressure, 01) cutaneous sensitivity, and (e) sensory information. These factors overlapped 123 considerably with Melzack and Torgerson’s (l97l) a priori classification of pain descriptors. The similarities were found in the factor groups loaded with items from Melzack and Torgerson’s Affective and Sensory dimensions. Leavitt, Garron, Whistler, and Sheinkop (T978) provided further support for the dimensions postulated by Melzack. Using T3T patients from the clinical services of three neurosurgeons and five orthopedic surgeons, the researchers determined the frequency with which particular words were used in describing back pain. A principal-component analysis with varimax rotation was applied to the correlation matrix for back pain. Seven interpretable factors were obtained, accounting for approximately 76% of the variance. The preponderance of affective-evaluative descriptors in two factors and specific sensory qualities in the remainder provide support for Melzack and Torgerson’s general taxonomy of pain. Reading (T979) administered the MPQ to T66 women complaining of dysmenorrhea vHu) were attending gynecological clinics. Subjects’ responses to the MP0 were subjected to factor analysis, and four factors were derived. Two of these factors reflected sensory qualities of time pain experience. Two factors relating to ‘the reaction component of the pain also were noted. Reading’s findings support the distinction between the sensory and reaction components of the pain experience. Prietro et al. (T980) reported the results of oblique and orthogonal rotation of factors for a sample of patients with back pain. In analyzing subjects’ responses to the MPQ, the researchers 124 found that four factors accounted for 5l% of the total variance in responses. Three of the factors were composed solely of the Sensory, Affective, and Evaluative descriptor subclasses. The fourth factor was defined by both the Sensory and Affective subclasses. The researchers asserted that their study was based on appropriate statistical methodology, overcoming the deficiencies of previous investigations. They viewed the results as providing strong support for Melzack’s (T975) three-factor conceptualization of the MP0. The close correspondence between the factors in their study auui Melzack’s classifications of pain descriptors provides assurance that investigators and practitioners may continue to use Melzack’s Sensory, Affective, and Evaluative pain rating indices. McCleary, Turner, and Dawson (l98l) examined the relationship between measures of emotional disturbance and the dimensions of the pain experience, in an attempt to discover the nature of the principal dimensions of pain. ‘They used the THU) to quantify the pain experience. Results (H: a factor analysis indicated that the relationships between pain descriptors and indicators of emotional disturbance were consistent. 0n the MPQ, patients with excessive somatic concern described their pain 'Hi terms (Hi evaluative and affective descriptors, whereas those with significant somatic concern, depression, and hysteria described their pain as frightening. Reading (T982) factor: analyzed the FUN) scores obtained from women experiencing acute postepisiotomy pain. A comparison of these 125 patients’ scores with those of chronic pain groups revealed specific sensory qualities and combined emotional-sensory dimensions. The results suggest that acute pain might involve less differentiation of the Sensory, Affective, and Evaluative dimensions than does chronic pain. The distinction between sensory and affective subgroups was confirmed in the above—mentioned studies. The results lend support to the practice of deriving scale scores. These researchers also demonstrated, but with less consistency, the existence of an evaluative component. Researchers investigating the relationship between MPQ scores and concomitant assessments of the psychological state also have addressed the construct validity of the MPQ. Kremer and Atkinson (1981) used a population of chronic pain patients to examine the relationship between subjects’ scores on the Affective dimension of the MPQ and independent measures of affect and infirmity. The major finding was that the Affective dimension of the MPQ appeared to have good construct validity. Subjects who had high scores on the Affective dimension of pain reported significantly greater scores on depression, anxiety, and somaticization than subjects who had low scores on the Affective dimension. This finding demonstrates that the Affective dimension of the MPQ provides a valid summary of the affective status of this population, supporting the interpretation that the Affective dimension has good construct validity. McCleary et al. (1981) studied MMPI profiles in relation to MPQ scores. In a multiple-regression analysis, the researchers used the 126 MMPI results. as criterion variables zuui the MPQ scales as predictors. McCleary et al. found that scores on the Affective dimension contributed to the prediction of MMPI profiles. They pointed out that patients who experienced more emotional disturbance tended to describe their pain as more intense, frightening, unbearable, and burdensome. The Affective dimension was related to signs of emotional disturbance, independent of descriptions of pain intensity. Pearce and Morely (1989) examined the construct validity of the MPQ through the use (M: an experimental technique attempting to bypass conscious processing. Researchers performing the Stroop task show subjects a series of colored words printed in another color. Red would be printed in blue ink, and subjects would be asked to name the color in which they are printed. If the subject is to name the correct color name, the competing response must be inhibited by the attentional process. Active inhibition is achieved at the expense of attentional resources. 'This reduces the attentional capacity available for processing the correct response, resulting in a slower response time. Pearce and Morely (T989) hypothesized that the Stroop interference effect would be most marked for pathological groups in tasks that use words that activate the predominant accessible constructs. They contended that the Stroop procedure may be used in an experimental analysis of the MPQ’s construct validity. 'The researchers predicted that chronic pain patients would show greater 127 interference on words representing the pain construct, on the assumption that the construct is chronically activated in pain patients. Pearce and Morely predicted that words representing the affective component of pain should produce greater interference than sensory words, on the assumption that one characteristic of chronic pain patients is the dominance of negative-pain-related affect. The differences between chronic pain patients and controls on the sensory and affective Stroop tasks supported the prediction that the pain group would have more cognitive representation of sensory and affective components of pain in comparison to the control group. Pearce and Morley concluded that chronic pain patients are more susceptible to interference effects when stimuli are pain related. Thus, through the use of an experimental technique from cognitive psychology, further validity was attributed to the MPQ. Concurrent validity. The concurrent validity of the MPQ has been supported in :1 number (If clinical trials. Reading (1979) investigated the concurrent validity of the MPQ by comparing the pain scores of women relying on pain killers and those denying the regular use of analgesics. The use of analgesics was related to the total questionnaire score, as well as 11) the score (”1 the factor composed of Evaluative and Affective dimensions, reflecting the reaction component of pain. Reading suggested that the reaction to pain (in terms of taking medication) was more closely related to the significance attached to the pain, as indicated by reaction words, than it as to the level of sensation experienced. 128 Van Buren and Kleinknecht (1979) used the 10%) to investigate pain reports following oral surgery (extraction of the third molar). The researchers found that higher scores on the Sensory, Affective, and Evaluative dimensions and on the PPI rating were related to longer recovery times and increased use of both narcotic and non- narcotic pain medication. The findings support the contention that all four subscales of the MPQ tap some responses indicative of the pain experience. Klepac, Dowling, and Hauge (1981) supported the validity of interpretations drawn from MPQ differences in clinical practice and research. They found that subjects describing cxflci presser pain (subjects immersed their left hands iri a container (H: ice water maintained at 2 +7“— 1 degrees Centigrade through refrigeraion and addition of ice chips) had significantly higher scores than did those exposed to tooth shock (electrical stimuli delivered to healthy Inaxillary incisors in low' current levels and increments required fin: tooth-pulp stimulation). Those subjects describing tooth shock received two ascending series of electrical shock below sensation threshold and in irmweasing increments through reported sensation threshold and to pain threshold levels. At this point, half of the subjects were stopped and asked to complete the MPQ to describe the most intense sensations they had experienced; the other half were continued to pain tolerance, after which they completed the MP0. Similar procedures were utilized with subjects describing cold presser pain. Half of the subjects removed their hands at pain 129 threshold and completed the MPQ. The other half continued to tolerance before completing the MPQ. Subjects who described tolerance levels of stimulation scored higher ("1 all measures than those exposed to threshold levels of pain (Klepac et al., 1981). The differences reached acceptable levels of statistical reliability on the Sensory, Evaluative, and Total dimensions, and on the PPI rating. The results clearly supported the validation of the MP0. The researchers found that pain threshold, as compared to pain tolerance, was associated with lower scores on two of the three global measures of intensity (PRI and PPI), as well as on three of the four subscale scores. The findings supported the utility of the MPQ in describing intensity and qualitative differences in pain, in both laboratory and clinical studies. The usefulness of the MPQ in comparing the various types of pain of interest to clinicians and researchers was also reinforced by the results of this study. Hunter and Philips (1981) obtained additional support for the concurrent validity of the MP0. They found significant correlations between subjects’ MPQ scores and diary-card ratings of the intensity and duration of headaches. The total intensity of pain quality and the affective response ix: pain were significantly associated with the intensity and duration of headaches. The sensory qualities of the experience and pain evaluation were associated only wfith pain intensity. Discriminant validity. Reading’s (T981) comparison of MPQ profiles of women experiencing pelvic pain supported the 130 discriminant validity of the MP0. He found that patients with acute pain, as compared to those with chronic pain, displayed greater use of sensory word groups, testifying to the pronounced sensory input from the damaged perineum. The chronic pain patients, in contrast, used affective and evaluative word groups more frequently. The results of this study reflect the diagnostic potential of the MPQ. This is consistent with Reading’s clinical observations that subjects will display distinctive score profiles according to the nature of the pain. Thus, the MP0 is efficient in distinguishing among patient groups. Conclusions. The MP0 has been found to have acceptable reliability, as well as face, construct, concurrent, and discriminant validity. Since its introduction in 1975, the MP0 has been recognized as a useful instrument with which to obtain data on both qualitative and quantitative aspects of pain. A major advantage of this instrument is that it scales pain multidimensionally. The widespread acceptance of the MP0 is demonstrated by its translation into Finnish (Ketovuori & Pontinen, 1981), German (Stein &. Mendl, 1988), Arabic (Harrison, 1988), and Dutch (Vanderiet, Adriaensen, Carton, & Vertommen, 1988). The relatively recent translations of this instrument have not IMHNT as extensively supported as the English version. The Department of Health and Human Services of the Commission on the Evaluation of Pain (1986) reviewed many of the principal 131 measures of pain behavior. Special attention was given to the MP0. The Commission reported that the MPQ is a widely used and accepted instrument with which to address the sensory, cognitive, and emotional aspects of pain. In a report by the Institute of Medicine (1987), the MPQ was described as "perhaps the most thoroughly evaluated multidimensional scaling device for pain." It is the standard by which other measures of subjective pain states are compared. The Affective Dimension of the McGill Pain Questionnaire Historically, psychological distress in chronic {xrhi patients had been viewed as the cause of the pain in these individuals. Excessive somatic concerns and personality traits commonly have been named as a cause of chronic benign low back pain. However, research has not supported this point of view. White and Gordon (1982) found these same characteristics in individuals with objective causes of back pain. Waddell, Pillowsky, and Bohn (1989) found that subjects with abnormal illness behavior had higher levels of objective physical impairment, pain, disability, psychological distress, time off from work, and previous surgeries than individuals with normal illness behavior. This finding supports the belief that psychological distress, abnormal illness behavior, and signs appear secondary to physical problems. Taking this line of inquiry a step further, Keefe, Crisson, Urban, and Williams (1990) found that patients having few physical findings showed the fewest pain behaviors. Conversely, patients with more physical findings exhibited time most pain behaviors. This suggests that the 132 combination of psychological distress, physical complaints, and physical findings is 1%”: more complicated than sttoricalTy perceived. 1 Findings (if the previously' mentioned research indicated the importance of evaluating psychological or affective distress. The Affective dimension of the MP0 is an effective tool for this purpose. It holds significant interest and implications for future research. The Affective dimension of the MPQ has been evaluated in a variety of clinical settings. Research results have demonstrated that pain patients with a medical diagnosis suffer as much affective distress as patients without identifiable organic etiology (Stanton, 1990). Researchers using the MPQ have been unable to discriminate between medical and psychiatric diagnoses (Kremer, 1983). The most reliable use of pain language, then, might be as an index of affective distress. Kremer (1983) demonstrated that affective descriptors of pain are better predictors of psychological distress than are patterns of sensory and affective descriptors. The also found that the MPQ measure of affect is important because it is part of a pain- assessment instrument and therefore is not presented as a measure of psychological distress. Many chronic pain patients are sensitive to any implication that their pain complaint is mental, and using an obvious measure of psychological distress could alienate them. Because it is not obvious what the Affective dimension of the MPQ is assessing, it can be of considerable clinical utility. 133 McCreary (1981) found that patients with high scores on the MMPI Hypochondriasis scale described their pain as more intense and as high in terms of affective and evaluative descriptors on the MPQ. The MMPI Depression and Hysteria scales, reflecting emotional disturbance, were also associated with a more intense description of pain and higher scores on. the Affective dimension of' the MPQ. Aronoff and Evans (1982) provided further support for the effectiveness of the Affective dimension of the MPQ as a measure of emotional distress. Their results corroborated the finding that the Affective and Sensory dimensions of the MPQ correlated with the MMPI Hypochondriasis scale. They further stated that the MMPI Depression scale was related only to the Affective dimension of the MPQ. Affect and Worker’s Compensation The combination of affective distress and receipt of worker’s compensation has been investigated in a number of studies. Mendelson (T982) and Pelz and Mersky (1982) found that patients with low back pain could not be categorized neatly into worker’s compensation and non-worker’s-compensation groups. Both groups, worker’s compensation and non-worker’s compensation, had higher levels of psychological disturbance than the normal population, but they did not differ from each other. In contrast, Melzack (1985) found that patients with low back pain who received worker’s compensation had significantly lower scores on the Affective dimension of the MPQ and fewer visits to health professionals, as compared to their counterparts who received no compensation. 134 Dworkin (T985, 1986) studied chronic pain cases and found that compensation predicted poor short-term outcome. In his 1986 study, Dworkin found that nondepressed individuals who were not receiving worker’s compensation improved with a greater number of treatments than did depressed individuals. Kleinke and Spangler (1988) found that elevations on the MMPI Depression and Hysteria scales predicted positive treatment success. They also found exceptions to this generality, leading them to conclude that recipients of worker’s compensation must be regarded individually. Similarly, Block et al. (1980) found that patients who were in litigation needed 1x1 be assessed carefully for motivation and also needed to have limited treatment goals. Summary This chapter contained a review of literature related to the physiology of chronic pain, the transmission of pain signals, and theories regarding pain. The writer also discussed the results of multidisciplinary treatment programs for low back pain and research on the prediction of treatment outcomes of nmltidisciplinary pain clinics (MPCs). The McGill Pain Questionnaire was reviewed in depth, as was the Affective dimension of the McGill Pain Questionnaire (MP0), and the relationship of affect to receipt of worker’s compensation. Results of outcome studies using an operant approach, the relaxation approach, the cognitive approach, a multimodal approach, and £1 prediction approach were summarized in tabular form. CHAPTER III METHODOLOGY Introduction The primary purpose of this study was ix) determine the existence and nature of any relationship between patients’ scores on the Affective dimension of the MP0 and the outcome of an MPC treatment progrmn. A secondary purpose was to determine what combinations (Hi MPQ .Affective score, demographic variables (age, educational level, spouse’s employment, marital status, employment status at intake, involvement in litigation, involvement in rehabilitation, and sources of income), and historical variables (length of chronicity, number of hospitalizations, number of surgeries, and number of past employers) best predict the outcome of treatment. The methodology used in conducting this observational study is described in this chapter. First, the research questions and hypotheses are restated. The study site and the treatment process are then discussed. Next, the sample-selection procedure is eXpTained, and the sample is described. The instruments employed in the research are discussed in detail. The data-collection and data- analysis procedures used in the study also are presented. 135 136 Research Questions The following research questions were posed to guide the collection of data for this study: 1 1. Will subjects’ scores on the Affective dimension of the MP0 correlate positively with successful outcome following treatment in an MPC? 2. Will the MPQ Affective dimension scores of subjects in the worker’s compensation group, tine group covered by no-fault automobile insurance, and the group receiving TH) financial support differ over the course of treatment: at intake, upon completion of the program, through follow-up? 3. Is there a difference in the treatment-success rates of the three groups (those receiving worker’s compensation, those covered by no-fault automobile insurance, and those receiving no financial support)? Research Hypotheses The following hypotheses were formulated to guide the analysis of data gathered in this study: Hypothesis 1: Chronic pain patients’ scores (”1 the Affective dimension of the MP0 at intake will correlate positively with successful outcome following treatment in an MPC. Hypothesis 2: Subjects with high scores on the Affective dimension of the MPQ at intake will have a nonlinear pattern of Affective dimension scores over the course of treatment, upon completion of the program, and at follow-up. Hypothesis 3: There will be no difference in the treatment- success rates of ‘the three study groups: those receiving worker’s compensation, those covered by no-fault automobile insurance, and those receiving no financial support. 137 Description of the Study Site and the Treatment Process Mary Free Bed (MFB) Hospital and Rehabilitation Center in Grand Rapids, Michigan, was the site of the study. MFB is a large, nonprofit. hospital and rehabilitation center that draws patients from all of western Michigan. MFB’s pain program is multidisciplinary. .A team approach is used; the team includes an orthopedic specialist, a family physician, a physiatrist, a physical therapist, a, psychologist, and a program Inanager. Because the chronic pain problem is so complex, a treatment approach is required that incorporates the various areas of medicine to provide patients with the tools they need to gain control of their pain. Patients are referred to MFB by a physician or a third-party insurance carrier. Selection of patients for the MFB pain program is a critical process. Prospective patients undergo a comprehensive evaluation designed to determine whether they have ongoing medical problems that can account for the pain. If a medical basis for the pain is identified, an appropriate referral is made. The prognosis for treatment is then determined. If it is decided that the patient is a candidate for MFB’s program, the duration and intensity of treatment are recommended. After a patient is accepted into the pain program, team members from the various disciplines converge to emphasize individual education. The average length of treatment is 4 to 6 weeks, but the program is designed to meet individual needs as assessed by the team. Some patients are seen as infrequently as once a week for 6 138 weeks, whereas others are seen as intensively as twice a day for 2 weeks. The patient’s visit to the clinic might consist of physical therapy, group or individual sessions with the team psychologist, biofeedback sessions, and education on body mechanics. Each patient’s improvement is evaluated weekly by the team in a conference designed to monitor the patient’s progress and to modify the treatment plan according to the individual’s needs. The pain program is viewed as a learning experience designed to help the patient develop independence. The patient learns exercise regimens, physical therapy, and body mechanics, as well as how to apply behavioral-medicine strategies and psychology in learning to cope with the pain. The ultimate benefit is realized 3 to 6 months after' completion of‘ the progranl when the patient applies these modalities in daily life outside the program. Selection and Description of the Sample Patients were obtained from the chronic pain program of Mary Free Bed Hospital and Rehabilitation Center (MFB). All records of patients from 1986 through 1990 were reviewed to identify patients who would be appropriate subjects for this study. Beginning from a list of 105 patients, the researcher applied several selection criteria. Patients selected for the study had to have suffered from low back pain. Low back pain was defined as pain originating or centering at or around the lumbar or sacral area of the back. The pain had to be of at least 6 months’ duration before the patient 139 began treatment. The subjects had to be voluntary participants in the program. A favorable recommendation had to have been made by the multidisciplinary team to allow the client to participate in the program. Medical records had to indicate that no additional medical treatment was needed. Successful completion of the program was required for the patient to be included in the sample. The subjects had to have completed the program at least 6 months before the follow-up procedure. Further, a review of the charts had to demonstrate that the MPQ had been administered at intake, at least three times during the course of treatment, and at least once through follow-up. The chart review reduced the number of subjects to 85. A letter was sent to each subject in the sample (Appendix B), describing the study and advising them that their participation was voluntary. MFB’s follow-up questionnaire was enclosed, along with the Productivity of Life Questionnaire. ll stamped, self-addressed envelope was attached for the subjects’ convenience. Approximately 30 days after the initial follow-up letter was sent out, a second request, duplicating the previous mailing, was sent. After approximately 30 days, the researcher attempted to telephone the subjects who had failed to respond to the nailed request. Three individuals declined to participate in the study. An additional seven subjects declined 11) provide fOTlow-up information over the telephone or could not be contacted. 140 Of the 85 patients in the sample, 75 provided responses to the follow-up questionnaire. This resulted in a response rate of 88%. The 75 subjects were divided into three groups for comparison purposes. Forty-three subjects were receiving worker’s compenSation benefits, 11 were receiving automobile no-fault benefits, and 21 were not receiving any form of subsidized financial support. The sample was representative of other follow-up studies (Guck, Skultety, Meilman, & Dowd, 1985; Turner & Clancy, l986) and prediction studies (Aronoff & Evans, 1982; Carlsson, l984; Guck et al., 1986; Keefe, Block, Williams, & Surwit, 1981; Kleinke & Spangler, l988; Maruta, Swanson, & Swenson, 1979; Painter, Seres, & Newman, 1980) when considering variables of age, education, gender, marital status, length of chronicity, and number of surgeries for pain. This sample was similar to Keefe et al.’s (T981) sample, which the researchers described as fitting "most of the characteristics commonly ascribed 11) the chronic low back pain syndrome in North America" (p. 233). Demographic characteristics of the sample are presented in Chapter IV. Instrumentation The instruments used in the study are discussed in this section. The three instruments used in the research were the McGill Pain Questionnaire, the Productivity of Life Questionnaire, and a follow-up questionnaire used by MFB. 141 The McGill PainEQuestionnaire The McGill Pain Questionnaire was reviewed in depth in Chapter II. A summary of its reliability and validity is given in the fol- lowing paragraphs. The MPQ is a reliable instrument (Hunter & Philips, 1981). As a result of their study, Hunter and Philips demonstrated the ability of patients to remember pain and report it consistently. The MPQ has been found to have face validity (Kremer, Atkinson, & Ignelzi, T981). Construct validity has been demonstrated by a number of researchers (Kremer & Atkinson, 1981; Leavitt, Garron, Whisler, & Sheinkop, 1978; McCreary, Turner, & Dawson, 1981; Prieto et al., 1980; Reading, 1979, 1982). The MPQ has been used in a number of clinical trials to provide information that supports its concurrent validity (Buren & Kleinknecht, 1979; Hunter & Philips, 1981; Reading, 1979, 1981, 1982). Discriminant validity has been shown in distinguishing between patient groups (Dubuisson & Melzack, 1976; Reading, 1982). The affective dimension of the MPQ was used to measure the level of affective distress acknowledged by each individual patient. Patients’ scores on the MP0 Affective dimension were obtained at intake, weekly throughout the course of treatment, and at follow-up. The Productivity of Life Questionnaire To measure the subjects’ level of productivity, the Productivity of Life Questionnaire developed by DeJong and Hughes (1980) was used. (A copy of the Productivity of Life Questionnaire 142 is contained in Appendix C.) This questionnaire was one of two outcome measures used in the study. DeJong and Hughes developed the Productivity of Life Questionnaire (PLQ) to evaluate a range of rehabilitation outcomes among persons with spinal cord injury (SCI). Traditionally, programs have been judged solely on the number of patients who are able to return to work following treatment. As important as this is, it is an incomplete criterion. DeJong (1981) pointed out that many disabled people make uncompensated, yet substantial, contributions to both home and community. In an effort to evaluate people’s ability to live more productively following rehabilitation, the PLO was developed. Productivity was evaluated in terms of contributions made to family and community life, not solely in terms of gainful employment. The questionnaire examines five dimensions of productivity by evaluating respondents’ participation in (a) gainful employment—~whether they worked full time, part time, or not at all; (b) homemaking--includes one or more of meal preparation, house-cleaning, food shopping, and supervising children or dependent adults; (c) school or educational programs--full time, part time, or not at all; ((1) formal organizations-~inc1uding both disability- related organizations and nondisability organizations, i.e., the local Rotary Club, school, PTA, fraternal organizations, and so on; (e) leisure-time activities, active-—where the individual leaves home to attend recreational events or to visit friends, and passive 143 -—where the individual remains at home to watch television or have friends visit. Consideration was given to the degree of participation in each activity. The individual’s productivity outcome was defined according to the activities in which the person participated and the degree of participation. The best possible outcome is an individual’s participating as fUlly as possible ir1.all activities. The 'worst possible outcome is the individual’s failure to participate in any of the activities except a form of passive leisure. The results for the 111 persons in DeJong and Hughes’s (1980) study group were collapsed into 12 outcomes, based on similarity and frequency of occurrence. A 42-member panel from the Massachusetts Interagency Council on Independent Living (ICIL) ranked and weighted the outcomes. The ICIL members ranked outcomes involving employment or school as the most productive; those involving formal organizations or homemaking were ranked the next most productive. Outcomes involving only active leisure or passive leisure were ranked the lowest. The productivity outcomes were classified as most productive, moderately productive, and least productive. (The rankings and weightings of productivity outcomes are shown in Appendix D.) To determine whether there were any significant differences in the weighting of outcomes among ICIL members, data were analyzed based (”1 members’ age, gendery and disability status. No statistically significant differences in weightings were found, 144 based on age and disability status. There was a difference on the basis of gender. Women tended to give more weight to productivity outcomes involving homemaking when it was the major nonleisure-time activity than did men. These results indicate that the PLQ has satisfactory interrater reliability (DeJong & Hughes, 1980). The Follow-Up Questionnaire The follow-up questionnaire used by MFB is regularly sent out at l-month, 3-month, 6-month, and lZ-month intervals. The questionnaire contains items asking the former patients whether they were working at the time of discharge, are presently working--if so, list employer, type of job, length of employment, and hours per week. If they are not working, subjects are asked whether they are involved in a job club or job search. (A copy of this questionnaire may be found in Appendix E.) A special section of four questions relates to worker’s compensation or automobile no-fault coverage. These questions relate to vocational rehabilitation and the type (fl: training, if any, with which subjects were involved. The clinic was also concerned with whether the subjects had obtained second-injury certification. The questionnaire contains the MPQ. MFB also questions the individuals about whether they are using the stress-reduction techniques taught in the program. They are also queried as to the level of depression, if any. The follow—up form has questions on pain medication, physician treatment, runv pain problems, and 145 surgery. There is 21 section (NT flexibility, exercise, and body mechanics. The questionnaire concludes with a section on participa— tion in light cleaning, meal preparation, laundry, heavy cleaning, and yard work. Return to work was defined as an individual’s having success- fully obtained employment and maintained it for at least 3 months. Participation in rehabilitation was defined as :1 subject’s taking part in a structured vocational rehabilitation program with a spe— cific vocational goal. Subjects’ return to work and participation in a rehabilitation program constituted a dichotomous outcome measure for this study, while the Productivity of Life Questionnaire provided a continuous outcome measure. Data—Collection Procedures Before beginning to collect the data for this study, the investigator submitted the research proposal 1x1 the Michigan State University Committee on Research Involving Human Subjects (UCRIHS). Upon receiving approval from UCRIHS (see Appendix F), the researcher sought and obtained permission to conduct the study from MFB’s Education and Research Committee and the hospital’s Human Subjects Review Committee. The initial predictive data were gathered fflwmi the Pain Rehabilitation Progranl Attachment (see Appendix G), contained in patients’ files at MFB. Patients’ scores on the MP0 (which had been administered weekly during treatment, as well as demographic and historical data, were obtained by reviewing this attachment. 146 Information obtained from the chronic back pain clinic and that received from the subjects was recorded on a data summary form (Appendix FD. All data obtained from these procedures were coded and recorded by medical number only, to protect patients’ anonymity and to ensure that their responses would remain confidential. Patients’ names were not used in this study. Data-Analysis Procedures Correlations were calculated among many subject characteris- tics, including scores on the MP0 and treatment outcome. The formula used for the correlation coefficients is as follows: N *_ .21(Xi - X) (yl ' Y) 1: r = S /S S = _ _ XY X y { I 1 (x, - x)’] I 1 (y, - y>21 } 1/2 i=1 l=1 where: x- = the ith observation of variable x l y, = the ith observation of variable y n = number of observations mean of variable x x I II x _J. \ 2 11 mean of variable y ‘< II II M 2 II. [‘12 H "< —lo \ 2 1| —J. H This formula is to be used when examining linear relationships between variables. It is based on the following assumptions: 1. For each value of x, the distribution of the associated y values is a normal distribution and vice versa. 147 2. The y means for each value of x fall on a straight line-- the relationship is linear. The same is true for each value of y. 3. The scatterplots possess homoscedasticity--the variance in the y values is uniform across all values of x. Conversely, the variance in x values is constant for all values of y. Regression analysis was used to determine the relationship between sets of predictor variables and outcomes. The outcome (y) measures used in this analysis were productivity of life, return to work, and participation in a rehabilitation program. The regression model used in this analysis is as follows: y = bo + blxl + bzxz + . . . + bnxn + e where: b0 is the intercept and b1 through bn are regression coefficients. This model is based on the following assumptions: 1. The y scores are normally distributed at all points along the regression line; that is, the residuals are normally distrib- uted. (There is no assumption that the independent variables are normally distributed.) 2. There is a linear relationship between the x’s and y’s at all points along the straight regression line; the residuals have a mean of zero. 3. The variance of the residuals is homogeneous at all points along the regression line. 148 4. The independent variables are fixed. The second hypothesis evaluated the effect of group membership on productivity by using the fixed-effects ANOVA procedure. This was used to evaluate continuous and categorical variables. The formula for this procedure is as follows: m= “+05" 811' where: yij = productivity score of subject i in group j, u = grand mean, aj = effect of group j (by type of compensation), and €ij = error. This procedure is based on the following assumptions: 1. The outcome observations are drawn from normal populations. 2. The observations are random samples from the population. 3. The error terms for any group are independent and normally distributed with a mean of zero and variance 02. The second assumption must be qualified iri that the subjects were not completely randomly sampled from 21 definable population. Howevery this limitation is; inherent in studies of“ chronic pain patients. The sample used in this study is typical of patients with chronic low back. pain, treating with MPC’s across the country. Keefe et al. (1981) stated that the characteristics are commonly ascribed to subjects suffering from chronic low back pain syndrome 149 in North America. It is not expected that the sample is biased in any way. The second hypothesis was further evaluated by utilizing. the hierarchical linear model (HLM). The HLM is a two-level model involving repeated measures expressed as a growth model. HLM models can represent any number of levels. This research uses HLM in considering a two-level problem. The hierarchy involves individual growth parameters for subjects (level 1) EHKi between-person varia- tion among growth parameters (level 2). 11w: general formula for the HLM results at the individual (first) level is as follows: . = n - + n ~t- + n -to2+-e - yt1 01 11 1 21 1 ti for i = 1 to 69 people, observed on t1 occasions (varies from 4 to 12 observations yti = affect of subject i at time t n01 = intercept nli = instantaneous growth rate for person i nz- = the curvature or acceleration in each person’s growth trajectory for affect Pretest-posttest designs traditionally have been used to evaluate the results of' MPC programs. Such designs have been recognized as inadequate for studying individual change (Bryk & Weisberg, l977; Kleinke & Spangler, l988; Rogosa, Brand, & Zimowski, 1982). However, the HLM has been found to be a satisfactory means 150 for measuring change and assessing multilevel effects (Bryk & Raudenbush, 1988). The HLM has developed out of three themes of methodological work (Bryk 81 Raudenbush, 1988): the mixed model ANOVA, random coefficient regression, and the statistical theory of covariance components. The two-level HLM model requires specification of two interrelated equations. This research uses :1 within-unit (person) and a between-person model. The HLM for research on individual growth provides an adequate measure of individual change. The yit is the observed status of individual i at time t. This is a function of a systematic growth trajectory (or growth curve) plus random error. It is assumed that the systematic growth over time is represented as a polynomial of degree k; in this case, k is 2. It is assumed that errors, eit’ are normally distributed with a mean of zero and constant variance (Bryk & Raudenbush, 1988). An important feature of the above equation is the assumption that the growth parameters vary across individuals (Bryk & Raudenbush, 1988). The between-person model represents this variation. For each of the three (k = 0 to 2) individual-growth parameters, we let “ ki = bko + bklxkli + bkzxk2i + - . . + bijkji t ”ki where iji = jth measured characteristic of the individual, either background (e.g., gender or marital status) or group membership (type of compensation) bkj = the effect of xj on the kth growth parameter 151 ”ki = the random effects (errors or residuals) with full covariance matrix, T, dimensioned (3 x 3), for i = I to 69, j = I to 22, and k = O to 2. Thus we have T 01 = boo + b01X011 + b02x02i + - ° - + bOpXOpi + “01’ W 11 = blo + bllxlli + b12X121 + . . . + blpxlpi + uli, and I 21 = b20 + b21x21i + b22x22i + - . - + prXZpi + "21° This model is used in research on change to perform a number of functions. This model describes the structure oi: the mean growth trajectory. It is used to estimate the extent and character of individual variation around mean growth. "Hue model allows us to assess the reliability of measures of both status and change. It also allows estimation of the correlation between a subject’s entry status and rate of growth, and examines correlates of status and change. It fUrther allows assessment of the adequacy of between- subject models by estimating the reduction in unexplained parameter variance. The model also improves estimates for each individual’s growth trajectory and prediction about individual future growth (Bryk & Raudenbush, 1988). The important component for this part of the statistical analysis is that the growth parameters differ over time between subjects. One growth curve is not assumed to describe the growth of every individual. The second-level model then tries to explain differences in the curves. 152 The third hypothesis was analyzed by use of the chi-square test of association. This is based on a 2 x 3 contingency table in the case examined here. The formula is as follows: 2 3 A 2 X2 = 2 2 n c (Prc ' N r.) r=1 c=1 ° 3 r. Each cell’s contribution is the square of the discrepancy between the observed and expected proportions (Prc -;; r.)2° Prc is the proportion of the observations in each column that falls into each row, divided by the expected proportion (€ r.) and multiplied by the number of observations in the column n.C for that cell. The sum of this result for all cells is then the chi-square statistic for the contingency table (Glass & Hopkins, 1984, p. 287). This is used to test hypotheses about the association between the two variables, success and group membership. Observational Research This study was designed to avoid methodological problems that often characterize observational research. In an effort to overcome limitations of collecting observational data, a reliable and valid self-report instrument, the MP0, was used. Thus, objective data were obtained through self-report. These data related to the complex set of behaviors referred to as chronic low back pain. A problem that exists in observational research is the presence of the observer. This problem was minimized as a result of 153 impartial collection of data through telephone calls and the mailing of questionnaires. The difficulty associated with the time factor was overcome by using existing data on the subjects from MFB records. This provided data collected by independent observers not previously associated with the research. Descriptive variables were used in this research. They have an advantage over inferential and evaluative variables because they require little inference on the part of the researcher (Borg & Gall, 1971) because inferential variables require inference from a sample to a population, which might or might not be justified. Recording observations was straightforward. The observational variables identified to be used in the study were listed on the data-collection form. This was easy to use because the categories were well defined and did not require a high degree of inference. The issue of training observers was avoided in this study as a result of the use of the MP0 and MFB’s intake and follow-up questionnaire. The MPQ was scored in the traditional manner by individuals with a number of years of experience with the procedure. MFB’s questionnaire provided straightforward descriptive data. The MPQ is viewed as an unobtrusive measure. This effectively minimized observer bias. There was minimal exposure for the researcher to have an influence on the subjects. This effectively eliminated the subjects being influenced by the researcher’s purpose. Observer bias was not involved because the descriptive variables did not call for conclusions or inferences. 154 Rating errors were avoided through the use of the standardized measure, the MPQ. The error of central tendency was avoided through the use of variables measured on a dichotomous scale. The halo effect also was avoided because personal contact with subjects in the study was avoided. It is believed that contamination was avoided in this study as a result of the data-collection procedures and the clear definitions of success stated before collection of data began. Summary The methodology of the study was explained in this chapter. The research questions and hypotheses were restated, and the sample and study site were described. Sample-selection, data-collection, and data-analysis procedures were discussed. The instruments used in the study' were described, with emphasis on the development, reliability, and validity of the MPQ. The results of the data analyses conducted for this study are presented in Chapter IV. CHAPTER IV RESULTS Introduction An analysis of the data is presented in this chapter. The statistical analyses. were calculated at time Michigan State University Computer Center. The Statistical Package for the Social Sciences (SPSS-X) computer program was used to analyze the data. The HLM package (Bryk, Raudenbush, Congdon, & Seltzer, 1986) was used for the hierarchical analysis. Demographic Characteristics of the Sample The demographic results of the study are presented in this section. The results are summarized in Tables 4.1 and 4.2. A total of 75 subjects (33 men and 42 women) participated in this study. Forty-three subjects (57.3%) were receiving worker’s compensation when they participated in MFB’s program. Eleven subjects (14.7%) were receiving auto no-fault benefits at the time of their participation in MFB’s program. The other 21 subjects (28%) were not receiving any financial assistance as a result of their physical impairment. The subjects ranged in age from 17 years to 78 years at the time of their involvement in MFB’s program. The average age of the participants was 39.7 years. 155 156 Table 4.1.-—Distribution of dichotomous variables used in analysis. Variable Yes (%) No (%)‘ Worker’s compensation 57 43 Automobile no-fault 15 85 High school graduate 77 33 Married 69 31 Spouse employed 28 4O Retired or employed 25 75 Litigation 23 76 Rehabilitation involvement 43 49 Table 4.2.——Distribution of continuous variables used in analysis. Variable Minimum Maximum Average Age 17 years 78 years 39.7 years Education 6 years 18 years 12.2 years Length of pain complaints 6 months 20 years 36.5 months No. of days hospitalized 0 days 42 days 2.5 days Surgeries for pain 0 surg. 9 surg. 1.3 surgeries Number of employers l emp. 6 emp. 2.3 employers Seventeen of the subjects (22.9%) had not completed high school, whereas 58 of the subjects (77.1%) were high school graduates. The average level of education of the subjects was 12.3 years. Fifty-two subjects (69%) were married; 23 (31%) were not. The spouses of 21 subjects (28%) were not employed. Thirty of the spouses (40%) were employed at the time of participation in the study. Spouse-employment data were missing for 24 subjects (32%). 157 A distinction was not made between not employed and seeking work and chronically unemployed. ‘ Employment status was unrelated to gender, as indicated by :1 correlation coefficient of .04. Men and women tended ix) be unemployed in roughly equal proportions. Nineteen of the subjects (25%) were employed or retired at the time of participation in the program. Fifty-six subjects (75%) were not employed. The average number of employers for subjects before entering the program at MFB was 2.3. Fifty-six subjects (80%) had three or fewer' employers before» entering the progranr Fourteen subjects (20%) had between four and six employers. This is a lower number of employers than one would expect to see; the hospital records only the number of major employers during adult life. Also, the number of employers is a factor in evaluating candidates for participation in the program, so applicants with many short-term employers may be screened out (Kremer, 1991). Thirty—seven subjects (49%) had one source of income at the beginning of their participation in MFB’s program. Thirty-three subjects (44%) had two sources of income in the household, and five subjects (7%) had three sources of income. The average length of chronic pain was 36.5 months. The range of chronicity was from 6 months to 20 years. The average number of days hospitalized for pain in the past year was 2.5 days. However, 45 subjects had not been hospitalized for pain. Eleven subjects had missing data in this area. Those hospitalized ix": pain spent an average of 8.5 days in the hospital. 158 The average number of surgeries for pain relief was 1.3. Thirty-one subjects (41%) did not have any surgery. Nineteen subjects (25%) had one surgery. Fifteen subjects (20%) received two surgeries for pain relief. Five subjects (7%) had three surgeries, and five subjects (7%) had four or more. Fifty-seven subjects (76%) did not have litigation pending when they began their participation in MFB’s program. Seventeen subjects (23%) were involved in some form of litigation. One subject did not report this information. McGill Pain Questionnaire Results The McGill Pain Questionnaire results are summarized in Table 4.3. A Present Pain Intensity (PPI) of 72.8 (on a scale of O to 100) indicates a moderately high level of overall pain intensity. A sensory score of 51.3 indicates that they have selected a combination of words that reflect a moderate level of discomfort. This level of sensory discomfort seems proportionately lower than would be expected, given the average PPI of 72.8. This suggests that they may be experiencing some symptom magnification. An affective score of 25.1 is a mild elevation of affective distress. The evaluative score of 75.8 is moderately high. The number of words chosen (NWC) of 18 is a moderate number. The results on success of the rehabilitation process were computed in simple percentages. Success was defined as being employed or participating in a rehabilitation program at the time of the last follow—up. The unsuccessful group numbered 29 (38.7%). 159 The successful group included 38 (50%) individuals who returned to work and 8 (10.7%) individuals who were participating in rehabilita- tion, a total of 46 (61.4%). Table 4.3.-—McGill Pain Questionnaire results. Variable N Mean SE Min. Max. PPI 75 72.8 2.00 25 100 NWC 75 18.2 1.29 3 65 Sensory 75 51.3 2.36 7 100 Evaluative 75 75.8 3.40 10 100 Affective 75 25.1 2.65 O 86 Correlation Data The correlation coefficients are presented in Table 4.4. There were several significant correlations; however, the only large significant correlation was that between productivity and the dichotomous measure of outcome (r = -.93). The PLQ was reviewed in Chapter III. It should be noted that high productivity scores, or levels 9 through 12, were the least productive, levels 5 through 8 were moderately productive, and levels 1 through 4 were the most productive. The first hypothesis, "Chronic pain patients’ score on the Affective dimension of the MP0 at intake will correlate positively with successfu1 outcome following 'treatment in 2u1 MPC," was not substantiated by the data. The correlation coefficient was .01 for outcome and affect. The correlation coefficient for affect and productivity was .14. 160 oo._ poocc< «F. oo.— >u_>_uo330ca Po. mo.- oo.— aeoouno ow. op. ~—.- oo.— o>_um3.m>m mm. «o. _o.- 0.. oo._ scomcam vo- -. c—.- on. oo- oo.. U32 mo. —P.- no. o—. «F. «P. oo.— _Qa 3o.- P... so. mN. -. mp. no. oo._ measo.n=w wo.- co. no.- -.- mo.- oo.- hp.- mF.- oo.— >comcam mo. ~F. c—.- 9.. ea. mc. no. mo.- Fm. oo.— .~__mu_amo= po.- m~.- op. «c. c~.- p—.- «F.- mo. mm. mo.- oo.. >u_o_c0czu 09. mp. ¢~.- up. m—. pp. om. No. N.. «o. op. oo.F «Eco:— po.. oo.- so. no.- oo- mo. co. 0.. «Q. «9.- o~.- «o. oo.F .nmnom mo.- ow. up.- "N. op. 39.. mm. mo. mo.- QN. PF.- QF. 00.. oo.— co_umm_u_4 «F. mp. mo.- op. NP. ON. o~. op. ~o.- me. QF.- Fo.- Nm. ~—. oo.. .QEo aco_.u up. m0. m6.- mm. mo.- «F. -.- u—.- mo. ON. —~.- mm. «o.- w_. mo. oo.— .QEo «macaw mp.- «9.. up. "9.- NF.- v—.- n—. No. N..- no. a..- o~.- FF. no. NF. om.- oo.F .mu_cmz om.- q~.- sp- uo. 09.- op.- F~.- cc. No. no. oo.- op. mp.- mo.- om.- ~N. mo.- oo.P co_umo:nw mp.- q~. o~.- No.- o¢.- ~o.- 09.. m—.- pm. o—.. «M. m_. m—.- op.- -.- oq.- -.- No. 00., om< op. no.- «a.- «o.- oo.- we. «0.- mo.- .0. m—.- no. mo.- mm. cm.- «0. m~.- 59.- cc.- no. oo.F concou u. .d no :4 c. N” .d :4 “S H” «o I. M. a: «J c. ”w as w. n. 1: J n A a M d w n O U- U a L. L d e o- 6 9 I: O 3. P. U 3 .I. d J S J O U. 1. L. 0 J n 8 U a D. 0 L S L 5 d O O 9 L. a n L. O 0- 3 n O n 0 O 3 L. U w G- 6 U S 1. 9 9 3 0 w 9 J A J 3. L. 9 L. D. 1. 8 Q .4 J 3 a 1. K a K e O L 1+ L L. L. I.- J L I... I... I: 9 a 0 A A S L. 1 1 o w w u L. 9 Z rA P. U .0 d and. D. «1.- L L ,A 14 L. o o L. 0 K rA o u a a U D- D- S .muco_o_**006 co_quoccou--.c.¢ o.nm_ 161 Regression Analysis In an attempt to determine whether any combinations of variables could be used to best predict successful outcome, a regression analysis was used to evaluate the various combinations of variables noted in Table 4.5. The variables that were most significant were used in the analysis. This outcome was best predicted by chronicity, education, marital status, and age. Table 4.5 shows the regression analysis. 'The results are presented in Table 4.6. Table 4.5.—-Regression analysis for outcome. Sum of Mean dfi Squares Square F 9 Regression 4 2.257 .564 Residual 66 14.221 .215 2°62 '042 Total TO Table 4.6.-—Variables in the equation for outcome. Variable b SE(b) p Chronicity .003 .001 .095 Education .052 .028 .069 Marital .207 .122 .096 Age -.012 .006 .058 (Constant) .099 .462 .832 Note. Alpha was set at .10 for individual coefficients (b). 162 Although the regression test was significant, the overall percentage of variance explained was very low (32 = .14). This regression analysis demonstrated that the younger an individual was, the more likely he or she was to succeed. The longer the individual had suffered from the condition, the more likely he (n' she was to succeed. The higher the individual’s level of education was, the greater potential that person had for success. If an individual was married, his or her chance for success was greater. The nonsignificant variables were gender, spouse employed, client. employed, litigation, rehabilitation, income, hospitaliza- tion, surgery, employers, PPI, NWC, sensory, evaluative, and affect scores on the MPQ. Table 4.7 shows the regression analysis for productivity. The best model explained only 29% of variability in productivity. The significant predictors (alpha == .10) were evaluative, chronicity, client employment, marital status, and sensory measure. Table 4.7.--Regression analysis for productivity. Sum of Mean df Squares Square 5 9 Regression 5 231.935 46.387 Residual 51 576.906 11.311 4'10 '0003 Total 56 Another regression analysis was conducted to determine the best predictors for successful outcome, as measured by productivity. 163 Table 4.8 shows the coefficients and standard errors for the above- named variables. It shouhd be noted that the signs of the coefficients ought to be opposite those for outcome because low productivity scores are "good." Table 4.8.-—Variables in the equation for productivity. Variable b SE(b) p Evaluative .038 .016 .021 Chronicity - .038 .014 .008 Client employed 2.857 1.079 .011 Marital -2.635 1.048 .015 Sensory - .044 .026 .097 (Constant) 3.156 3.008 .299 Note. Alpha was set at .10 for individual coefficients (b). Past research has suggested that several of these predictors would be significant, such as client employment (Carlsson, 1984; Dworkin et al., 1985; Maruta et al., 1979) and marital status (Maruta et al., 1979). Gender (Maruta et al., 1979) and age (Arnoff & Evans, 1982) were not significant as predictors, as previous studies implied. The reason these predictors were not significant may be explained, in part, because the other studies did not use evaluative and sensory as predictors. The inclusion of additional variables will affect the results of a study. The nonsignificant variables in this analysis were gender, age, education, spouse employment, litigation, rehabilitation, income, hospitalizations, surgery, employers, PPI, NWC, and affect. Litiga— tion and compensation historically have been viewed as negative 164 characteristics, as pointed out by the following researchers: Block et al. (1980), Dworkin et al. (1985, 1986), Guck et al. (1986), Kleinke and Spangler (1988), and Melzack et a1. (1985). Previous studies also supported the concept that pain duration or chronicity was a predictor of limited success (Dworkin, 1986; Keefe et al., 1981; Maruta et al., 1979). High evaluative scores on the MPQ were significantly related to high productivity scores, with other variables held constant. High productivity means people are not successful in being active participants in life activities. Individuals with high evaluative scores may perceive themselves as significantly disabled by pain and not able to perceive themselves achieving significant change. High chronicity was related to a low productivity score, which means the longer someone suffered from chronic pain, the more likely he or she was to recover. Those individuals suffering from high chronicity may have been more motivated to practice the modalities presented to them. If subjects were unemployed at intake, this related to a high (poor) productivity score. Perhaps clients who were not employed might not have wanted to increase their levels of productivity or to be more active; they might have adjusted to the lifestyle of the "disabled." Married clients also seemed to get back into life more quickly. Those individuals demonstrating high sensory scores at intake appeared to increase the quality of life, as demonstrated on the Productivity of Life questionnaire. 165 Missing Data The regression analysis was hindered by missing data on several of the ‘variables. The Productivity (Hi Life questionnaire information was missing on 15 subjects. 0n spouse employment, 24 subjects provided no information. Eleven subjects provided no information on hospitalizations. Six subjects were missing information on rehabilitation. Several other variables were missing information regarding one or two subjects. An examination of subjects who were missing data seemed to indicate the data were not missing at random. For example, those missing information on the productivity variable were, on average, 1.5 years younger, had fewer hospitalizations (2.1 vs. 2.7), and were less likely to have succeeded in the program (47% vs. 65%). Therefore, the regression analySis excluded cases and variables that may have altered the results. Affect by Time The second hypothesis, "Subjects with high scores on the Affective dimension of the MPQ at intake will have a nonlinear pattern of Affective dimension scores over the course of treatment, upon completion of the program, and at follow up," was evaluated by analysis of variance. Scores were combined across. weeks within six time periods. Thus, there were six combined scores called yti’ for t = O to 6 for the ith person. The periods were: 0 for intake, 1 through 3 weeks, 4 through 10 weeks, 11 through 26 weeks, 27 through 52 weeks, 53 166 through 73 weeks, and 79 or more weeks. The highest number of weeks recorded was 188. Table 4.9 provides the means across subjects for each time period, and the means are plotted in Figure 4.1. Figure 4.2 shows mean affect for the successful and unsuccessful groups. These means, especially those for week 26 and thereafter, were based on very low numbers of subjects, and differences were not significant. Table 4.9.—-Mean scores on the Affective dimension of the MP0 for each time period. Week Midpoint Week Index Affect Mean ("ti) Actual Weeks (t) E 0 O O 75 25.19 2 1-3 1 187 18.85 7 4-10 2 236 14.87 18 11-26 3 72 13.64 40 27-52 4 34 15.94 65 53-78 5 38 14.58 122 79+ 6 42 16.60 The week 0 score was the intake score obtained on the Affective dimension of the MP0. The affect mean score of 25.19 is a mild elevation of affective distress. Weeks 1 through 3 were the begin- ning stages of treatment for most individuals. At 18.85, the mean level of Affect was reduced quickly from that of intake. Treatment was completed for most subjects between weeks 4 and TO. Continued reduction in the level of affect occurred during this time period. The measure designated weeks 11 through 26 was taken immediately following completion of the MPC. The level of affect continued to 30 25 20 15 r+catp -h-+1I> 10 3O 25 20 15 e-c-nro-h-h> 10 167 O 1 2 3 4 5 6 Weeks Index Figure 4.1: Weeks by average affect (all subjects). Failure Success 0 1 2 3 4 5 6 Weeks Index Figure 4.2: Average affect by success/failure. decrease. However, when the subjects’ levels of affect were measured between 27 and 52 weeks, an increase of 2.31% was noted. This was followed in weeks 53 through 78 by a decrease of 2.37%. 168 This level was still higher than the period measured at weeks 11 through 26. Finally, in week 79 and thereafter, affect scores increased to an average level of 16.60, £1 higher level of affect than seen since subjects were in treatment. The means depicted in Table 4.9 were examined by use of repeated-measures ANOVA. The results are shown in Table 4.10. There was a significant quadratic effect and a significant cubic effect. The means showed a downward trend, but affect seemed to sink and then rise slowly toward the end of the time measured. Table 4.10.--Analysis of variance: affect over time. Source df SS MS E p Between weeks 6 7804.89 1300.81 3.20 .004 Linear 1 490.54 490.54 1.21 .270 Quadratic 1 1867.27 1867.27 4.59 .030 Cubic 1 2315.24 2315.24 5.69 .020 Within weeks 677 275295.95 406.64 Total 353 283100.84 A problem with this analysis resulted from the recoding of weeks into unequal intervals. The test for trends was not as precise as it could have been if time had been coded into equal intervals. However, the data did not lend themselves to even groupings because measures were obtained with less frequency as time passed. Another related problem is that data did not exist for all subjects in all weeks. The number of people in any week decreased in later weeks. 169 The Hierarchical Linear Model The hierarchical linear model (HLM) analysis was used to determine which variables best predicted individual patterns of affect over time. The formula for the HLM vfithin-unit model for affect level is as follows: yti = TT01' + “1i (”ti ' L) + Tr21 ("ti ' L12 + eti t goes from O to t~, where t1 is the number of scores avail- able for person 1 (in weeks 1 through 20) for i = 1 to 69 people L = O, 4, 8, and 12 weeks yti = the tth affect score for person i «01 = intercept “1i = instantaneous growth rate for person i at time L 021 = the curvature or acceleration in each person’s growth trajectory of affect ”ti = number of weeks past intake, when affect score yti is obtained eti = error term One of the strengths of the HLM is that it allows for growth modeling. The individual growth model for the affect level of subject i on occasion t is represented in the above equation. Each of the growth parameters in the above equation has meaning. The intercept 1101- represents the status of person i at time L. The linear' component, 1111, is the instantaneous growth rate (linear 170 slope) for person i at time L. '021 represents the curvature or acceleration iri each person’s growth trajectory' of affect. Acceleration is a characteristic of the entire trajectory. The status and instantaneous rate parameters depend (n1 the particular choice of the location parameter L. An example would be: If L is set at 8 weeks, then '00; and '011 represent the status and instantaneous rate for person i at that particular time point. The scores across all time points were modeled in fOur ways. Each of the four ways of' modeling involved anchoring the time variable, i.e. (weeks - L), at a different point, represented by L. The form of the time values (weeks - L) will differ when weeks are anchored at different points, which can allow one to see different aspects of how score patterns change over time. The linear and quadratic terms for L = 0, (weeks) and (weeks)2 are highly correlated. By anchoring weeks at L and creating two predictors (weeks - L) and (weeks - L)2 at different values of L, the dependence between the two predictors may be reduced. These analyses reveal the differences in slopes and in intercept at all four time points. These time points represent L. The first time point was 0 weeks, which was the intake reading of the MPQ. Subjects were evaluated at that time point as potential candidates for the MPC. The second time point was at 4 weeks. This time point was approximately 1 month after the subjects were in treatment. It was thought this time point represented a minimal amount of time in treatment, after which the effect of treatment on affect could be 171 measured. The 8-week time point is toward the end of treatment in the MPC. This time point measures affect at the close of treatment and represents immediate feedback on the level of affect. Twelve weeks is the final time point measured because it is approximately 1 month after treatment has ended for most of the subjects. An important feature of the above equation is the assumption that the growth parameters vary across individuals. The between- patient model is formulated to represent this variation. The between-patient model is as follows: Wkl bk 0 + bklxkl + bkzxkz + . . . bk7xk7 + Ukl where: x1 = the subject’s litigation status x2 = the subject’s length of chronicity x3 = the subject’s number of surgeries x4 = the subject’s number of words chosen on the MP0 at intake x5 = the subject’s level on the evaluative scale of the MPQ at intake x6 = the subject’s level of affect on the MPQ at intake x7 = the subject’s membership in the auto no—fault group bk = the effect of xp on the kth growth parameter P uki = the random effects with full covariance matrix, T, dimen- sioned (K + I x 1) 172 The uki are random effects that represent the deviations in the growth parameters for subject i from the respective means. It is assumed that the effects are normally distributed with mean 0 and variance-covariance, T. The growth parameters are viewed as random among individuals, but the sources of variation are unknown. The 3 statistics provide statistical evidence that each of the parameters in the mean growth trajectory is different from zero. The total parameter variance for each growth parameter is estimated by the diagonal elements of T. The HLM provides a large-sample chi-square test of homogeneity, that is, of Ho: kk = 0. HLM Analyses This analysis was performed with the original coding of weeks. The weeks equaled the number of weeks following intake. However, values greater than 20 weeks were dropped because the scores were so few and widespread. Another reason for dropping the later scores was to reduce the possibility that influences outside of treatment were altering the results. HLM analyses were conducted at four time points—~intake, 4 weeks, 8 weeks, and 12 weeks. That is, L was allowed to equal 0, 4, 8, and 12 weeks, respectively, in the four analyses. The average intercept and slope for the within—subjects models are presented in Table 4.11. The intercepts are plotted in Figure 4.3. 173 Table 4.Tl.--Average slope coefficients across individual growth models. Intake 4 Weeks 8 Weeks 12 Weeks Intercept 23.77* 16.58* 12.45* 11.38* Linear effect -2.18* -1.42* -O.65* 0.00** Quadratic effect 0.10* 0.10* 0.10* 0.10* *p < .001. **p = .59 (ns). 25 " 23.77 20 - 165m A f 15 - f 12°45 11.38 e 10 ~ c t 5 » O O 4 8 12 Weeks After Intake Figure 4.3: Average affect intercepts. The intercept represents the average affect score at each time point. It rapidly decreased from intake to 4 weeks and slowly decreased thereafter. This indicates that the subjects were obtaining information and training that was beneficial to them. The 174 subjects’ level of affect seemed to be moderated to some degree by their participation in the MPC. The average slope for the linear effects was negative at the first three time points, indicating a decrease in affect. This decrease was strongest at time 0. After 12 weeks, this average slope was positive, and insignificant. This indicates that the slope for a linear trend in affect was flat, and not significantly different from zero. The average quadratic term was positive, significant, and constant across all time points. The indicates an overall curvature upward in the affect patterns. While the tendency for affective scores to increase is small, it is significant. Perhaps subjects in the MPC initially experience reduced affect because they are getting relief from pain-management techniques. At the same time, they also are required to face new issues that have been avoided to date, and this may explain the later increase in affect. The variability in both the intercepts and linear slopes were significant at all four time points. This indicates that the intercept and linear slope differed significantly between subjects. The HLM allows the modeling of these differences. The variability in the quadratic term was not significant at any time point, and therefore that variance was fixed at zero (and unmodeled) for the remaining analyses. 175 Overview of results. At each time point-—intake, 4, 8, and 12 weeks--a11 22 predictors of the intercept and slope were examined. Affect at intake was included because it would influence prediction at intake. But, more important, it is involved in predicting affect at the various time points. To keep the same variables constant across all four time points, affect at intake (labeled "base affect") was included at this time point. For the intake analysis, the variables litigation, chronicity, surgeries, number (H: words chosen, evaluative score, affect at intake, and membership in auto no-fault group were significant predictors, at the CH) level, of either the intercept, the slope, or both. Insignificant variables were gender, education, age, marital status, membership in worker’s compensation group, spouse employed, client employed, rehabilitation, hospitalizations, sensory, PPI, productivity, outcome, and employers. In the analyses for 4, 8, and 12 weeks, the variables found to be significant. were litigation, chronicity, surgery, evaluative, base affect, and membership in the auto no-fault group. Number of words chosen was no longer significant. The four analyses appear in Tables 4.12 through 4.19. Table 4.12.~-Statistical 176 table for HLM coefficients at intake (L = 0). Q _E(b) t-Statistic p-Value Intercept INTERCEPT 16.11 7.98 2.02 .05 LITIGATION -5.04 2.89 -l.75 .09 CHRONICITY 0.00 0.03 0.00 .39* SURGERY -0.45 0.72 -0.63 .33* NWC 0.24 0.14 1.74 .09 EVALUATIVE 0.02 0.04 0.44 .36* BASE AFFECT 0.55 0.06 8.69 .00 GROUP 2 1.49 3.53 0.42 .36* Slope INTERCEPT -l.73 0.98 -l.76 .09 LITIGATION -O.64 0.35 -1.84 .08 CHRONICITY -0.0l 0.00 -l.82 .08 SURGERY 0.29 0.08 3.65 .00 NWC -0.02 0.02 -1.11 .21* EVALUATIVE 0.01 0.01 1.79 .08 BASE AFFECT -0.04 0.01 -4.29 .00 GROUP 2 0.92 0.49 1.84 .07 Quadratic Intercept 0.08 0.02 4.28 .00 *Not significant at p < .10. Table 4.13.--Variance components table at L = 0. Estimated Random Parameter Parameter df Chi-Square p-Value Variance Base coefficient 45.67 60 97.29 .00 Week slope 0.21 60 74.17 .10* *Not significant at p < .10. 177 Table 4.14.--Statistica1 table for HLM coefficients at L = 4 weeks. b S§(Q) t—Statistic p-Value Intercept INTERCEPT 10 53 7.98 1.34 .16* LITIGATION -7.60 2.92 -2.60 .02 CHRONICITY —0.03 0.03 -O.95 .25* SURGERY 0.71 0.74 0.97 .25* NWC 0.16 0.14 1.17 .20* EVALUATIVE 0.05 0.04 1.29 .17* BASE AFFECT 0.41 0.06 6.37 .00 GROUP 2 5.17 3.53 1.47 .14* Slope INTERCEPT -l.06 0.95 -l.12 .21* LITIGATION -0.64 0.35 -1.84 .08 CHRONICITY -0.01 0.00 -1.82 .08 SURGERY 0.29 0.08 3.65 .00 NWC -0.02 0.02 -l.ll .21* EVALUATIVE 0.01 0.01 1.79 .08 BASE AFFECT -0.04 0.01 -4.29 .00 GROUP 2 0.92 0.49 1.84 .07 Quadratic Intercept 0.08 0.02 4.28 .00 *Not significant at p < .10. Table 4.15.--Variance components table at L = 4. Estimated Random Parameter Parameter df Chi-Square p-Value Variance Base coefficient 71.33 60 244.11 .00 Week slope 0.21 60 74.12 .10* *Not significant at p < .10. Table 4.16.--Statistical table 178 for HLM coefficients at L = 8 weeks. Q S_(Q) t—Statistic p-Value Intercept INTERCEPT 7.71 9.65 0.80 .28* LITIGATION -10.19 3.63 —2.81 .01 CHRONICITY -0.06 0.04 -l.52 .12* SURGERY 1.87 0.89 2.08 .05 NWC 0.08 0.18 0.47 .35* EVALUATIVE 0.09 0.05 1.72 .09 BASE AFFECT 0.26 0.08 3.19 .00 GROUP 2 9.04 4.62 1.95 .06 Slope SLOPE —O.39 0.96 -O.4l .36* LITIGATION -O.64 0.36 —l.79 .08 CHRONICITY -0.01 0.00 -l.79 .08 SURGERY 0.29 0.08 3.49 .00 NWC -0.02 0.02 —1.08 .22* EVALUATIVE 0.01 0.01 1.73 .09 BASE AFFECT -0.04 0.01 -4.22 .00 GROUP 2 0.95 0.51 1.85 .07 Quadratic Intercept 0.08 0.02 4 25 .00 *Not significant at p < .10. Table 4.17.--Variance components table at L = 8. Estimated Random Parameter Parameter df Chi-Square p-Value Variance Base coefficient 109.50 60 198.650 .00 Week slope 0.28 60 74.168 .10* *Not significant at p < .10. 179 Table 4.18.--Statistical table for HLM coefficients at L = 12 weeks. p S_(Q) t-Statistic p-Value Intercept INTERCEPT 7.38 11.98 0.62 .33* LITIGATION ~12.72 4.53 -2.81 .01 CHRONICITY -0.09 0.05 -l.77 .08 SURGERY 3.04 1.09 2.77 .01 NWC 0.00 0.23 0.01 .39* EVALUATIVE 0.13 0.06 1.93 .06 BASE AFFECT 0.12 0.10 1.17 .20* GROUP 2 12.53 6.02 2.08 .05 Slope INTERCEPT 0.27 0.95 0.29 .38* LITIGATION -0.64 0.35 -l.84 .07 CHRONICITY -0.01 0.00 -l.82 .07 SURGERY 0.29 0.08 3.65 .00 NWC -0.02 0.02 -l 11 .21* EVALUATIVE 0.01 0.01 1.79 .08 BASE AFFECT -0.04 0.01 -4.29 00 GROUP 2 0.92 0.49 1.84 07 Quadratic Intercept 0.08 0.02 4.28 00 *Not significant at p < .10. Table 4.19.--Variance components table at L = 12. Estimated Random Parameter Parameter df Chi-Square p-Value Variance Base coefficient 142.78 60 146.64 .00 Week slope 0.21 60 74.08 .11* *Not significant at p < .10. 180 Intake Analysis In this section, the results are described in further detail. At intake, the variables significantly predicting the intercept were litigation, number of words chosen, and base affect. The last variable makes sense because the intercept should be higher if the subjects have high affect at intake. More words chosen would also typically lead to higher affect at intake. Being involved in litigation indicated a lower affect at intake. Perhaps the individuals represented by legal counsel received more information on their rights regarding their situations. They might not have been as apprehensive about legal and financial matters as the other subjects. This may be reflected in their affect scores. The variables that predicted slope at intake were litigation, chronicity, surgeries, evaluative, base affect, emu! membership in Group 2 (the auto no-fault group). ‘Those patients involved in litigation showed steeper negative slopes, indicating a quicker drop-off in affect score. Also leading to steeper negative slope were chronicity and base affect. Those patients with more surgeries, higher evaluative scores, and those receiving no-fault benefits saw the overall negative slope reduced (less steep) by these factors. That is, their affect did not drop off as fast. The reason these factors were significant may relate to reasons already stated for litigation, but those individuals with long-term chronicity and surgeries for pain relief may find ways of managing their pain. Those individuals with chronicity may be more ready to 181 use the principles taught at the MPC. It is important to note that this' may be the first time that a group of health professionals recognize and accept their pain problem as real. The acceptance of their pain by health professionals may help to explain the reduction in affect. An explanation for the significance of Group 2 membership (the auto no-fault group) is that these individuals typically have limited wage support in Michigan, 3 years by statute. These individuals may have support from their employer for return to work, or they may have a job to return to once medical rehabilitation is complete. The individuals receiving auto no—fault benefits are aware of the 3-year limitation of wage support. There is a clearly defined limit to the resources from which no-fault recipients can draw. Individuals receiving worker’s compensation have vague references as to the length of their disability and the financial support they may receive. Week 4 Analysis The analysis for the 4-week, mid-treatment interval demonstrated that litigation and base affect were significant. NWC was no longer significant at this time point. Again, the same reasons would apply for litigation and base affect being significant. The same variables were significant for predicting the slope at this time point. It appears that these factors were remaining constant during treatment. 182 Week 8 Analysis Examining the 8-week interval, which occurred at the end of treatment for most subjects, litigation and base affect were sig- nificant for predicting the intercept. But added were surgeries, evaluative, and membership in auto no-fault. Based on the lower level of affect, it would appear that emotional issues were being resolved or dealt with in a positive or constructive manner. Individuals who had a history of surgery showed a significantly lower level of affect, indicating that they were making use of the modalities being presented in the MPC. Individuals who previously received surgery for relief of low back pain are individuals who previously relied on an external variable for the relief of their pain. Since surgery did not relieve them of their pain complaints, the MPC is a viable alternative. These individuals are perceived as having real pain difficulties. The health professionals confirm that they are not the only ones to suffer from these types of prolonged complaints. The techniques provided in an MPC require that the individuals assume responsibility for active participation in the program. They are no longer passive subjects awaiting the surgeon’s scalpel. For pain relief in) be effective, these individuals must assume more responsibility for their behavior than they did in the past. The people who previously had surgery are now more significantly motivated toward an internal locus of control, as opposed to the external locus of control. 183 Week 12 Analysis The lZ-week analysis had litigation zuui base affect as significant variables for the intercept. Chronicity and surgery were additional predictors at this time point, as was evaluative. It seems likely that the combination of surgeries and chronicity were again significant because the individual subjects recognized that they had control over the physical complaints and behaviors they exhibited. Evaluative was having an influence at this time point because of the need for the subjects to make difficult choices at this stage of their treatment. Again, the same predictors were significant for the slope at this time point. ngmary of the Four Time- Point Analyses The most significant predictor for the intercept at intake was base affect. At 4 and 8 weeks, the most significant predictors were litigation and base affect. At 12 weeks, the most significant predictors were litigation and surgery. Base affect was a significant predictor in the weeks of treatment, but not in the weeks immediately following treatment. Litigation appeared in) be important not so much at intake, but when treatment was provided. Following treatment, litigation was a significant variable in the prediction equation. Apprehension over the outcome of treatment may account for this finding. The most significant predictors for linear slope an; all time points were surgery and base affect. This demonstrates that affect at intake was a significant component for predicting decrease in F;— 184 affect through the course of treatment and at follow-up. The surgery predictor had a positive slope for the linear trend at all time points. This suggests that higher slopes (or a tendency for their affect scores not to decrease) were found for people who had surgery. The chi-square homogeneity—of-variance tests indicated that the differences in the intercept were not fully explained in any of the four tables. The differences in the week slope were explained by the variables. That is, it cannot be fully explained why someone was higher or lower on affect, but one can explain the linear pattern for everyone. Success-by-Compensation Analysis Hypothesis 3, "Success rates will not differ among the worker’s compensation group, the auto no-fault group, and the other group," was supported by chi-square analysis of the contingency-table data. The results of this procedure are presented in Table 4.20. Table 4.20.--0utcome for patients in three compensation groups. Worker’s Auto No- Row Outcome Comp. (%) Fault (%) Other (%) Total (%) Failure 17 ( 39.5) 3 ( 27.3) 9 ( 42.8) 29 ( 38.7) Success 26 ( 60.5) 8 ( 72.7) 12 ( 57.1) 46 ( 61.3) Column total 43 (100.0) 11 (100.0) 21 (100.0) 75 (100.0) 185 The chi—square test of association (chi-square = 0.77, df = 2, p = .68) was not significant, indicating no relationship between compensation group and outcome. An ANOVA was conducted to determine whether group differences existed on the productivity variable for the three groups. The results of that analysis are summarized in Table 4.21. Table 4.21.—-Productivity by worker’s compensation. Sum of Mean Sig. of gf Squares Square F E Main EffECtS 2 10.88 5.44 0 35 7" Residual 57 899.85 15.79 ° ° Total 33 910.73 15.44 No significant differences were found. The three groups did not differ on either outcome variable. These results are similar to those of Mendelson (1982) and Pelz and Mersky (1982), demonstrating that patients with low back pain do not fall neatly into worker’s compensation groups and non-worker’s compensation groups. The results in this study differed from those of Dworkin (1985, 1986), who found that compensation cases predicted poor short-term outcome. CHAPTER V SUMMARY, DISCUSSION, AND IMPLICATIONS Summary Pain is a complex problem as opposed to a simple sensation. Scientific measurement procedures have focused (”1 pain as an individual sensory quality that varies only in intensity. Traditional pain theories, such as the specificity and pattern theories, focused on the transmission of pain signals but failed to describe the role of psychological factors in pain perception. Evidence has supported the concept that pain is influenced by activities of the "higher central nervous system." Historically, descriptions of pain mechanisms have failed to account for the psychophysiological process. This is the cognitive- affective component that motivates the organism to avoid, minimize, or stop the pain as quickly as possible. Melzack and Wall (1965) proposed the gate-control theory of pain, which recognizes the obvious sensory dimension of pain as described tn! the specificity and pattern theories. The gate-control theory also provides the basis for considering the motivational-affective dimension of pain. Failure to consider this dimension of pain has seriously limited the total picture of the pain experience. The motivational dimension is crucial to the concept of pain as a perception, comprising past experience, attention, and sensory quality and intensity. 186 187 Review of Literature The literature reviewed in this research focused on two areas in which the motivational-affective concept is used as a foundation. The first area was the pioneering work Fordyce did in investigating the psychology of chronic pain. The second area was that of pain measurement, specifically that of the McGill Pain Questionnaire (MPQ). Fordyce (1968) described the use of behavioral-management techniques for problems associated with treating chronic pain. His publications (Fordyce et al., 1968a, 1968b, 1978) caused a rapid increase in the use of these techniques for treating chronic pain. A discipline known as behavioral medicine has developed over the past 15 to 20 years. The goal of 'traditional medical treatment is to remove or relieve the pathogenic process. Since this is often impossible with chronic low back pain, the condition presents a challenge to traditional medical scientists. To meet the demand for relief from chronic pain, multidisciplinary pain clinics (MPCs) have been developed. 'The various treatment approaches used iri these clinics were reviewed in Chapter II. The operant approach views pain as a set of oyert responses, such as medication taking, limping, and reporting pain. It recognizes that overt responses are controlled by reinforcers, such as attention or medications, if the reinforcers are given contingent on the pain behaviors. Treatment with this approach concentrates on eliminating the reinforcement of pain behaviors and increasing 188 healthy behaviors. The operant programs have resulted in clinically significant increases in activity levels and reduction in analgesic intake. The relaxation approach assumes that organic processes are relevant and are influenced by learning. The idea for treatment of chronic pain is to break the vicious pain cycle that exists. This is accomplished by reducing muscle tension and psychological stress to control the pain. The data from studies in this group indicated that many patients may benefit from relaxation treatment. The cognitive-behavioral approach sees pain as an experience that is mediated by cognitions. Because pain is a subjective experience, treatment at the cognitive level is appropriate. If this approach successfully mediates pain, there might be a concurrent change in other pain-related behaviors, such as activity and medication intake. This approach holds that, regardless of the origin of pain, modifying cognitions may reduce pain levels and provide the patient with a better method of dealing with pain. There is little evidence that cognitive strategies are effective in treating chronic pain. The multimodal approach is an attempt to improve the treatment of pain by using several techniques to control as many pain variables as possible. The multimodal approach includes operant, relaxation, and cognitive strategies, in addition to a wide range of other techniques. The clinical significance of the improvements noted as a result of this treatment approach is difficult to judge 189 because of weak designs and because several techniques are employed. The promise of multimodal treatment lies in combining the ‘most effective aspects of various approaches. Over the past 20 years, considerable support has developed for the effectiveness. of' behavioral pain-management techniques. Researchers have demonstrated exclusively positive results. However, it is difficult to rehabilitate individuals to their pre- pain level of functioning. Treatment should be oriented to helping the patient live as normally and productively as possible under the circumstances. The management of chronic pain allows individuals to learn to lead useful and satisfying lives despite their pain. In an effort to determine which patients are most likely to benefit from treatment, a number of researchers have undertaken prediction studies. The initial work in this field was done by Maruta et al. (1979). The results of that study demonstrated that the likelihood of success in pain management decreased for individuals with an increase in prior duration of pain, work time lost, number of previous operations, and level of' pain at 'the beginning of ‘the program. Over the years, various researchers have found that subjects with healthy life styles, indices of depression, shorter duration of pain, and gainful employment. were favorable candidates for participation in pain clinics. Individuals with litigation pending, receiving compensation, and elevations on the Hypochondriasis scale of the MMPI were judged less likely to succeed. 190 Researchers investigating the Affective component of the MP0 have found that high scores on the MMPI Hypochondriasis scale related to high scores on the Affective and Evaluative dimensions of the MPQ. Elevated scores on the MMPI Depression and Hysteria scales were associated with high scores on the Affective dimension of the MPQ. A number of investigators have evaluated the combination of affective distress and worker’s compensation. 'Hwe results of the various studies led the researchers to conclude that subjects with litigation pending or receiving worker’s compensation needed to be assessed individually and very carefully for motivation. The motivational affective component of the perception of pain was formalized in Melzack and Wall’s (1965) gate-control theory of pain. Melzack and Torgerson (1971) developed a new approach to describe and measure pain. Melzack (1975) formalized and refined the work he and Torgerson initiated. The MP0 is a verbal—adjective questionnaire that is used to quantify subjects’ perceptions of pain on three dimensions: sensory, affective, and evaluative. The MPQ has been found to have acceptable reliability and face, construct, concurrent, and discriminant validity. Since its introduction in 1975, the MP0 has been recognized as a useful instrument with which to obtain data on both qualitative and quantitative aspects of pain. Purpose The researcher’s purpose in this study was to evaluate the relationship of scores on the Affective dimension of the MP0 to the 191 outcome of subjects who had participated in Mary Free Bed’s multidisciplinary pain clinic. The subjects were grouped by type of insurance—based financial support: those receiving worker’s compensation, those receiving automobile no-fault benefits, and individuals not receiving any financial support from the insurance industry. This researcher also attempted to determine whether the three treatment groups (worker’s compensation, no-fault automobile insurance, and no financial support) differed significantly in their scores on the Affective dimension of the MPQ at intake, over the course of treatment, and at follow-up. The level of scores on the Affective dimension of the MPQ at various points over the course of treatment was evaluated using the HLM statistical procedure. Three research hypotheses were developed to evaluate the concepts of interest in this study: Hypothesis 1: Chronic pain patients’ scores on the Affective dimension of the MP0 at intake will correlate positively with successful outcome following treatment in an MPC. Hypothesis 2: Subjects with high scores on the Affective dimension of the MP0 at intake will have a nonlinear pattern of Affective dimension scores over the course of treatment, upon completion of the program, and at follow—up. Hypothesis 3: There will be no difference in the treatment— success rates of the three study groups: those receiving worker’s compensation, those covered by no—fault automobile insurance, and those receiving no financial support. The Study Sample To test the hypotheses, 85 patients with chronic low back pain were selected after a review of patients’ charts. The MP0 had been 192 administered to the subjects at intake, at least three times during the course of treatment, and at least once during follow-up.‘ A letter was sent to all subjects in the sample, explaining the purpose of the study and asking them to furnish follow-up information. MFB’s follow—up questionnaire was enclosed, along with the Productivity of Life questionnaire. A stamped and addressed envelope was provided for subjects’ convenience in returning the questionnaires. Approximately 30 days following the initial request, a second request for information was mailed. After approximately 30 days, the researcher attempted to telephone the subjects who had failed to respond to the mailed request. Seventy—five subjects provided follow-up information. For comparison purposes, these subjects were divided into three groups, according to the type of insurance support they received. The three groups were (a) worker’s compensation (p = 43), auto no-fault (p = 11), and (c) no form of subsidized financial support (p = 21). Method The overall design consisted of measuring 18 dependent variables and two outcome measures. The 18 variables were historical, demographic, and related to the MPQ. Correlation coefficients were computed to determine the linear relationship between scores on the Affective dimension of the MPQ and the outcome of the subjects who had completed treatment in the MPC (Hypothesis 1). Regression analysis was used to determine the relationship between predictor variables and outcome. 193 The second hypothesis was evaluated by using the fixed-effects ANOVA procedure to evaluate continuous and categorical variables. This hypothesis was further evaluated by using time HLM procedure. The important aspect of this analysis is that ii; is an attempt to explain the growth parameter over time between subjects. The third hypothesis was analyzed by use of the chi-square test of association. The analysis procedures yielded the following results. Correlation and Regression Data The outcome data did not substantiate that chronic pain patients’ scores on the Affective dimension of the MPQ would correlate positively with successful outcome. The correlation coefficient was .01 for outcome and affect. The correlation coefficient for affect and poor productivity was .14. The only large significant correlation coefficient was that between productivity and outcome (r_ = - 93). This means there was a significant relationship between individuals returning 11) work or entering a rehabilitation program and those individuals who were productive, as productivity was measured on the PLQ. In the regression analysis, the overall percentage of variance explained was very low (32 = .14). The analysis demonstrated that, the younger an individual was, the more likely he or she was to succeed. The longer an individual had suffered from the condition, the more likely he (H’ she was in) succeed. The higher the individual’s level of education, the greater potential that person 194 had for success. If an individual was married, he (n: she had a greater chance for success. The regression analysis for productivity explained only 29% of the variability in productivity. The significant predictors in this analysis were evaluative, chronicity, client employment, marital, and sensory. Affect by Time The results of the ANOVA procedure evaluating affect over time demonstrated a significant quadratic effect and a significant cubic effect. There was a downward trend, but it seemed to sink and then rise toward the end of the time measured. HLM analyses were conducted at four time points: intake, 4 weeks, 8 weeks, and 12 weeks. The average affect level decreased from intake to 4 weeks and slowly decreased thereafter. The variability in both the intercept and the slope was significant at all four time points, indicating that the slope and intercept differed significantly between subjects. The variability in the quadratic term was not significant at any time point and was fixed at zero for the analyses. For the intake analysis, the variables litigation, chronicity, surgeries, number of words chosen, evaluative, affect at intake, and membership in time auto no-fault group *were significant predictors of either the intercept, the slope, or both. In the analysis for 4, 8, and 12 weeks, the variables found to be significant were litigation, chronicity, surgery, evaluative, base affect, and membership in the auto 195 no-fault group. The results indicated that the differences in the intercept were not fully explained at any of the four time points. The differences in the week slope were explained by the variables. Success-by-Compensation Analysis The result of the chi-square test of association was not significant (chi-square = .77). This indicated there was no relationship between compensation group and outcome. An ANOVA was performed to determine whether group differences existed on the productivity variable. No significant differences were found. This means that the worker’s compensation group, the auto no-fault group, and the no-financial-support group did not differ on either outcome variable. Discussion This section contains a discussion of the conclusions drawn from the results of the statistical analyses, within the limitations of the study. A major conclusion that can be drawn from this research is that no single variable was significant in predicting successful outcome following treatment in an MPC. To further evaluate factors relating to successful outcome, a regression analysis was performed, using the most significant variables; this analysis explained a very low overall percentage of variance. The regression analysis demonstrated that, the younger an individual was, the more likely he or she was to succeed. The longer the individual had suffered from 196 the condition, the more likely he or she was to succeed. The higher an individual’s level of education, the more likely that person was to succeed. If an individual was married, his or her chance for success was greater. The variables age and chronicity have been negatively correlated in previous studies. The percentage of variance explained in this regression analysis was very low ([2 = .14). It appears that the significance of age and chronicity may be independent of each other. The very low percentage of explained variance indicates idiosyncratic characteristics may be more important in predicting outcome. The results of this research supported the findings of some previous prediction studies. The studies it supported were those by Arnoff and Evans (1979), whose results indicated a negative correlation of age to success, and Maruta et al. (1979), who found that the fewer operations a patient had, the better was his or her chance for success. In contrast, the results of the present research refuted those of previous research regarding the length of time an individual had suffered from the impairment. Maruta et al. (1979) said that individuals with pain of fewer than 3 years’ duration did better than individuals with pain complaints of longer duration. Keefe et a1. (1981) indicated that individuals with continuous pain for a shorter period of time experienced the greatest pain relief. The findings of the present research suggest that individuals will benefit from an MPC pain—treatment program, regardless of the length of time they have suffered from chronic pain. 197 The controversial concept of compensation has been reviewed in four previous studies. Dworkin et a1. (1985) stated that compensation benefits predicted poor short-term outcome but that compensation and litigation did not predict successful long-term outcome. Dworkin et al. (1986) stated that nondepressed compensation recipients did not do as well as nondepressed subjects who were not receiving worker’s compensation. Guck et al. (1986) indicated that successfully treated patients were less likely to receive compensation than were unsuccessfully treated patients. Melzack et a1. (1985) indicated that patients receiving compensation had lower affective scores than those receiving no compensation. The results from the present study indicated that there was no relationship between affect and outcome. The results also indicated there was no relationship between compensation group and outcome. A major conclusion that can be drawn from the findings of this research is that the level of affect of chronic low back pain patients at intake cannot be used as a predictor of successful outcome following treatment in an MPC. It does appear that the level of affect, as measured on the Affective dimension of the MP0, will vary over the course of treatment. The HLM analyses demonstrated that the variables predicting intercept were litigation, number of words chosen, and base affect. The variables that predicted slope at intake were litigation, chronicity, surgeries, the evaluative dimension of the MP0, base affect, and membership in the auto no-fault group. The analysis for 198 the 4-week treatment interval demonstrated that litigation and base affect were significant. At the 8—week level, litigation and base affect were significant in predicting the intercept; also included were surgeries, evaluative, and membership in the auto no-fault group. In the 12-week analysis, litigation and base affect were significant variables for the intercept. Chronicity, evaluative, and surgery were also added as predictors. The most significant predictor for the intercept at intake was base affect. At 4 and 8 weeks, the most significant predictors were litigation and base affect. At 12 weeks, the most significant predictors were surgery and litigation. Base affect was significantly involved in the affective score during the weeks of treatment, but it did not appear to be significant in the weeks immediately following treatment. The significant involvement of affect during treatment may occur as a result of emotional issues being addressed at that time. The existence of emotional issues in conjunction with the possible termination of benefits, the return- to-work issue, and a significant change in lifestyle may be reflected by the affective dimension. Litigation appeared to be important, not so much at intake, but as treatment was provided, as well as after. The significance of litigation at these points may occur as the result of individuals obtaining legal advice about their rights and reducing the level of affect that they have over possible uncertainty in their future. The most significant predictors for slope at all time points were surgery and base affect, demonstrating that affect at intake 199 was a significant component for predicting decrease in affect through the course of treatment and at follow—up. The level of affect at intake may also mean that, the higher the level of affect, the further it is able to decrease, as a function of a regression to the mean. Higher slopes were found for people who had had surgery. The chi-square homogeneity—of—variance tests indicated that the differences in the intercept were not fully explained in any of the four tables. In contrast, the differences in the slope were explained by the significant variables identified above. The amount of variability explained as a result of the chi-square homogeneity- of—variance tests did not tell why someone was higher or lower on affect. However, the linear pattern could be explained by the significant variables. The researcher could not fully explain why someone was higher or lower on affect, but the linear pattern could be explained for everyone. At this point it is important to relate the findings of this research to the field of pain treatment. The demographic characteristics of this sample accurately represent the population of individuals suffering from chronic low back pain. This typical sample was analyzed and the results did not demonstrate any significant variable or characteristic that predicted successful outcome. This demonstrates that chronic pain is a complicated issue. The various theoretical approaches and treatment modalities demonstrate that there is no single correct or singularly appropriate approach to the treatment of chronic pain. 200 Historically, we have seen worker’s compensation recipients omitted from research because they were believed to lack the proper motivation to succeed. Since the inclusion of worker’s compensation recipients in MPCs, researchers have obtained mixed results over time, leading to the inclusion of a favored population. The population studied appears to be characteristic of those included in other research projects analyzing chronic low back pain. The important point to remember here is that the individuals participating in the study all had insurance benefits to cover the cost of treatment. If individuals suffering from chronic low back pain without insurance benefits were offered the same type of treatment as offered to this population, would the results be similar? Another question that comes to mind is whether the individuals rejected as candidates for the program fared as well as those accepted in the program. Obviously, the comparison of accepted versus rejected candidates poses a complicated research design and follow-up procedure that would be extremely expensive as far as time and resources were concerned. But, if different people were admitted into treatment programs for chronic pain, would the results be the same? It does not seem as if the demographic variables would be so different for those individuals not considered for the program as a result of limited resources. It appears that administrators of MPCs are very selective in allowing access to the specific programs. One of the criteria involved in selecting a treatment population involves an attempt to predict those individuals who have characteristics deemed 201 appropriate by the administrators of the program, i.e., those who will be successful. Of course, we then have the difficulty of defining success. The first problem this points out is that researchers and treaters have decreased the population in an effort to increase success levels. Another area of concern in the selection process of MPCs is the emphasis on return to work as a measure of success. The successful\‘ return to work of subjects receiving treatment in an MPC can be perceived as a financial return for the investment in the program. A financial payback for the funding source allows the investment in the program to be returned. The attention that funding sources attach to ifinancial return (n1 their investment diverts attention from the multitude of benefits an MPC is able to provide to a more diverse population. This research presented aoua zuoo go: oz: z w.z o.n w.n 0.0 0 oz 02 ummwz on c« pounouuaouom 02 m j Locos so; n._ 0.0 n.m 0.5 w mesa” n. mo» oz 02 oz 0 o._ Q.m m.m N_.nn 0.5 mac 00> oz mu» 02 oz K mm": '7 A1 5:9: .6: m._ 0.n m.m Locke m.0 n mosfiu n_ moz oz oz oz 0 o>quusooum zfiuumcuno: . c._ ._.m n.n 0 m m we» may we» oz oz n _._ 0.~ 0.5 I I n.q q mo> oz no we» 02 mw> oz 4 N.0 0.0 0.0 w.~ n no» mo> mm» may oz n Nm.m~ II III ~_u: moaou>auom n we use Aueuuiuunav m.o 0.0 0.0 V. 0:0uo .A m.~ N no» N “mooz no :0 owuomquzuumm mo» N w>auo=vob~ “no: moauq>auue n Jan go: as; Aesop-Joaav 10 0.0_ 0.0. _4 _ my» 0 gnaw“ on 5 032:0:me 2; z .>mo muzmawz muzwfiuz ozone xcaz xcmz Hoczmfiwg o>aou< wwcfixmsoeor Nmzofiumuacmmuo ~Aoonum ~ucoexOAoau uuouo emu . c . . v m. nave: _ can: >uq>auuzuocc coo: Hmooz “a“ mcfiumo«ufiuzoo compo; mo: xcmz ucuuzufioz mwcfixcoz maauu>auu< mo5o0uao zuu>wuo=00um 0o mwcfiuzuqoz vca mmcuxcmx APPENDIX E MARY FREE BED FOLLOW—UP QUESTIONNAIRE 212 PAIN REHABILITATION PROGRAM FOLLOW-UP INFORMATION DATE: PATIENT'S NAME: 1 MONTH 3 MONTHS 6 MONTHS 12 MONTHS MEDICAL RECORD H: DATE OF DISCHARGE: Please fill out the information or check the box which is most appropriate. If you a:e not able to answer a question or do not care to answer a question, please Indicate by checking the N/A box. were you working at the time of discharge from our program? No I 1 N/A I l 2. Are you presently working? Yes I No I l 1 N/A I I If yes, list employer, job title, length of employment, toleration of job duties, and hours worked per week. If not working, are you actively involved in a job search or job club? Yes I T No [ 1 N/A I l ANSVER QUESTIONS 4-7 ONLY IF YOUR COVERAGE HAS THROUGH WORKER'S COMPENSATION OR AUTO-NO-FAULT. If you are not working, are you involved with vocational rehabilitation? Yes I 1 No I 1 N/A I 1 If no, explain. 1 Are you working with a rehabilitation specialist? Yes I No I 1 N/A [ 1 Are you involved in a retraining program? Yes I 1 No I 1 A I 1 If yes, list type and length? Have you obtained second-injury certification? Yes I 1 No I 1 N/A I ] 10. 11. 213 Rate the Intensity of your pain on a 0-100 scale where 0 is no_pain and 100 is pain as bad as it could be. Average pain over past week Worst pain over past week Least pain over past week Present pain intensity Comfort level NOVICE) Circle all those words that describe Leave out any group where there are Some of the words below describe your pain. your pain at any time during the last week. no words that describe your pain. 1 2 3 4 Flickering Jumping Pricking Sharp Quivering Flashing Boring Cutting Pulsing Shooting Drilling Lacerating Throbbing Stabbing Beating Lancinating 5 6 7 8 Pinching Tugging Hot Tingling Pressing Pulling Burning Itchy Gnawing Wrenching Scalding Smarting Cramping Searing Stinging Crushing 9 10 ll 12 Dull Tender Tiring Sickening Sore Taut Exhausting Suffocating Hurting Rasping Aching Splitting Heavy 13 14 15 16 Fearful Punishing Wretched Annoying Prightful Grueling Blinding Troublesome Terrifying Cruel Miserable Vicious Intense Killing Unbearable 17 18 19 20 Spreading Tight Cool Nagging Radiating Numb Cold Nauseating Penetrating Drawing Freezing Agonizing Piercing Squeezing Dreadful Tearing Torturing Are you using the stress management techniques which you learned in the program? Frequently I 1 Have you had periods Sometimes I 1 time since Seldom I l leaving the Never I I program where you have been depressed or anxious to the point where it interfered with your daily functioning? (If you have not been questionnaire, answer No.) Yes I No I l depressed or anxious since N/A I you last completed a follow~up 12. 214 Are you currently using any pain medication? Yes I No I I N/A I I I If yes, please list: Medication Dosage How often Have you seen a physician for your pain problem since discharge? Yes No I I N/A I I If yes, how many times? Are you sleeping well at night? Yes I No I I N/A I I Have you developed any new pain problems? Yes I No I I I N/A I I If yes, explain. Have you had any further surgery for your pain problem or are you planning on it? Yes I I No I I N/A I I If yes, date and procedure: APPENDIX F APPROVAL LETTER FROM THE UNIVERSITY COMMITTEE ON RESEARCH INVOLVING HUMAN SUBJECTS 215 MICHIGAN STATE UNIVERSITY OFFICE Of VICE PRESIDENT FOR RESEARCH EAST LANSING I MICHIGAN ' ARIN-1N6 AND DEAN Of THE GRADUATE SCHOOL February 20, 1991 Mr. Paul Delmar 7922 Pine Edge Court Alto, MI 49302 RE: RELATIONSHIP OF AFFECT TO TREATMENT AND OUTCOME IN CHRONIC PAIN, IRB#9l—063 Dear Mr. Delmar: The above project is exempt from full UCRIHS review. The proposed research protocol has been reviewed by another committee member. The rights and welfare of human subjects appear to be protected and you have approval to conduct the research. You are reminded that UCRIHS approval is valid for one calendar year. If you plan to continue this project beyond one year, please make provisions for obtaining appropriate UCRIHS approval one month prior to February 11,1992. Any changes in procedures involving human subjects must be reviewed by UCRIHS prior to initiation of the change. UCRIHS must also be notified promptly of any problems (unexpected side effects, complaints, etc.) involving human subjects during the course of the work. Thank you for bringing this project to my attention. If I can be of any future help, please do not hesitate to let me know. Sincerely, Da id E. Wright, Ph. D. Chair, UCRIHS DEW/deo cc: Dr. Nancy Crewe AIYIInmAI/ ,- .1 r m , APPENDIX G MARY FREE BED PROGRAM-—ORIGINAL INFORMATION SHEET Please print or type information on this form. 216 PAIN REHABILITATION PROGRAM Mary Free Bed Hospital and Rehabilitation Center 235 Wealthy SE Grand Rapids, Michigan 49503 616-242-9204/242-0482 Check boxes where appropriate. Name Please use black ink if possible. Date Address Type of Insurance Coverage: (Street) (City) (State) (Zip) Telephone ( ) Social Security I Date of Birth Age 1 Where in your body do you experience pain? 2. From whom did you hear about this program? 3. a. Is your pain problem the result of an injury? Yes No b. If yes, date of injury Location Description c. If no, when did your pain problem start? Cause? b. Dates of surgery (if any) How many times have you had surgery for your pain problem? Who has treated you for your pain problem? a. Who is your present treating physician? b. Who is your family physician? What have your doctors told you about the pain? 217 Application, Continued Page 2 8. Have you had a pain problem in the past (prior to this current problem)? Please explain: 9. a. How many days have you been in the hospital for pain problems in the past year? days b. How many times do you estimate you went to your doctor's office because of your pain problem during the last 12 months? times c. How many times do you estimate you went to the emergency room for treatment of your pain problems in the last 12 months? times d. Please estimate how many times in the last 12 months that you received physical therapy treatments for your pain problems. times 10. Please list all prescription and non-prescription medications you now take. I 11. Do you smoke? Yes No 12. a. Estimate the number of hours you are lying down each day. hours b. Do you feel that currently you do at least 80% of the homemaking tasks around around the house? Yes No c. Do you consider yourself active? Yes No d. Do you consider yourself retired? Yes No 13. How often do you complete an exercise program? Daily Weekly Infrequently Not at all 14. Please list the treatments you have had for your pain problem. Put an * by any treatment that seemed to help. 218 Application, Continued Page 3 15. a. Circle any of the following that made your pain worse. sitting standing running lying down driving a car coughing straining bending lifting twisting b. Add any activity which worsens the pain (not listed above): c. Circle any of the following that helps your pain. rest heat massage exercise brace adjustments hot baths medicines (Please list): d. Add anything else that helps (not listed above): 16 If the pain was not such a major focus of your life, what would you do differently with your time? \ a. I ‘ d. b. e. c. f. 17. a. Some of the words below, which are divided into 20 categories, may describe your pain. Please read each group carefully. Circle all words in each group that describe your pain. Leave out any groups where there are no words that describe your pain. 1 2 3 4 Flickering Jumping Pricking Sharp Quivering Flashing Boring Cutting Pulsing Shooting Drilling Lacerating Throbbing Stabbing Beating Lancinating 5 6 7 8 Pinching Tugging Hot Tingling Pressing Pulling Burning Itchy Gnawing Wrenching Scalding Smarting Cramping Searing Stinging Crushing 9 10 ' 11 12 Dull Tender Tiring Sickening Sore Taut Exhausting Suffocating Hurting Rasplng Aching Splitting Heavy 21S) Application, Continued Page 4 13 14 15 ' 16 Fearful Punishing Wretched Annoying Frightful Grueling Blinding Troublesome Terrifying Cruel Miserable Vicious Intense Killing Unbearable 17 I 18 19 20 Spreading Tight Cool Nagging Radiating Numb Cold Nauseating Penetrating Drawing Freezing Agonizing Piercing Squeezing Dreadful Tearing Torturing b. Are there any words which are not listed above which you think describe your pain? Yes No 18. a. On a scale of 0-100, where 0 is no pain and 100 is pain as bad as it could be, what would you estimate your average or usual pain level has been in the past week? b. Using the same scale of 0—100, what would you estimate your least pain has has been in the past week? ‘ c. Again, using the same scale of 0-100, what would you estimate your worst pain has been in the past week? 19. What do you expect to achieve by participating in this program? 20. a. Marital Status: Single Married Separated b. If married, is your Spouse employed? Yes No c. If yes, what is your spouse's occupation? d. Number of children? e. List all people besides yourself that live in your home: RELATIONSHIP 21. a. Are you currently working? Yes No b. If no, has your pain problem affected your ability to work? Explain: 21. 220 Application, Continued Page 5 c. If you are employed, how many hours per week are you working? hours d. If you are not working, how many months has it been since you worked last? months a. Present or most recent employer: Name: Supervisor: Address: Street City State Zip Phone Job Title: Wages: b. Dates Employed Fro _________ To: How many days absent in the past month. c. Previous Employment: Name of Employers Supervisor Job Title Wages Dates I I I I | I I T i I I I I I I. I I d. Job enjoyed the most: w ‘5 N L11 N m Please check as appropriate: plan on going back to my old job. plan on going back to the same company but a different job. need to look for a new job. consider myself totally and permanently disabled and thus unable to work. choose not to work. ther, please explain: I I I I I 0 Please check as appropriate: an to eventually work full time (40-hour week). I plan to eventually work part time (less than 40 hours a week). The biggest barriers to my going back to work are: (Check as appropriate) My strength and endurance My employment history My education My work skills The fast pace of my old job. Fear of losing compensation or The company's attitude disability income My attitude My lawyer My family's attitude Loss of income My health history Fear of hurting myself further Other, please explain: The economy IIIIII School: a. Circle the highest grade completed Grammar/High School 1 2 3 4 5 6 7 8 9 10 ll 12 College 2 3 4 5 6 Trade/Business School other Application, Continued Page 6 26. c. Are you currently involved in a school program? Yes No d. If yes, please describe: 27 a. Are you currently involved in any legal activity as a result of oyour pain? (For example. Lawsuit, compensation litigation) Yes If yes, please describe: b. Lawyer's Name: (___I Street City State Zip Telephone c. Date of pending legal action: d. Do you plan any future legal action? Yes No If yes, please explain: Are you currently a client of Michigan Rehabilitation Services or a private ‘ rehabilitation company? Please specify: a. Please check all sources of income that apply to you and your spouse. Wages (earned income) Applicant Worker‘s Compensation Applicant Spouse Auto No-Fault Applicant Spouse Social Security Disability Applicant Spouse Unemployment Benefits Applicant Spouse Social Security Retirement Applicant Spouse Pension Applicant Spouse Public Assistance Applicant Spouse other (please specify): Applicant Spouse b. Total Monthly Income Before Taxes : 3 Please show your current medical insurance coverage: Be sure to include all policy numbers. a. Blue Cross/Blue Shield: Group I Service Code Plan I Contract I b. Medicaid: IDI c. Medicare: Policy I d. Private Insurance: Policy I Name of Insurance Company Address Adjuator's Name Telephone ( 222 Application, Continued Page 7 e. Worker's Compensation/Auto No Fault: (Please circle one.) Policy I Na e of Insurance Company Address Adjustor's Name: Telephone ( ) 31. Please list all doctors, hospitals and clinics from whom you will be requesting medical information. (We have enclosed medical release forms that need to be signed and forwarded to the appropriate facilities.) Name Address Phone I Last Seen On (DO-00"” 32. Please list all x-ray departments, hospitals, or doctors from whom you will be requesting x—rays. (We have enclosed x-ray release forms that need to be signed and forwarded to the proper facility.) \ Date x—rays were Name Address Type of x-ray taken (month/year) I understand the above information is confidential and will be shared only with the Mary Free Bed Pain Rehabilitation Team to aid them in evaluating my pain problems. Furthermore, I also understand that it is my responsibility to contact my doctors and arrange for them to forward to you all pertinent medical records. I understand once you have received this form, my completed medical histories from doctors I have seen for my pain problems, x-rays, and insurance clarification, that you will contact me to set up my initial evaluation. Finally, I have discussed with my spouse the requirement that he/she will actively be involved with my treatment program. He/she is prepared to participate in my initial evaluation and will attend meetings later at the hospital depending on treatment recommendations. Applicant's Signature Spouse's Signature Date 223 Application, Continued Page 8 Mark the areas on your body where you feel the described sensations. Use the appropriate symbol. Include all affected areas. Also, using the scale at the bottom of the page, place an X on the scale where you would put your average daily amount of pain. Numbness ----- Pins and Needles Aching . 000 000 000 000 000 Stabbing / / / / / ///// / / / / / V A O\ 10\ 20\ 30\ 40x 50\ 60% 70% 80\ 90% 100‘ No Unbearable Pain Pain APPENDIX H DATA SUMMARY FORM 224 DATA SUMMARY FORM NAME: ADDRESS: TELEPHONE ‘_ DATE OF INTAKE: AGE: EDUCATION: MARITAL STATUS: SPOUSE EMPLOYED: CLIENT EMPLOYED AT INTAKE: LITIGATION: REHABILITATION INVOLVED: SOURCES OF INCOME: LENGTH OF CHRONICITY: HOSPITALIZATIONS: NUMBER OF PRIOR SURGERIES: NUMBER OF PRIOR EMPLOYERS: WC/ NO FAULT / OTHER: MPQ PPI: NWC: S A E M NUMBER OF WEEKS IN PROGRAM DATE/ AFFECT SCORE % FOLLOW UP INFORMATION REFERENCES REFERENCES Anderson, T., Cole, T., Gullickson, G., Hudgens, A., & Roberts, A. (1977). Behavior modification of chronic pain: A treatment program by a multidisciplinary team. Clinical Orthopaedics, 129, 96-100. Arawaka, K. (1981). Outpatient pain clinic. Journal of the Kansas Medical Society, 82, 292-294. Aronoff, G. M., & Evans, w. 0. (1982). The prediction of treatment outcome at a multidisciplinary pain center. Pain, 14, 67-73. Atkinson, J. H., Kremer, E. F., & Ignelzi, R. J. (1982). Diffusion of pain language with affective disturbance confound differen— tial diagnosis. Pain, 12, 375-384. Bandura, A. (1969). Principles of behavior modification. New York: Holt, Rinehart, & Winston. Beecher, H. K. (1946). Pain in men wounded in battle. Bulletin of the U.S. Army Medical Department, 5, 445. Beecher, H. K. (1956). The relationship of significance of wound to pain experienced. Journal of the American Medical Associa- tion, 161, 1609. Beecher, H. K. (1959). Measurement of subjective responses. Quantitative effects of druqs. New York: Oxford University Press. Belar, C., & Cohen, J. (1979). The use of EMG feedback and pro— gressive relaxation in the treatment of a woman with chronic back pain. Biofeedback and self-regulation, 4, 345—353. Bigos, S. J., Battie, M. C., Spengler, D. M., Fisher, L. D., Fordyce, w. E., Hanson, T. H., Nachemson, A. L., & Nortley, M. D. (1991). A prospective study of work perceptions and psychosocial factors affecting the report of back injury. Spine, 16, 1-6. Bishop, G. H. (1940). Neural mechanisms of cutaneous sense. Physiological Review, 26, 77-102. 225 226 Block, A., Kremer, E., & Gaylor, M. (1980). Behavioral treatment of chronic pain: Variables affecting treatment efficacy. Pain, 8, 367-375. Bonica, J. J. (1973). Fundamental considerations of chronic pain therapy. Post-Graduate Medicine, 88, 81-85. Bonica, J. J. (1974). Advances in neurology. New York: Raven Press. Borg, N. R., & Gall, M. D. (1971). Educational research: An introduction. New York: David McKay. Britton, N. F., & Skevington, S. M. (1989). A mathematical model of the gate control theory of pain. Journal of Theoretical Biology, 137, 91-105. Bryk, A. S., & Raudenbush, S. W. (1988). Towards a more appropri- ate conceptualization of research on school effects: A three- level hierarchical linear model. America Journal of Education, 81, 65-108. Bryk, A. 3., Raudenbush, S. H., Congdon, R., & Seltzer, M. (1986). An introduction to HLM: Computer proqram and user’s guide. Chicago: University of Chicago, Department of Education. Bryk, A. S., & Weisberg, H. I. (1977). Use of the nonequivalent control group design when subjects are growing. Psychological Bulletin, 88, 950-962. Buren, J. V., & Kleinknecht, R. A. (1979). An evaluation of the McGill Pain Questionnaire for use in dental pain assessment. Pain, 8, 23-33. Cairns, D., & Pasino, J. (1977). Comparison of verbal reinforce- ment and feedback in the operant treatment of disability due to low back pain. Behavior Therapy, 8, 621-630. Cairns, D., Thomas, L., Mooney, V., & Pace, 0. (1976). A compre- hensive treatment approach to chronic low back pain. Pain, 8, 301-308. Carlsson, A. M. (1984). Assessment of chronic pain: II. Problems in the selection of relevant questionnaire items for classifi- cation of pain and evaluation and prediction of therapeutic effects. Pain, 18, 173-184. Cautela, J. (1977). The use of covert conditioning in modifying pain behavior. Journal of Behavior Therapy and Experimental Psychiatry, 8, 45-52. 227 Chapman, C. R., Casey, K. L., Dubner, R., Foley, K. M., Gracely, R. H., & Reading, A. E. (1985). Pain measurement: An over- view. Pain, 88, 1-31. Chapman, S., Brena, S., & Bradford, A. (1981). Treatment outcome in a chronic pain rehabilitation program. Pain, 11, 255-268. Chapman, w. P., Finesinger, J. E., Jones, C. M., & Cobb, S. (1947). Measurements of pain sensitivity in patients with psychoneurosis. Archives of Neurology and Psychiatry, £1, 321-331. Cicala, R. S., & Wright, H. (1989). Outpatient treatment of patients with chronic pain: An analysis of cost savings. Journal of Pain, 8, 223-226. Cinciripini, P. M., & Floreen, A. (1982). An evaluation of a behavioral program for chronic pain. Journal of Behavioral Medicine, 8, 375-389. Connally, G. H., & Sanders, S. H. (1991). Predicting low back pain patients to lumbar sympathetic nerve blocks and interdiscipli- nary rehabilitation: The role of pretreatment overt pain behavior and cognitive coping strategies. Pain, 88, 139-146. Crockett, D. J., Prkachin, K. M., & Craig, K. D. (1977). Factors of the language of pain in patient and normal volunteer groups. Pain, 8, 175—182. De Benedittis, G. (1979). A new strategy for chronic pain control: Preliminary results (Part 11). Journal of Neurosurqical Sciences, 88, 191-200. Dworkin, R. H., Handlin, D. S., Richlin, D. R., Brand, L., & Vannucci, C. (1985). Unraveling the effects of compensation, litigation, and employment on treatment response in chronic pain. Pain, 88, 49-59. Dworkin, R. H., Richlin, D. M., Handlin, D. S., & Brand, L. (1984). Chronic pain and depression. Pain, Supplement, 8, $208. Dworkin, R. H., Richlin, D. M., Handlin, D. S., & Brand, L. (1986). Predicting treatment response in depressed and non-depressed chronic pain patients. Pain, 81, 343-353. Ellis, A. (1962). Reason and emotion of psychotherapy. New York: Lyle Stuart. Fey, S. G., & Fordyce, w. E. (1983). Behavioral rehabilitation of the chronic pain patient. Annual Review of Rehabilitation. 228 Fields, H. L. (1987). Pain. New York: McGraw-Hill. Flor, H., Haag, G., Turk, D. C., & Koehler, H. (1983). Efficacy of EMG biofeedback, pseudotherapy and conventional medical treat- ment for chronic rheumatic back pain. Pain, 11, 21-31. Fordyce, W. E. (1976). Behavioral methods for chronic pain and illness. St. Louis: C. V. Mosby. Fordyce, W. (1982). A behavioural perspective on chronic pain. British Journal of Clinical Psychology, 21, 313-320. Fordyce, W., Brena, S., de Lateur, B., Holcombe, S., & Loeser, J. (1978). Relationship of patient semantic pain descriptors to physician diagnostic judgments, activity level measures and MMPI. Pain, 8, 293-303. Fordyce, W. E., Fowler, R. S., & de Lateur, B. (1968). An appli- cation of behavior modification techniques to a problem of chronic pain. Behavior Research and Therapy, 8, 105-107. Fordyce, W., Fowler, R., Lehmann, J., & de Lateur, B. (1968a). Some complications of learning in problems of chronic pain. Archives of Physical Medicine and Rehabilitation, 88, 399-408. Fordyce, W., Fowler, R., Lehmann, J., & de Lateur, B. (1968b). Some implications of learning in problems of chronic pain. Journal of Chronic Disease, 81, 179-190. Fordyce, W. E., Fowler, R., Lehmann, J., de Lateur, 8., Sand, P., & Treischman, R. (1973). Operant conditioning in the treatment of chronic pain. Archives of Physical Medicine and Rehabilita— tion, 88, 399-408. Fordyce, W. E., Lansky, D., Calsyn, D. A., Shelton, J. L., Stolou, W. C., & Rock, D. L. (1984). Pain measurement and pain behavior. Pain, 8, 53-69. Fordyce, W. E., Roberts, A. H., & Steenbach, R. A. (1985). The behavioral management of chronic pain: A response to critics. Pain, 88, 113-125. Freeman, C., Calsyn, D., Paige, A., & Halar, E. (1980). Biofeed- back with low back pain patients. American Journal of Clinical Biofeedback, 8, 118-122. Gentry, W., & Bernal, G. (1977). Chronic pain. In R. Williams & W. Gentry (Eds.), Behavioral approaches to medical treatments. Cambridge, MA: Ballinger. 229 Gerard, R. W. (1951). The physiology of pain: Abnormal neuron states in causalgia and related phenomena. Anesthesiology, 18, 1-13. Glass, G. V., & Hopkins, K. D. (1984). Inferences regarding proportions. In Statistical methods in education and 95V- chology, (pp. 285-290). Englewood Cliffs, NJ: Prentice-Hall. Goldfried, M. R. (1977). The use of relaxation and cognitive relabeling as coping skills. In R. B. Stuart (Ed.), Behavioral self-management: Strategjes. techniques. and outcome. New York: Brunner/Mazel. Gottlieb, H., Alperson, B., Koller, R., & Hockersmith, V. (1979). An innovative program for the restoration of patients with chronic back pain. Physical Therapy, 52, 996—999. Gottlieb, H., Strite, L., Koller, R., Madorsky, A., Hockersmith, V., Kleeman, M., & Wagner, J. (1977). Comprehensive rehabilita- tion of patients having chronic low back pain. Archives of Physical Medicine and Rehabilitation, 88, 101-108. Graham, C., Bond, S. S., Gerkousch, M. M., & Cook, M. R. (1980). Use of the McGill Pain Questionnaire in the assessment of cancer pain: Replicability and consistency. Pain, 8, 377-387. Greenhoot, J., & Sternbach, R. (1974). Conjoint treatment of chronic pain. In J. Bonica (Ed.), Advances in neurology: Pain (vol. 4). New York: Raven Press. Grzesiak, R. (1977). Relaxation techniques in the treatment of chronic pain. Archives of Physical Medicine and Rehabilita- tion, 88, 302-304. Guck, T. P., Skultetly, F. M., Meilman, P. W., & Dowd, E. T. (1985). Multidisciplinary pain center follow-up study: Evaluation with a no-treatment control group. Pain, 81, 295-306. Guck, T. P., Skultetly, R. M., Meilman, P. W., & Dowd, E. T. (1986). Prediction of long-term outcome of multidisciplinary pain treatment. Archives of Physical Medicine and Rehabilita- tion, 81, 293-296. Guyton, A. C. (1971). Textbook of medical physiology (4th ed.). Philadelphia: W. B. Saunders. Hammonds, W., Brena, S., & Unikel, I. (1978). Compensation for work-related injuries and rehabilitation of patients with chronic pain. Southern Medical Journal, 18, 664-666. 230 Harrison, A. (1988). Arabic pain words. Pain, _8, 239—250. Hartman, L., & Ainsworth, K. (1980). Self-regulation of chronic pain. Qgpadian Journal of Psychiatry. 25, 38-43. Haugen, F. P. (1956). Current concepts of the pain process. Journal of Chronic Diseases, 4, 9. Head, H. (1920). Studies in neurology. London: Kegan Paul. Heaton, R. K., Getto, C. J., Lehmann, R. A. W., Fordyce, W. E., Braver, E., & Groban, S. E. (1982). A standardized evaluation of psychosocial factors in chronic pain. Pain, 18, 165-174. Hendler, H., Derogatis, L., Avella, J., & Long, D. (1977). EMG biofeedback in patients with chronic pain. Diseases of the Nervous System, 88, 505-509. Herman, E., & Baptiste, S. (1981). Pain control: Mastery through group experience. Pain, 18, 79-86. Hill, R. (1952). Studies of anxiety associated with anticipation of pain. Archives of Neurological Psychiatry, 61, 612. Hudgens, A. (1979). Family-oriented treatment of chronic pain. Journal of Marital and Family Therapy, 5, 67-78. Hunter, M., & Philips, C. (1981). The experience of headache--An assessment of the qualities of tension headache pain. Pain, 18, 209-220. Hunter, M., Philips, C., & Rachman, S. (1979). Memory for pain. Pain, 8, 36-46. Hurri, H. (1989). The Swedish back school in chronic low back pain. Part 11: Factors predicting the outcome. Scandinavian Journal of Rehabilitation Medicine, 88(1). & Timmermans, G. (1977). The pain ward Ignelzi, R., Sternbach, R., Pain, 8, 277-280. follow-up analyses. Jones, A., & Wolf, S. (1980). Treating chronic low back pain. Physical Therapy, 88, 58-63. Keats, A. S., & Lane, M. (1963). The symptomatic therapy of pain. Boston: Yearbook Medical Publishers. Keefe, F. J., Block, A. R., Williams, R. B., & Surwit, R. S. (1981). Behavioral treatment of chronic low back pain: Clinical outcome and individual differences in pain relief. Pain, 11, 221-231. 231 Keefe, F. J., Crisson, J., Urban, B. J., & Williams, D. A. (1990). Analyzing chronic low back pain: The relative contribution of pain coping strategies. Pain, 88, 293-301. Keefe, F. J., Schapira, B., Brown, C., Williams, R. B., & Surwit, R. S. (1981). EMG-assisted relaxation training in the manage- ment of chronic low back pain. American Journal of Clinical Biofeedback, 3, 93-103. Kelly, D. D. (1985). Central representations of pain analgesia. In E. R. Kandel & J. H. Schwartz (Eds.), Principles of neural science (2nd ed.). New York: Elsevier. Ketovuori, H., & Rush, J. (1978). A pain vocabulary in Finnish: The Finnish Pain Questionnaire. Pain, 18, 247—253. Khatami, M., & Rush, J. (1978). Pilot study of the treatment of outpatients with chronic pain. Pain, 8, 163-172. Khatami, M., Woody, G., & O’Brien, C. (1979). Chronic pain and narcotic addiction: A multitherapeutic approach. Comprehen- sive Psychiatry, §£, 165-168. Kleinke, C. K., & Spangler, A. S. (1988). Predicting treatment outcome of chronic back pain patients in a multidisciplinary pain clinic: Methodological issues and treatment implications. Pain, 88, 41-48. Klepac, R. K., Dowling, J., & Hauge, E. (1981). Sensitivity of the McGill Pain Questionnaire to intensity and quality of labora- tory pain. Pain, 18, 199-207. Klepac, R. K., Dowling, J., Rokke, P., Dodge, L., & Schafer, L. (1981). Interview vs. paper and pencil administration of the McGill Pain Questionnaire. Pain, 11, 241-246. Kremer, E. (1991, May 21). Mary Free Bed Hospital and Rehabilita- tion Center, Grand Rapids, MI. Personal communication. Kremer, E., & Atkinson, J. H. (1981). Pain measurement: Construct validity of the affective dimension of the McGill Pain Ques- tionnaire with chronic pain patients. Pain, 11, 93-100. Kremer, E., & Atkinson, J. H. (1984). Pain language: Affect. Journal of Psychosomatic Research, 88, 125-132. Kremer, E., Atkinson, J. H., & Ignelzi, R. J. (1981). Measurement of pain: Patient preference does not confound pain measure- ment. Pain, 18, 241-248. 232 Kremer, E., Atkinson, J. H., & Ignelzi, R. (1982). Pain measure- ment: The affective dimensional measure of the McGill Pain Questionnaire with a cancer pain population. Pain, 18, 153- 163. Kremer, E., Atkinson, J. H., & Kremer, A. (1983). The language of pain: Affective descriptors of pain are a better predictor of psychological disturbance than pattern of sensory and affective descriptors. Pain, 18, 185-192. Leavitt, F., Garron, D. C., Whistler, W. W., & Sheinkop, M. B. (1978). Affective and sensory dimensions of pain. Pain, 8, 273-281. Levendusky, P., & Pankratz, L. (1975). Self-control techniques as an alternative to pain medication. Journal of Abnormal Psy- chology, 88, 165-168. Lewis, T. (1942). Pain. New York: Macmillan. Linton, S. (1982). A critical review of behavioral treatments for chronic benign pain other than headache. British Journal of Clinical Psychology, 21, 321-337. Linton, S. (1986). Behavioral remediation of chronic pain: A status report. Pain, 88, 125—141. Linton, S. J., & Gotestam, K. G. (1984). A controlled study of the effects of applied relaxation and applied relaxation plus operant procedures in the regulation of chronic pain. British Journal of Clinical Psychology, 88, 291-299. Linton, S. J., & Melin, L. (1983). Applied relaxation in the management of chronic pain. Behavioral Psychotherapy, 11, 337-350. Linton, S. J., Melin, L., & Gotestam, K. G. (1984). Behavioral analysis of chronic pain and its management. Progressive Behavior Modification, 18, 1-42. Linton, S. J., Melin, L., & Stjernlof, K. (1985). The effects of applied relaxation and operant activity training on chronic pain. Behavioural Psychotherapy, 18, 87-100. Livingston, W. K. (1943). Pain mechanisms. New York: Macmillan. Loeser, J. D. (1980). Concepts of pain. In Chronic low back pain (pp. 1-46). New York: Elsevier. Loeser, J. D., & Egan, K. J. (1989). Managing the chronic pain patient. New York: Raven Press. 233 Lutz, R. W., Silbert, M., & Ohlson, N. (1983). Treatment outcome and compliance with therapeutic regimens: Long-term follow-up of a multidisciplinary pain program. Pain, 11, 301-308. Martin, J. J. (1985). Anatomical substrates for somatic sensation. In E. R. Kandel & J. H. Schwartz (Eds.), Principles of neural science (2nd ed.). New York: Elsevier. Maruta, T., Swanson, D. W., & McHardy, M. J. (1990). Three-year follow-up of patients with chronic pain who were treated in a multidisciplinary pain management center. Pain, 81, 47—53. Maruta, T., Swanson, D. W., & Swenson, W. M. (1979). Chronic pain: Which patients may a pain management program help? Pain, 1, 321-329. McCauley, J. P., Thelen, M. H., Frank, R. G., Willard, R. R., & Callen, K. E. (1983). Hypnosis compared to relaxation in the outpatient management of chronic low back pain. Archives of Physical Medicine and Rehabilitation, 88, 548-552. McCreary, C., Turner, J., & Dawson, E. (1981). Principal dimen- sions of the pain experience and psychological disturbance in chronic low back pain patients. Pain, 11, 85-92. Meichenbaum, D., & Turk, D. (1976). The cognitive-behavioral management of anxiety, anger, and pain. In Davidson (Ed.), The behavioral management of anxiety. anger. and pain (pp. 163- 182). New York: Brunner/Mazel. Melin, L., & Linton, S. J. (1988). A follow-up study of a compre- hensive behavioural treatment programme for chronic pain patients. Behavioural Psychotherapy, 1§(4), 313-321. Mellin, G., Jarvikoski, A., & Verkasalo, M. (1984). Treatment of patients with chronic low back pain, comparison between rehabilitation centre and outpatient care. Scandinavian Journal of Rehabilitation Medicine, 18, 77-88. Melzack, R. (1961). The perception of pain. Scientific American, 888, 41-49. Melzack, R. (1968). Pain. In D. L. Sills (Ed.), International encyclopedia of the social sciences. New York: Macmillan. Melzack, R. (1973). The puzzle of pain. London: Penguin Books. Melzack, R. (1975). The McGill Pain Questionnaire: Major proper- ties and scoring methods. Pain, 1, 275-299. 234 Melzack, R. (1983). Pain measurement and assessment. New York: Raven Press. Melzack, R., & Casey, K. L. (1968). Sensory, motivational and central control determinants of pain: A new conceptual model. In D. Kenshald (Ed.), The skin senses. Springfield, IL: Thomas. Melzack, R., Katz, J., 8 Jeans, M. E. (1985). The role of compen- sation in chronic pain: Analysis using a new method of scoring the McGill Pain Questionnaire. Pain, 88, 101-112. Melzack, R., & Perry, C. (1975). Self-regulation of pain: The use of alpha feedback and hypnotic training for the control of chronic pain. Experimental Neurology, 18, 452-469. Melzack, R., & Torgerson, W. S. (1971). On the language of pain. Anesthesiology, 88, 50-59. Melzack, R., & Wall, P. D. (19 ). Pain mechanism--A new theory. Science, 150, 971-979. Melzack, R., & Wall, P. D. (1970). Psychophysiology of pain. In H. Yamamura (Ed.), Anesthesia and neurophysiology. Interna- tional Anesthesiological Clinician, 8, 3-34. Miller, C., & LeLieuvre, R. B. (1982). A method of reduce chronic pain in elderly nursing home residents. Gerontologist, 88, 314-317. Newman, R., Seres, J., Yospe, L., & Garlington, B. (1978). Multi- disciplinary treatment of chronic pain. Long-term follow-up of low back pain patients. Pain, 8, 283-292. Nigl, A., & Fisher, M. (1980). Treatment of low back strain with electromyographic biofeedback and relaxation training. Psycho- somatics, 81, 496-499. Noordenbos, W. (1959). Pain. Amsterdam: Elsevier Press. Noordenbos, W., & Wall, P. D. (1981). The failure of nerve grafts to relieve pain following nerve injury. Journal of Neurologi- cal and Neurosurgical Psychiatry, 39, 1008-1073. Nouwen, A. (1983). EMG biofeedback used to reduce standing levels of paraspinal muscle tension in chronic low back pain. Pain, 11, 353-360. Nouwen, A., & Solinger, J. (1979). The effectiveness of EMG biofeedback training in low back pain. Biofeedback and Self- Regulation, 8, 103-111. 235 Painter, J. R., Seres, J. L., 8 Newman, R. I. (1980). Assessing benefits of the pain center: Why some patients regress. Pain, 8, 101-113. Pavlov, I. P. (1927). Conditioned reflexes: An investigation of the physiological activity of the cerebral cortex. G. V. Anrep (Ed.). England: Dover. Pearce, J., 8 Morley, S. (1989). An experimental investigation of the construct validity of the McGill Pain Questionnaire. Pain, 88, 115-121. Peck, C., 8 Kraft, G. (1977). Electromyographic biofeedback for pain related to muscle tension. Archives of Surgery, 112, 889- 895. Prieto, E. J., Hopson, L., Bradley, L. A., Byrne, M., Geisinger, K. F., Midax, D., 8 Marchisello, P. J. (1980). The language of low back pain: Factor structure of the McGill Pain Ques- tionnaire. Pain, 8, 11-20. Raudenbush, S. W. (1988). Educational applications of hierarchical linear models: A review. Journgl of Educational Statistics, 18, 85-116. Reading, A. E. (1979). The internal structure of the McGill Pain Questionnaire in dysmenorrhea patients. Pain, 1, 353-358. Reading, A. E. (1980). A comparison of pain rating scales. Journal of Psychosomatic Research, 88, 119-126. Reading, A. E. (1982). An analysis of the language of pain in chronic and acute patient groups. Pain, 18, 185-192. Roberts, A., 8 Reinhardt, L. (1980). The behavioral management of chronic pain: Long-term follow-up with comparison groups. Pain, 8, 151-162. Rogosa, D. R., Brand, D., 8 Zimowski. (1982). The measurement of change. Psychological Bulletin, 88, 726-748. Rose, J. E., 8 Mountcastle, V. B. (1959). Touch and kinesthesis. Handbook of Physiology, 1, 387-429. Rybstein-Blinchik, E. (1979). Effects of different cognitive strategies on chronic pain experience. Journgl of Behavioral Medicine, 8, 93-101. 236 Rybstein-Blinchik, E., 8 Grzesiak, R. (1979). Reinterpretative cognitive strategies in chronic pain management. Archives of Physical Medicine and Rehabilitation, 88, 609-612. Sanders, S. (1979). Behavioral assessment and treatment of clinical pain. In M. Hersen, R. Eisler, 8 P. Miller (Eds.), Progress in behavior modification (vol. 8). New York: Academic Press. Sanders, S. (1983). Component analysis of a behavioral treatment program for chronic low-back pain. Behavior Therapy, 18, 697- 705. Seres, J., 8 Newman, R. (1976). Results of chronic low-back pain at the Portland Pain Center. Journal of Neurosurgery, 58, 32-36. Seres, J., Newman, R., Yospe, L., 8 Garlington, D. (1977). Evalua- tion and management of chronic pain by nonsurgical means. In J. Lee (Ed.), Pain management. Baltimore: Williams 8 Wilkins. Sherrington, C. S. (1906). The integrative action of the nervous system. London: Constable. Sinclair, D. C. (1955). Cutaneous sensation and the doctrine of specific nerve energies. Brain, 18, 584-614. Sinclair, D. C. (1967). Cutaneous sensation. London: Oxford University Press. Stanton, D. (1991, May 22). Clinical Center, Michigan State University, East Lansing. Personal communication. Stein, C., 8 Mendl, G. (1988). The German counterpart to the McGill Pain Questionnaire. Pain, 88, 251-255. Sternbach, R. (1974). Pain patients: Traits and treatment. London: Academic Press. Swanson, D., Swenson, W., Maruta, T., 8 McPhee, M. (1976). Program for managing chronic pain: I. Program descriptions and charac- teristics of patients. Mayo Clinic Proceedings, 81, 401-408. Swanson, D., Maruta, T., 8 Swenson, W. (1979). Results of behavior modification in the treatment of chronic pain. Psychosomatic Medicine, 81, 55-61. Tan, S. Y. (1982). Cognitive and cognitive-behavioral methods for pain control: A selective review. Pain, 18, 201-228. 237 Taylor, C., Zlutnick, S., Corley, M., 8 Flora, J. (1980). The effects of detoxification, relaxation, and brief supportive therapy on chronic pain. Pain, 8, 319-329. Timming, R., Cayner, J., Malec, J., Harvey, R., Schwettmann, K., 8 Chosey, J. (1980). In-patient treatment program for chronic pain. Wisconsin Medical Journal, 18, 23-26. Todd, J., 8 Belar, C. (1980). EMG biofeedback and chronic low back pain. American Journal of Clinical Biofeedback, 8, 114-117. Turk, D. C., Meichenbaum, D., 8 Genest, M. (1983). Pain and behavioral medicine: A cognitive-behavioral perspective. New York: Guilford Press. Turkat, 1., 8 Adams, H. (1978). Issues in pain modification: The Cautela model. Journal of Behavior Therapy and Experimental Psychiatry, 8, 135-138. Turner, J. A. (1982). Comparison of group progressive-relaxation training and cognitive-behavioral group therapy for chronic low back pain. Journal of Consultinqpand Clinical Psychology, 88, 757-765. Turner, J. A., 8 Chapman, C. R. (1982). Psychological interven- tions for chronic pain: A critical review: I 8 11. Pain, 18, 1-46. Turner, J. A., 8 Clancy, S. (1986). Strategies for coping with chronic low back pain: Relationship to pain and disability. Pain, 88, 355-364. Turner, J. A., 8 Clancy, S. (1988). Comparison of operant behav- ioral and cognitive-behavioral group treatment for chronic low back pain. Journal of Consulting and Clinical PsychologY, 88(2), 261-266. Turner, J. A., Robinson, J., 8 McCleary, C. P. (1983). Chronic low back pain: Predicting response to nonsurgical treatment. Archives of Physical Medicine and Rehabilitation, 88(11), 560-563. Turner, J. A., 8 Romano, J. M. (1989). Cognitive behavioral therapy for chronic pain patients. In J. D. Loeser 8 K. J. Egan (Eds.), Managing the chronic pain patient. New York: Raven Press. Tyre, T., 8 Anderson, D. (1981). In-patient management of the chronic pain patient: One-year follow-up study. Journal of Family Practice, 18, 819-827. 238 Vanderiet, K., Adriaensen, H., Carton, H., 8 Vertommen, H. (1987). The McGill Pain Questionnaire constructed for the Dutch lan- guage: Preliminary data concerning reliability and validity. Pain, 88, 395-408. Varni, J. (1981). Self-regulation techniques in the management of chronic arthritic pain in hemophilia. Behavior Therapy, 18, 185-194. Varni, J., Bessman, C., Russo, D., 8 Cataldo, M. (1980). Behav— ioral management of chronic pain in children. Archives of Physical Medicine and Rehabilitation, 8;, 375-378. Varni, J., Gilbert, A., 8 Dietrich, S. (1981). Behavioral medicine in pain and analgesia management for the hemophilic child with factor VII inhibitor. Pain, 11, 121-126. Waddell, G. (1955). Somesthesis and the chemical senses. Annual Review of Psychology, 8, 119-136. Waddell, G., Pilowsky, I., 8 Bohn, M. R. (1989). Clinical assess- ment and interpretation of abnormal illness behavior in low back pain. Pain, 88, 41-53. Wang, J., Ilstrup, D., Nauss, L., Nelson, D., 8 Wilson, P. (1980). Outpatient pain clinic: A long-term follow-up study. Minne- sota Medicine, 88, 663-666. White, B., 8 Donovan, A. (1980). A comprehensive programme for the assessment and treatment of low back pain in Western Australia. In C. Peck 8 M. Wallace (Eds.), Problems in pain. New York: Pergamon. White, A. A., 8 Gordon, S. L. (1982). Synopsis: Workshop on idiopathic low back pain. Spine, 1, 141-149. Wolf, S. L., Nacht, M., 8 Kelly, J. L. (1982). EMG feedback training during dynamic movement for low back pain patients. Behavior Therapy, 1;, 395-406. Zotterman, Y. (1939). Touch, pain and tickling: An electrophysi- cal investigation on cutaneous sensory nerves. Journal of Physiology, 25, 1. MICHIGQN STQTE UNIV LIBRQRIES I: III I. I I ;‘ IU IEI 1. I II I I (I I I I II 29 008826327 ‘ ~ .‘-.,, at ”a," «u -- “w 35.... .. ,..‘... m n-» ,,\ .. ... u... ... , “1 n- . ..... 1‘ ‘ ‘ mum-um... w... “um?” . . _. .. i ., .1» ‘ v........‘....._.;‘, _. . k H \ ‘N‘ZHHMI . I‘M \ .8. -- w h”... J-‘va'lpulvn n . u. \.w.......uy \nve. “H...” H” I“. .m~...-. ... . . |-‘;Iyt\ - H., N" mu. ‘\‘\~-t) m. , v i .-.. \ H a. . I ~--..-m......‘:, v-.' x» \\ In. .. .. K‘ a. ,. ..«uL F.,. 'n . .2.31:... ... , -'.r......, .. , N ‘ ‘ I W. H n. mu. :~v:ev|l'>> . . .. .“HN‘H ‘1 " m, 1 . ‘ \ m u c. C ‘U n ‘ “mm, ,. a _ \ “mum“ “- ' . ”4.. . . .' . My}... 1 \ ~,, . ..-.. H )“I“::W:'E'\':- .. . m. ... .\ r . __ _ -.‘....i.. . I‘ I \ . .. y , ‘ . - s I .. . ‘3 . I I I. I I U .f... . ‘ -~~.,. 2.. ‘ - .. . n .» .. p.31. , ‘ u . .n. -« I .ww'u mi“. I-w- mu. p... -. 3m ‘3‘“ u ~~ 8"" . u - . -.,. 1- “mm”... INN-1...... w «u i \‘u‘ "J ' ' I “i . ”I; . ‘ I ‘ ' I 'l"'lv‘n'v‘wau r:‘~'1-’II'Y.>\»--yr .. v.1...“ ..- . v v ~"ww-‘hun. my. 'u {hw- n. \ u. u , u" ' “ "MI. .. . . .m. m.” .1... ._ H M'- ‘ ' . ”M“ H ' ‘ V- . . ~31 m , M ":3“ H ”may...“ . 1 . 1- -..- ”IA. . ...‘.. . ”1.13....” t a . {HUB-u... .~ 1 ‘ . I I H ' "In-uh”... ’71 “'L" ' ‘ ~ I . n»... . .k...‘,:.L.~.-..uu.--. any ., . . ."H "n “ML ‘ ‘ ' ' -- i -. . . . . ‘ v n. « “I", n4' . wan-U... H.,“ I; " -I Am“... .. :w‘ n u. wo- J.»- ‘ r ‘ ‘ . ....y ' ‘ mm”... - H.,...”I..." ...u..'...:pn-........u u M.,,» . .. . ’ - sun-W.,”... "I. “”1”...” N a: -. . ,. .. l]. "“M’ c or fly...“ i»7:~..-,..,.,..,,,._h “_ ' “mu-v...“ wn-A ., , . 1."... v. ”I. z.‘ .I- w... ~ . . ~“..u.,,..... " ' ' u : .-~.r-.I'.,...,!.,‘ " I. . u... . u . . I. ‘ '-- "NJ-:01”! w v.