PREDICTING SELF-REPORTED DISABILITY in CHRONIC PAIN PATIENTS with the MMPI-2-RF a Thesis Submitted to Kent State University in P

Home , Emotionality

PREDICTING SELF-REPORTED DISABILITY IN CHRONIC PAIN PATIENTS WITH THE MMPI-2-RF

A thesis submitted

To Kent State University in partial

Fulfillment of the requirements for the

Degree of Master of Arts

Jesica Leigh Rapier

December, 2014

Except for previously published materials Thesis written by

Jesica Leigh Rapier

B.A., Eastern Kentucky University, 2011

M.A., Kent State University, 2014

Approved by

Yossef S. Ben-Porath, Professor, Ph.D., Psychological Sciences, Masters Advisor

Maria Zaragoza, Department Chair, Ph.D., Psychological Sciences

James L. Blank, Interim Dean, Ph.D., College of Arts and Sciences TABLE OF CONTENTS

TABLE OF CONTENTS…………………………………………………………………..iii

LIST OF TABLES..……………………………………………………………………….. v

ACKNOWLEDGMENTS..………………………………………………………………..vi

CHAPTERS

I. Introduction..………………………………………………………………… 1

Anxiety, Depression, and Chronic Pain Treatment Outcome…………… 4

The MMPI and Chronic Pain Treatment Outcomes…………………… 6

Chronic Pain and the MMPI-2………………………………………… 8

The MMPI-2-RF and Chronic Pain…………………………………… 10

Current Study………………………………………………………...… 12

II. Method…………………………………………………………………..… 15

Measures..……………………………………………………………... 16

Procedures..…………………………………………………………… 16

Analysis Plan………………………………………………………….. 17

III. Results..…………………………………………………………………. 19

One-Way Repeated Measures ANOVA..……...…………………….... 19

Correlational Analyses..…………….……………………………….... 19

Hierarchical Linear Modeling (HLM)..……………………………….. 20

IV. Discussion………………………………………………………………… 23

Clinical Implications …………………………………………………. 25

Limitations……………………………………………………………. 27

Conclusion……………………………………………………………. 28

REFERENCES …………………...……..…………………………………………... 29

iii APPENDICES

A. Pain Disability Index.……………………………………………………… 41

iv LIST OF TABLES

Table 1. ANOVA: PDI And Total Score at Each Time Point……………………………… 36

Table 2. MMPI-2-RF and Intake Pain Disability Index: Correlations……………………… 37

Table 3. Final Estimation of Fixed Effects: H-O, RC, and Somatic/

Cognitive Specific Problems Scales………………………………………………………… 38

Table 4. Final Estimation of Fixed Effects: Internalizing and Interpersonal

Specific Problems Scales…………………………………………………………………… 39

Table 5. Final Estimation of Fixed Effects: PSY-5 Scales. ………………………………… 40

v ACKNOWLEDGEMENTS

There are many people to whom I owe my gratitude for their support and wisdom in the completion of this project. First and foremost, I would like to thank my advisor, Dr. Yossef S.

Ben-Porath, for his continued support and guidance throughout this process. Through encouraging independence and your confidence in me, I have come to realize some of my capabilities and increased my own confidence as a researcher.

Furthermore, I would like to extend my appreciation to my lab members, Ryan Marek and Anthony Tarescavage, for their insights and encouragement. Their brotherly companionship and academic wisdom has been and continues to be a great source of support. I would also like to thank Emily Gathright and Kate Zelic for the many hours we shared at the library “co- ruminating”- I believe they owe us our own booth. Lastly, I would like to send a special acknowledgement to Dr. David Kalmbach, my personal academic support system and better half.

May you continue to be successful in your pursuits and provide me continued guidance through your knowledge and love.

vi INTRODUCTION

Chronic pain is problematic not only because of its aversive physical sensation, but also because of its association with occupational and social functioning, ability to live independently, and overall quality of life. Though advances in chronic pain rehabilitation have offered increasingly effective treatment options for individuals with chronic pain, a large number of individuals do not respond to current treatment protocols (Han, Geffen, Browning, Kenardy, &

Geffen, 2011). It is important to investigate characteristics associated with treatment response in order to improve outcomes. Identifying psychological factors associated with both good and poor treatment response could facilitate better-informed and appropriate patient referrals, as well as future improvements in chronic pain treatment.

By identifying characteristics associated with good response, providers can improve treatment planning for chronic pain patients. Similarly, by identifying vulnerabilities to treatment-refractory pain, clinicians may consider potentially beneficial adjunctive treatment.

Taken together, this research will allow for chronic treatment that is more individually tailored, thus improving outcomes. However, to evaluate psychological factors associated with pain treatment outcomes, it would likely prove beneficial to measure a wide range of psychological and personality characteristics. The Minnesota Multiphasic Personality Inventory-2 Restructured

Form (MMPI-2-RF; Ben-Porath & Tellegen, 2008/2011) is a commonly used and well- established assessment tool that assesses a broad range of personality and psychopathology constructs. Recent research has supported its predictive abilities of treatment outcomes in both

1 mental health and medical treatment settings, including bariatric and spine surgery candidates

(Marek, Ben-Porath, Merrell, Ashton, & Heinberg, 2014; Block, Ben-Porath, & Marek, 2013).

Along the lines of this emerging literature, the present study aims to evaluate the predictive abilities of the MMPI-2-RF in a chronic pain rehabilitation program.

The multiple aspects of chronic pain include the physical sensation component, as well as biological, psychological, and environmental factors (Turk & Wilson, 2010; Scascighini, 2008).

Because of the complex nature of chronic pain, a multidisciplinary approach to treatment is most effective. Patients engaged in Multidisciplinary Pain Management Programs (MPMP) receive therapeutic services for the psychological influences of pain, physical therapy for rehabilitation, and medication management for pain control, which requires the concerted efforts of providers in multiple disciplines (Wright & Gatchel, 2002; Flor, Fydrich, & Turk, 1992). Flor, Fydrich, and

Turk’s (1992) review supported the efficacy of MPMP, and their results indicated that patients who undergo MPMP are twice as likely to return to work as compared to patients treated with single modality treatments. Despite support for MPMP, Han, Geffen, Browning, Kenardy, and

Geffen (2011) found that many patients who received this treatment reported no change in pain

(~60%), and a small subset of patients reported an increase in pain (~5%). The large number of individuals who respond poorly to these treatments highlights the importance of identifying treatment refractory patients. Furthermore, comparisons between patients engaged in chronic pain rehabilitation versus treatment in general medical settings, the largest distinguishing factors were psychosocial problems and difficulties in functioning (Crook, Weir, & Tunks, 1989). Thus, understanding the psychological characteristics of individuals who do not respond well to treatment will allow for the improvement of current chronic pain treatment practices, as well as improvement in the individual tailoring of treatment to patients’ psychological profiles.

However, as addressed in Morely, Williams, and Eccleston’s (2013) review on evaluating psychological treatments for chronic pain, there are a number of considerations for examining the chronic pain literature. One of the primary considerations is the diversity in diagnoses in chronic pain treatment clinics. The authors argue that this heterogeneity in treatment settings has resulted in controlled trials focusing on single diagnostic groups, such as depression (see Ang,

Bair, Damush, Wu, Tu, & Kroenke, 2010 for example), which ultimately neglects a number of patients with various other psychiatric illnesses and patients who do not reach the threshold for a clinical diagnosis. They suggest that describing the psychological characteristics of chronic pain patients may be more useful in treatment planning rather than targeting symptoms of specific diagnoses.

A second notable limitation of prior research concerns inconsistency of how treatment progress is measured in these studies. As patients, clinicians, and insurance companies have differing views of the definition of successful pain treatment (Turk & Melzack, 2011), treatment outcomes vary widely between studies. As previously discussed, pain is comprised of multiple subjective and objective components. Subjective measures of improvement focus on pain severity or discomfort and perceived functional impairments in various aspects of life. However, given the naturally subjective appraisal of pain by patients, some clinicians prefer objective indications of pain improvement, such as the ability to complete physical tasks (e.g., climbing stairs, lifting light weights) and the number of visits to primary care physicians (i.e., fewer visits indicating good treatment response). This inconsistency in treatment outcome measurement makes comparisons between studies difficult (Thorn, Cross, & Walker, 2007). As such, it is important for investigations studying pain to maximize uniformity between studies to allow for comparison.

Anxiety, Depression, and Chronic Pain Treatment Outcome

The impact of comorbid psychiatric illness on chronic pain has been a primary focus of treatment research. Of chronic pain patients, 52% present with comorbid depression (Bair,

Robinson, Katon, & Kroenke, 2003). Prior research has demonstrated a bidirectional association between depression and chronic pain. Chronic pain patients often report a history of depression that precedes the onset of their pain. Additionally, chronic pain patients without a history of depression are at greater risk of developing depressive symptoms (Brown, 1990; Katon, Egan, &

Miller, 1985). An important consideration of the depression-pain relationship is that perceived functional impairment may be a critical underlying mechanism (Holzber, Robinson, Geisser, &

Gremillion, 1996). That is, pain can lead to lower levels of perceived functionality, which, in turn, increases depressed affect (Bair, Robinson, Katon, & Kroenke, 2003). Chronic pain patients are more likely to self-report physical disability, occupational, recreational, and social impairment. Importantly, these patients are less likely to return to work following pain treatment, as compared to patients without comorbid depression.

In comparison, comorbid anxiety has received less attention in the literature, despite being equally prevalent within medical populations as depression (Roy-Byrne, Davidson,

Kessler, Asmundson, Goodwin, Kubzansky, Lydiard, Massie, Katon, Laden, & Stein, 2008).

Research into comorbid of chronic pain and anxiety is beginning to make headway. Rates of anxiety in chronic pain samples have varied from 12.5-15.7% for specific phobia, 8.3-1.8% for social anxiety, and 7.3-10.7% for PTSD (Roy-Byrne, Davidson, et al., 2008). Furthermore, rates of anxiety disorders have been found to be associated with different types of chronic pain. For example, anxiety rates in fibromyalgia patients have been reported at 60% (Arnold, Hudson,

Keck, Auchenbach, Javaras, & Hess, 2006), whereas rates of anxiety in spinal pain patients have

4 been found to be between 11-27% (Von Korff, Crane, Lane, Miglioretti, Simon, Saunders, Stand,

Brandenburg, & Kessler, 2005; Dersh, Gatchel, Mayer, Polatin, & Temple, 2006). In research focusing on the association between anxiety and pain, much attention has been given to anxiety that is secondary to pain. Specifically, fear of pain is a common phenomenon in pain patients, presenting as fear of unknown causes of pain, the prospect that they cannot be helped, that others will disbelieve their pain, and that they will re-injure themselves (Gatchel, Yuan Bo Peng, Peters,

Fuchs, & Turk, 2007). This fear ultimately leads to poorer perceived functionality and the avoidance of situations that might initiate or worsen pain symptoms. This fear commonly results in a restriction of recreational, social, and occupational activities. Furthermore, Vlaeyen, Kole-

Snijders, Rotteveel, Ruesink, and Heuts (1995) found that fear of pain predicted self-reported disability, whereas pain sensation severity did not. Additionally, pain-related anxiety was shown to be a more robust predictor of disability than self-reported pain severity in a sample of head pain patients (Nash, Williams, Nicholson, & Trask, 2006). Importantly, McCracken and Gross

(1998) found that when pain-related anxiety is reduced, patients report better functioning and activity. Like the literature on depression, pain, and perceived functionality, anxiety has been shown to be the driving force behind impaired functionality, thus contributing to poorer treatment outcomes.

Though prior research has shown anxiety and depression to predict poor treatment outcomes, few investigations have utilized a nuanced and comprehensive assessment of psychiatric illness. Further, the association between personality characteristics and pain treatment outcome has not been explored. Though the lack of standard assessment batteries has been a barrier to this line of research, one self-report measure of personality and psychopathology, the

Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1943), and its

5 subsequent forms, the Minnesota Multiphasic Personality Inventory-2 (MMPI-2; Butcher,

Graham, Ben-Porath, Tellegen, Dahlstrom, & Kaemmer, 2001) and the MMPI-2-RF (Ben-Porath

& Tellegen, 2008/2011), have been routinely utilized in chronic pain settings. The MMPI and

MMPI-2 have long histories in chronic pain and treatment outcome research; however, there has been little up-to-date research with the newest revised form, the MMPI-2-RF, and chronic pain.

The MMPI and Chronic Pain Treatment Outcomes

As a self-report measure of multiple aspects of both psychopathology and personality, the

MMPI is well suited for assessing the psychological dimension of chronic pain. The original

MMPI, developed in the 1940s, was a true-false 566-item questionnaire designed to differentiate between individuals with psychiatric diagnoses from “normal” individuals. To this aim, an empirical keying approach was used during the construction of the MMPI, leading to the development of ten Clinical Scales. The ten original Clinical Scales of the MMPI assessed

Hypochondriasis, Depression, Hysteria, Psychopathic Deviant traits, Masculinity-Feminity traits,

Paranoia, Psychasthenia, Schizophrenia, Hypomania, and Social Introversion. Of these scales, the Hypochondriasis (Hs) and Hysteria (Hy) Scales focused on physical symptoms and expression of pain, respectively.

The first approaches in studying chronic pain patients using the MMPI focused on identifying total group personality characteristics using correlates of MMPI mean profiles. Two main profiles emerged through this method of examination. Gough (1946) examined MMPI profiles in soldiers with psychoneurosis (i.e., those without a physical origin to their pain) and discovered a “Neurotic Triad” profile with high elevations on Hy, Hs, and Depression (D). Other research yielded the same “Conversion V” MMPI profile with high elevations on the Hy and Hs

Scales, and low D scores, resulting in a “v-shaped” profile. One such study by Hanvik (1950)

6 found that Hs, Hy, Depression (D), Psychopathic Deviate (Pd), Psychasthenia (Ps), and

Schizophrenia (Sc) differentiated organic pain patients, that is, patients with pain derived from physical origins, from functional pain patients, or patients whose pain had no apparent physical origin. However, this method of examining chronic pain patient characteristics as a single, unified group neglected the heterogeneity of pain patients and possible homogenous subgroups within pain patient samples, such as patients with similar complaints or those who respond differently to treatment (Bradley, Prokop, Margolis, & Gentry, 1978).

To account for this heterogeneity of pain patients and in an effort to improve treatment, studies began to focus on predicting outcome in various chronic pain subgroups. Using the

MMPI, Sternbach (1974) identified four distinct subgroups of pain patients including: 1) hypochondriasis, 2) reactive depression, 3) somatization reaction, and 4) manipulative reaction groups. This finding highlighted not only the importance of psychological factors in pain patients, but also the complexity with which different patients presented. Prokop, Bradley,

Margolis, and Gentry (1980) sought to identify subgroups of differing MMPI profiles within male and female patients presenting with multiple pain complaints. Across the male and female groups, a homogeneous group emerged consisting of elevations (defined consistently as T-score

>70) on Hypochondriasis, Depression, and Hysteria. A second group in both samples was identified as having a lack of elevations on any MMPI scale. Among men, a subgroup consisting of an elevated Depression Scale and another subgroup of male patients with elevations on the majority of MMPI scales were identified. The female pain sample yielded a “Conversion V” subgroup, with elevations on Hypochondriasis and Hysteria. These findings revealed clinically significant psychopathology among pain patients, and showed that these presentations can

7 widely vary. That is, to capture the complexity of between-person psychopathology associated with chronic pain, it may necessitate a more comprehensive and nuanced assessment measure.

While research with chronic pain patients and the MMPI seemed promising, a notable shortcoming of the original MMPI subgroups is the lack of prediction of treatment outcomes. For example, examination of the association between MMPI-defined sub-groups and extra-test data and demographic information, yielded few consistencies. In one such study, McGill, Lawless,

Selby, Mooney, and McCoy (1982) were able to replicate the subgroups found by Prokrop et al.

(1980). The authors also found significant differences between the subgroups in subjective pain estimates; however, they were unable to find differences between these groups and other outcome data, including time out of bed, range of motion, and pain medication use.

Furthermore, other researchers were unable to find associations between outcome data and subgroups including Moore, Armentrout, Parker, and Kivlahan (1986) and Guck, Meilman,

Skultety, and Poloni (1988).

Chronic Pain and the MMPI-2

In the development of the MMPI-2, great effort was taken to make as few changes as possible from the original MMPI. The original aims included updating the norms and editing and updating the language of certain items. Research with the MMPI-2’s Clinical Scales in the assessment of chronic pain samples has included utilizing cluster analysis to identify subgroups of chronic pain patients (Keller & Butcher, 1991). Keller and Butcher identified four MMPI-2 profile clusters within chronic pain patients including depressed-pathological, conversion, neurotic triad, and within-normal-limits profiles. These four profile types have received criticism due to the small sample these profiles classified, in other words many pain patients MMPI-2

8 profiles did not match one of these clusters, and lack of generalizability to specific pain samples, such as low-back pain patients (Slesinger, Archer, & Duane, 2002). Other studies have found fewer subgroups, such as Nordin, Eisemann and Richter’s (2005) Hysteria and Hypomania subgroup, similar to Keller & Butcher’s Conversion group, and a Generally Elevated subgroup, with elevations on most clinical scales. Gatchel, Mayer, & Eddington (2006) identified a similar profile in spine patients, labelled the “Disability Profile”, with profiles consisting of four or more elevated scales (T-score>65). In a follow-up study examining the Neurotic Triad, Conversion-V,

Disability, and Normal Profiles and various psychosocial measures, significant mean differences were found among the four profiles with measures of depression and functional impairment, whereas pain type was not significantly related to such measures (Haggard, Stowell, Bernstein,

& Gatchel, 2008). This suggested that MMPI-2 profiles were important to understanding pathology and functional impairment at intake into treatment moreover the type of pain.

Although the item content of the MMPI-2 has not been revised, various scale construction and reconstruction has taken place. To improve discriminant validity, Restructured

Clinical Scales (RC; Tellegen, Ben-Porath, McNulty, Arbisi, Graham, & Kaemmer, 2003) were developed by extracting a demoralization factor from the ten original Clinical Scales. As a result, a Demoralization scale was created (RCd) as well as measures labeled Somatic

Complaints (RC1), Low Positive Emotions (RC2), Cynicism (RC3), Antisocial Behavior (RC4),

Ideas of Persecution (RC6), Dysfunctional Negative Emotions (RC7), Aberrant Experiences

(RC8) and Hypomanic Activation (RC9),, which measure similar, but distinct, constructs assessed by the original Clinical Scales (Graham, 2012). The development of these scales, in addition to new validity scales, content scales, supplementary scales, and the Personality

Psychopathology Five (PSY-5; Harkness, McNulty, Ben-Porath, & Graham, 2002) provided new avenues for research in chronic pain assessment with the MMPI-2.

Chronic pain research with the MMPI-2 RC Scales has capitalized on improved scale construction. One study hypothesized that fewer elevations would be identified on the RC Scales compared to the Clinical Scales due to lack of demoralization saturation and removal of item overlap between scales (McCord & Drerup, 2011). Results supported their hypothesis, with elevations of T-scores greater than 65 on RCd, RC1, and RC2, as to be expected given the relationship between pain and depression. A subgroup of pain patients with depression symptomatology and a group without depression were identified, and the RC scales were determined to more parsimoniously discriminate between these groups. MMPI-2 Clinical Scales

7 (Psychasthenia) and 8 (Schizophrenia) were elevated in the depressed group, which was explained to be due in part to demoralization saturation. Though several profiles have emerged in regard to chronic pain subgroups, there has been a gap in literature in terms of examining how the MMPI-2 can use this information to predict treatment outcomes.

The MMPI-2-RF and Chronic Pain

With the development of the RC scales and to address some of the criticisms of the

MMPI-2, including the number of items and ease of interpretability, the MMPI-2-RF was developed. The MMPI-2-RF consists of 338-items from the MMPI-2 item pool (enabling the

MMPI-2-RF to be scored from the MMPI-2) and is interpreted under a hierarchical structure. As the MMPI-2-RF is composed of MMPI-2 items, the MMPI-2-RF can be scored from the MMPI-

2. The three Higher-Order (H-O) Scales consist of Emotional/Internalizing Disorders,

Behavioral Dysfunction, and Thought Dysfunction, with the nine RC scales fall under the H-O

Scales. The Specific Problems (SP) Scales follow after the RC scales. There are five groups of

SP Scales: Somatic/Cognitive Specific Problems Scales: Malaise (MLS), Gastro-Intestinal

Complaints (GIC), Head Pain Complaints (HPC) Neurological Complaints (NUC), and

Cognitive Complaints (COG). Nine scales comprise the Internalizing Specific Problems Scales:

Suicidal/Death Ideation (SUI), Helplessness/Hopelessness (HLP), Self-Doubt (SFD), Inefficacy

(NFC), Stress/Worry (STW), Anxiety (AXY), Anger Proneness (ANP), Behavior-Restricting

Fears (BRF), and Multiple Specific Fears (MSF). Externalizing Specific Problem Scales consist of Juvenile Conduct Problems (JCP), Substance Abuse (SUB), Aggression (AGG), and

Activation (ACT), and lastly the Interpersonal Specific Problems Scales are Family Problems

(FML), Interpersonal Passivity (IPP), Social Avoidance (SAV), Shyness (SHY), and

Disaffiliativeness (DSF). There are also two Interest Scales: Aesthetic-Literary Interests (AES) and Mechanical-Physical Interests (MEC) and the revised PSY-5 scales: Aggressiveness-Revised

(AGG-r), Pyschoticism-Revised (PSYC-r), Disconstraint-Revised (DISC-r), Negative

Emotionality/Neuroticism- Revised (NEGE-r), and Introversion/Low Positive Emotionality-

Revised (INTR-r).

Although the MMPI-2-RF is presently being utilized with chronic pain patient populations, no research, to the knowledge of this author, has been published on predicting treatment outcomes or describing pain patients using the MMPI-2-RF. However, the MMPI-2-

RF has been shown to be useful in medical settings, including identification of psychological characteristics in bariatric surgery patients (Marek, Ben-Porath, Windover, Tarescavage, Merell,

Ashton, Lavery, & Heinberg, 2013) and prediction of treatment outcome in bariatric surgery candidates one and three-months post-surgery (Marek, Ben-Porath, Merrell, Ashton, &

Heinberg, 2014). Furthermore, the MMPI-2-RF has been utilized in the development of

11 algorithms for predicting treatment outcomes based on pre-surgical screening in spine surgery candidates (Block, Ohnmeiss, Ben-Porath, & Burchett, 2011). The MMPI-2-RF has also demonstrated utility in identifying psychological risk factors for poor treatment outcome in spine surgery candidates (Block, Ben-Porath, & Marek, 2013). Given the psychometric and time demand benefits of the MMPI-2-RF over prior versions, and the recent literature in other medical settings, it is likely that the MMPI-2-RF will be useful in the prediction of rehabilitation treatment outcomes in chronic pain populations.

Current Study

The MMPI-2 has proven to be promising in the area of chronic pain subgroup identification as a broad measure of personality and psychopathology; however, the literature base involving the MMPI-2-RF is lacking. Additionally, prior research has largely been cross- sectional, and there is a gap in the literature concerning predicting treatment outcomes in chronic pain patients using the MMPI-2-RF. The complex psychological factors related to chronic pain necessitate a comprehensive assessment of psychiatric symptoms and personality factors. As discussed, depression and anxiety have been identified as barriers to response, despite the psychotherapy component of treatment. Thus, it is important to identify specific targets for intervention in chronic pain patients that will optimize treatment outcome as defined by perceived functional improvement. Examining the MMPI-2-RF RC and Specific Problems

Scales in the prediction of perceived disability in areas of function is one way to identify targets for intervention. The MMPI-2-RF scales examined in this study are based on past literature on comorbid diagnoses, depression and anxiety, and their relation to treatment outcomes in chronic pain patients. For this reason, the MMPI-2-RF scales associated with depression, anxiety, and

12 somatic complaints were utilized including the Higher-Order (H-O) Emotional/Internalizing

Dysfunction (EID) EID scale, Restructured Clinical (RC) Scales Demoralization (RCd), Somatic

Complaints (RC1), Low Positive Emotions (RC2), and Dysfunctional Negative Emotions (RC7),

Somatic/Cognitive Specific Problem Scales Malaise (MLS), Gastrointestinal Complaints (GIC),

Head Pain Complaints (HPC), Neurological Complaints (NUC), and Cognitive Complaints

(COG), and Internalizing Specific Problem Scales Suicidal/Death Ideation (SUI),

Helplessness/Hopelessness (HLP), Self-Doubt (SFD), Inefficacy (NFC), Stress/Worry (STW),

Anxiety (AXY), Anger Proneness (ANP), Behavior-Restricting Fears (BRF), and Multiple

Specific Fears (MSF). Additionally, Specific Problems Scales assessing Interpersonal Problems,

Family Problems (FML), Interpersonal Passivity (IPP), Social Avoidance (SAV), Shyness

(SHY), and Disaffiliativeness (DSF), were examined given the possible social impairment due to chronic pain. Finally, the PSY-5 Scales, Aggressiveness (AGG-r), Psychoticism (PSYC-r),

Disconstraint (DISC-r), Negative Emotionality/Neuroticism (NEGE-r), and Introversion/Low

Positive Emotionality (INTR-r), were also examined to determine if there were possible underlying personality constructs associated with poor treatment outcome.

This study aimed to further the literature on chronic pain treatment outcome by examining prediction of self-reported disability due to chronic pain with the MMPI-2-RF. The dependent measure is the Pain Disability Index (PDI; Pollard, 1984), which measures functional impairment in several life domains. Research suggests that pain patients perceive chronic pain treatment to be more effective at improving function compared to pain sensation (Aronoff,

Evans, & Enders, 1985). As discussed, pain is subjective, and thus perceived lack of improvement is detrimental to treatment outcome. To examine the predictive abilities of the

MMPI-2-RF, descriptive information was first reported concerning the PDI at each study time

13 point, as are correlations between the PDI and the MMPI-2-RF. Longitudinal data analysis was used in order to identify MMPI-2-RF scales that were predictive of change in pain disability scores versus no change or worsening from intake into treatment, discharge, 6 months post- discharge, and 12 months post-discharge. Pain disability was expected to significantly decrease across the four time points (hypothesis 1). Higher levels of psychopathology were expected to be predictive of overall greater pain disability throughout treatment (hypothesis 2). Lastly, greater psychopathology was expected to be predictive of poorer treatment outcome. Specifically, reported disability was expected to remain unimproved or worsen over time for those with higher elevations on the MMPI-2-RF, while those low on the identified MMPI-2-RF scales at intake were expected to improve over time (hypothesis 3).

METHOD

This study used an archival data set collected from a Midwestern chronic pain rehabilitation center. Of the 1,494 total patients treated during the study period, 1,156 had valid

MMPI-2-RF protocols based on the MMPI-2-RF technical manual recommendations (CNS>18,

VRIN>80, TRIN>80, Fp-r>100, and F-r>120; Tellegen & Ben-Porath, 2008). For the purposes of obtaining complete data at our primary points of interest, patients were selected for this study based on a PDI Total score>0 at intake and 12 months post-discharge; however, some of these patients had missing data at discharge and 6 months post-discharge. A sample size of 232 chronic pain patients with valid MMPI-2-RF protocols and PDI data at intake and 12 months post-discharge remained. An independent t-test indicated that the valid sample (N=1156,

M=1.67, SD=1.60) had scored significantly higher compared to the final sample (N=232, M=

1.44, SD= 1.45) on Gastrointestinal Complaints (t(353.86)= -2.14, p<.05). The valid sample also scored significantly higher compared to the final sample (N=232, M=4.83, SD=3.18) on

Disconstraint (t(371.20)= -3.30, p<.01).

The sample was primarily female (64%) with a mean age of 48.70 (SD=12.83). The average amount of years of education was 13.17 (SD=4.12). According to the information available, 95.7% of the sample completed the rehabilitation program, with an average of 25 days between intake and discharge. In terms of years of pain duration from onset to treatment discharge, the sample had a mean of 11.16 (SD=10.09) years. Of those seeking treatment for a primary pain complaint, 43% had a primary complaint of lower back pain, 17% fibromyalgia, and 7% joint pain.

Measures

Minnesota Multiphasic Personality Inventory-2 Restructured Form (MMPI-2-RF; Ben-

Porath & Tellegen, 2011). The MMPI-2-RF is a 338-item true-false, self-report measure of personality and psychopathology. The scales of this measure are organized in a hierarchical fashion, with nine validity scales, three Higher-Order Scales, nine Restructured-Clinical Scales, and the Specific Problems Scales including the five Somatic/Cognitive Scales, nine Internalizing

Scales, four Externalizing Scales, five Interpersonal Scales, and two Interest Scales. The MMPI-

2-RF also contains the Restructured Psychopathology Five Scales.

Pain Disability Index (PDI; Pollard, 1984). The PDI is a 7-item self-report questionnaire assessing the degree to which pain interferes with seven life categories. The seven categories assessed are Family/Home Responsibility, Recreation, Social Activity, Occupation, Sexual

Behavior, Self-Care, and Life Support (i.e., eating, breathing, sleeping). Patients are asked to rate their perceived level of disability on a scale of 0 (no disability) to 10 (worst disability). A total disability due to pain score is derived, ranging from 0-70. Good internal consistency (α=.82) for the total score was found in the current sample.

Procedures

Patients entered into an intensive pain rehabilitation clinic in the Midwest, which consisted of 3-4 weeks of full-day intervention. The clinic exercised a multidisciplinary approach to treatment focusing on physical rehabilitation, individual and group cognitive behavioral therapy psychotherapy addressing emotional distress, and mindfulness and relaxation techniques, and monitored cessation of addictive substances, including opioids and benzodiazepines.

Patients underwent a battery of pre-treatment assessment measures, including the MMPI-2-RF

16 and the PDI. At discharge, patients were given a battery of outcome measures, including the PDI.

Those who participated in the research were mailed follow-up measures at 6 months post- discharge and 12 months post-discharge.

Analysis Plan

Descriptive analyses were first run to determine the nature of the data and justify the use of more complicated analyses. A One-Way Repeated Measures ANOVA was conducted examining the change in pain disability across the four time points. This analysis was necessary in order to examine the effect of treatment on pain disability. As stated in our hypothesis, it was predicted that pain disability would decrease throughout treatment and follow-up. Secondly, bi- variate correlational analyses were then examined with the MMPI-2-RF scales and the total PDI scores at intake, discharge, and 6- and 12- month discharge. It was expected that positive associations would be found between the identified scales and pain disability.

Due to the time-nested structure of the data, we employed hierarchical linear modeling in testing out substantive hypothesis. In doing so, the average time in days was determined between time points. Time between discharge and intake was 25 days, 205 days passed between 6 months post-discharge and discharge, and 191 days between 12 months post-discharge and 6 months post-discharge. Time was then centered at 12 months post-discharge, so that the Time variable indicating 12 months post-discharge equaled zero, 6 months post-discharge equaled -191, discharge equaled -396, and intake equaled -423. This was necessary to ensure that the central hypothesis of change in PDI total scores from intake would be examined.

An unconditional, null model regressing PDI onto an intercept without any predictors was run to describe the amount of both inter- and intra-individual variability found in the pain

17 disability outcome. Percent of inter- and intra-individual variance was calculated by dividing the

PDI standard deviation at the intercept by total variance (i.e., PDI standard deviation at the intercept plus error stand deviation). Next, MMPI-2-RF predictors were added to the model by scale-set in six separate models. The Higher-Order Scale Emotional Dysfunction (EID) was entered into a model (See Example 1). Next, the RC Scales RCd, RC1, RC2, and RC7 were entered into a model in a similar fashion. The Somatic/Cognitive Scales MLS, GIC, HPC, NUC, and COG were then entered into a model. Following the Somatic/Cognitive Scales, the

Internalizing Specific Problems Scales SUI, HLP, SFD, NFC, STW, AXY, ANP, and BRF were analyzed in a separate model. The interpersonal Scales FLM, IPP, SAV, SHY, and DSF were next into a model. Lastly, the PSY-5 Scales AGG-r, PSYC-r, DISC-r, NEGE-r, and INTR-r were entered into a model.

Example 1.

PDIit= β0i+β1(EIDi)+β2(TIMEit)+ β3(EID*TIMEit)+ζ0i + εit

PDIit Expected total PDI score for individual i at time t.

β0i The intercept for person i.

β1(EIDi) Expected difference in pain disability for greater

emotional and internalizing dysfunction at intake.

β2(TIMEit) Expected difference in pain disability over time for person i at time

β3(EID*TIMEit) The effect of the interaction between EID and Time

for person i at time t.

ζ0i Leve-1 stochastic parts.

εit Level-2 stochastic parts.

RESULTS

One-Way Repeated Measures ANOVA

A one-way repeated Measures ANOVA indicated significant mean differences in total

PDI score at the four time points (F(2.11, 194.30)=106.80, p< .01, η2=.54; see Table 1). Post-hoc analyses were performed using Tukey’s HSD test to determine which total PDI scores differed.

Analyses revealed that patients experienced a significant reduction in symptoms from intake to discharge (see Table 1). However, after discharge, patients reported increased disability at 6- and 12- month follow-ups, at which time PDI level remained stable. Notably, disability ratings at the follow-up assessments was significantly lower than the mean baseline PDI score, indicating that the treatment overall exhibited stable therapeutic benefit.

Correlational Analyses

Zero-order correlations were calculated between the selected MMPI-2-RF scales and the

PDI items and total score at each time point (see Table 2). At intake, significant associations between the PDI total score and the MMPI-2-RF were found for EID, RCd, RC1, RC2, RC7,

MLS, NUC, COG, SUI, HLP, SFD, STW, AXY, BRF, SAV, DSF, NEGE-r, and INTR-r. At treatment discharge, significant associations were found between the PDI total score and the

MMPI-2-RF Scales EID, RC2, COG, HLP, BRF, and INTR-r. At six months post-discharge, the total PDI score was significantly associated with RC1, RC2, HPC, NUC, COG, SUI, HLP, AXY,

BRF, SAV, and DSF. And lastly, at 12 months post-discharge, the PDI total score was significantly associated with RC1, NUC, COG, HLP, NFC, AXY, and BRF.

Hierarchical Linear Modeling (HLM)

An unconditional, or random-coefficient, model was tested first in order to determine whether there was significant inter- and intra-individual variability in PDI total scores. Results indicated that approximately 31% of PDI variance throughout treatment was accounted for by difference between individuals. This supports the importance of examining individual differences as predictors of treatment outcomes. Additionally, 69% of the variance in PDI was due to changes within individuals, which supports the use of HLM analysis of these time-nested data.

Pain Disability Total score was next regressed onto the Higher-Order scale EID (see

Table 3). Analysis revealed a significant main effect such that individuals who indicated greater emotional distress experienced greater disability overall than individuals low on EID. However, contrary expectations, time did not moderate the association between EID and self-reported disability. That is, though patients high on EID did report experiencing greater pain disability throughout and following treatment, the rate at which their symptoms decreased did not differ as a function of EID scores.

The predictive abilities of the RC Scales RCd, RC1, and RC2 were then examined.

Specifically, the scales were examined by regressing PDI onto RCd, RC1, RC2 and RC7 (see

Table 3). Analyses revealed that greater baseline Somatic Complaints (RC1) and Low Positive

Emotions (RC2) were predictive of greater PDI scores throughout treatment. Notably, as these predictors were entered simultaneously into the model, these associations are unique, rather than due to shared variance. However, again contrary to the hypotheses, individuals with high

Somatic Complaints or Low Positive Emotions did not exhibit a rate of change in PDI scores that differed from individuals low on either of these two RC scales. Demoralization (RCd) and

Dysfunctional Negative Emotions (RC7) were not found to be predictive of PDI scores, nor did they exhibit a rate of change in PDI scores that differed from individuals low on these scales.

Next, the predictive abilities of the Somatic/Cognitive Scales were examined. PDI was regressed onto MLS, GIC, HPC, NUC, and COG (see Table 3). Analyses revealed that greater baseline Neurological Complaints (NUC) and Cognitive Complaints (COG) were predictive of greater PDI scores throughout and following treatment. Again, because these predictors were entered simultaneously into the model, these associations are unique and not due to shared variance. However, individuals with high GIC, NUC, or COG scores did not exhibit a rate of change in PDI scores that different from individuals low on these scales. Malaise (MLS) and

Head Pain Complaints (HPC) were not found to be predictive of PDI scores, nor did they exhibit a rate of change in PDI scores that differed from individuals low on these scales.

Next the predictive abilities of the MMPI-2-RF’s Internalizing Specific Problems Scales were examined. The total PDI score was regressed onto SUI, HLP, SFD, NFC, STW, AXY,

ANP, BRF, and MSF (see Table 4). Of these scales, analyses revealed that greater

Helplessness/Hopelessness (HLP) was predictive of greater PDI score throughout treatment.

Additionally, trends existed with higher Anxiety (AXY) and Behavior-Restricting Fears (BRF) indicating possible predictive abilities with greater pain disability over treatment. These nine predictors were entered into the model simultaneously, and thus these findings are unique and

Stress/Worry (STW), Anger-Proneness (ANP), and Multiple Specific Fears (MSF) were not found to be predictive of PDI scores. However, individuals high on Multiple Specific Fears did exhibit a rate of change in PDI scores that differed from individuals low on this scale, though the other Internalizing Scales did not. Because Multiple Specific Fears did not have a significant

main effect, a separate model with only MSF was run to reduce the concern for a spurious effect.

This analysis confirmed a significant interactive effect (β=.004, t=2.96, p<.05).

The predictive abilities of the Interpersonal Scales were then examined. Again, the PDI was regressed onto FLM, IPP, SAV, SHY, and DSF (see Table 4). Results indicated that

Disaffiliativeness (DSF) was predictive of greater disability over treatment. However, Family

Problems (FML), Interpersonal Passivity (IPP), Social Avoidance (SAV) and Shyness (SHY) were not predictive of PDI scores. Furthermore, individuals high on these scales did not exhibit a rate of change in PDI scores that differed from those who scored low on these scales.

Lastly, the predictive abilities of the PSY-5 scales were examined. In other words, PDI was regressed onto AGG-r, PSY-r, DIS-r, NEGE-r, and INTR-r (see Table 5). Of these scales, high Psychoticism (PSY-r) and Introversion (INTR-r) was predictive of higher pain disability over treatment. However, Aggression (AGG-r), Disconstraint (DISC-r), and Negative

Emotionality (NEGE-r) were not predictive of pain disability over time. Additionally, individuals scoring higher on the PSY-5 scales did not exhibit a rate of change in PDI scores that differed from those who scored low on the PSY-5 Scales.

DISCUSSION

This study examined prediction of pre- and post-treatment self-reported disability in chronic pain patients using the MMPI-2-RF. Specifically, this study used the MMPI-2-RF scales assessing baseline somatic symptoms, internalizing symptoms, and interpersonal problems to differentiate between good and poor treatment responders. We predicted that scores on a measure of self-reported pain-disability would decrease over time (hypothesis 1). We also predicted that those with higher levels of psychopathology would also experience greater pain disability throughout treatment and at follow-up compared to those with lower levels of psychopathology

(hypothesis 2). Lastly, it was predicted that the rate of improvement in pain disability would differ based on levels of psychopathology. Namely, it was expected that those with greater psychopathology would experience slower or no improvement in disability due to chronic pain compared to those with lower levels of psychopathology (hypothesis 3).

This study revealed several significant findings that have important implications for chronic pain treatment. First, the results indicated that individuals entering into treatment with higher levels of emotional and internalizing dysfunction experienced higher levels of pain disability throughout treatment and follow-up compared to those with lower levels of emotional and internalizing dysfunction. Unsurprisingly, those with higher levels of somatic complaints at intake, as assessed by RC1, experienced greater disability due to chronic pain across treatment.

That is, patients who reported tendencies to somaticize also perceived greater functional impairment due to their chronic pain. Consistent with prior research on depression and chronic

pain (Holzber, Robinson, Geisser, & Gremillion, 1996; Vlaeyen, Kole-Snijders, Rotteveel,

Ruesink, &Heuts, 1995), patients who endorsed greater dysfunctional negative emotions or lower positive emotions were at risk for experiencing greater functional impairment secondary to their chronic pain. Importantly, positive affect and negative affect conferred independent vulnerabilities to greater pain disability, which is consistent with prior research indicating that these two constructs are uniquely related to a number of various health outcomes, including cardiovascular, immunological, and gastrointestinal health (Pressman & Cohen, 2005; Steptoe,

Wardle, & Marmot, 2005).

In examining more specific problems, patients with higher neurological complaints or cognitive complaints upon entering treatment experienced higher levels of disability due to chronic pain throughout treatment and follow-up. This is consistent with the findings regarding

RC1. Among specific internalizing problems, the more hopeless or helpless an individual felt was predictive of greater pain disability throughout and following treatment. Notably, pain patients who presented with greater anxiety may have experienced greater comorbid pain disability, as indicated by a statistical trend. Though patients with greater multiple specific fears did not report greater pain disability overall, they did respond more poorly to treatment. That is, the rate at which these patients’ pain disability symptoms was alleviated, was lower compared to patients with lower levels of multiple specific fears. Examination of specific interpersonal problems revealed that individuals with higher levels of disaffiliativeness, or preference to be alone and disliking others, experienced higher levels of disability across treatment and follow-up.

Lastly, regarding personality qualities, patients who present with greater psychoticism and patients who are more introverted were at greater risk of higher levels of pain disability. Notably, contrary to our predictions, no interactive effects were found. That is, though elevations on the

scales just discussed were predictive of poorer treatment outcome, we did not find evidence that the rate of decrease in pain disability symptoms varied as a function of the severity of psychological and personality factors. Rather, individuals with greater psychopathology presented with greater disability at intake, which continued throughout and following treatment.

Clinical Implications

The present study’s findings support the utility of the MMPI-2-RF in predicting treatment outcome for chronic pain patients. Overall, patients high on internalizing emotional distress, specifically as indicated by somatic complaints, low positive emotions, and high negative emotionality, are at greater risk for poorer treatment outcome. Given that individuals with these psychological factors report more disability throughout treatment and post-treatment follow-up compared to individuals with lower levels of these psychological factors,, these findings suggest that the psychotherapy component in chronic pain treatment may be insufficient for individuals with higher levels pathology. Thus offering adjunctive therapy to address any psychological disorder or symptoms as indicated by elevated MMPI-2-RF scales and other measures may improve perceived pain-related functional impairment.

Further, careful examination of scale elevations offers nuanced information on targets for intervention. For example, patients presenting with somatic complaints are, unsurprisingly, preoccupied with their physical health and are more prone to developing somatic symptoms in response to stress. These individuals are also likely to reject the notion that their physical complaints are of psychological origins. Thus, adjunctive therapy may focus on this aspect of the relation between somatization, psychological influences of the pain experiences, and pain disability. Furthermore, for the individuals with greater neurological (e.g., dizziness, coordination difficulties) and cognitive (e.g., memory difficulties, low frustration tolerance, poor

concentration) symptoms these symptoms may be manifestations of stress. Additionally, these findings could also be indicative of underlying cognitive difficulties that may influence both their somatic complaints, as well as their pain experience and perceived disability. Elevations on these scales, as well as clinical interview, may indicate appropriateness and need for neuropsychological evaluation to better individually tailor treatment.

Similarly, the findings with the more narrowly focused scales on the MMPI-2-RF provide more nuanced information on targets for intervention and understanding patients who score high on these scales. For example, individuals reporting higher levels of somatic complaints are, unsurprisingly, preoccupied with their physical health and are more prone to developing somatic symptoms in response to stress. These individuals are also likely to reject the notion that their physical complaints are of psychological origins. Likewise, treatment outcomes might be better improved by focusing on an individual’s pessimism and lack of energy, as indicated by significant findings.

In comparison, patients who present with greater comorbid anhedonia and greater negative emotionality may better respond to therapy that focuses on these depressive symptoms.

This is supported by our finding that individuals with greater pain disability also experience greater hopelessness and helplessness. Given that these feelings are related to poor motivation for change (Salomons, Moayedi, Weissman-Fogel, Goldberg, Freeman, Tenenbaum, & Davis,

2012), it is important to tend to these feelings as they may likely prove to be a barrier to pain treatment.

As introversion and disaffiliativeness were identified as risk factors of greater pain disability, individuals high on these scales may benefit from interpersonal therapy that focuses on developing relationships to increase social support, as well as focusing on maintaining strong

therapeutic alliances between the patients and providers during treatment. Finally, as pain disability was related to greater thought dysfunction, these symptoms should receive further evaluation for pain patients who endorse high levels of psychoticism.

Limitations

Though the present study adds to the current literature, its findings must be interpreted in the context of a number of limitations. The outcome measure only assessed self-reported disability. Given the literature on functional outcomes and depression and anxiety in chronic pain patients, this measure was deemed an important aspect of treatment outcome. However, other aspects of the pain experience, such as pan severity and objective functional outcomes, were not assessed. It may be appropriate for future investigations to examine how psychopathology not only predicts pain disability across treatment, but other aspects of the pain experience as well.

Doing so would provide a more comprehensive and nuanced depiction of how these psychological factors influence pain treatment outcome.

Finally, in effort to examine the independent effects of predictors, each set’s scales were entered into the HLM models simultaneously. Thus, some models included a large number of predictor variables (e.g., internalizing specific problem scales include nine scales, resulting in 19 predictors [i.e., time + 9 main effect predictors + 9 interactive effects predictors]). Given the large number of collinear predictors, these analyses may have lacked sufficient statistical power to detect some effects. However, we determined that the conservative approach of running the models by scale set was needed to reduce the likelihood of type I errors stemming from analyzing each scale in an individual model.

Conclusion

The present study identified risk factors for greater pain disability as assessed with the

MMPI-2-RF. Preoccupation with physical symptoms, low positive emotions, focus on neurological complaints, feelings of helpless and hopeless, the experience of disaffiliativeness, and introversion represent important targets for intervention that may improve chronic pain treatment outcome. Given the high rates of non-response or poor response in this setting, more individualized treatment is necessary to improve outcome in this population.

REFERENCES

Arnold, L.M., Hudson, J.I., Keck, P.E., Auchenbach, M.B., Javaras, K.N., & Hess, E.V. (2006).

Comorbidity of fibromyalgia and psychiatric disorders. Journal of Clinical Psychiatry,

67(8), 1219-1225.

Aronoff, G.M., Evans, W.O., & Enders, P.L. (1985). A review of follow-up studies of

multidisciplinary pain units. Pain, 16, 1-11.

Ang, D. C., Bair, M. J., Damush, T. M., Wu, J., Tu, W., & Kroenke, K. (2010). Predictors of

pain outcomes in patients with chronic musculoskeletal pain co-morbid with depression:

Results from a randomized controlled trial. Pain Medicine, 11(4), 482-491.

Bair, M.J., Robinson, R.L., Katon, W., & Kroenke, K. (2003). Depression and pain comorbidity:

A literature review. Archives of Internal Medicine, 163(20), 2433-2445.

Block, A.R., Ben-Porath, Y.S., & Marek, R.J. (2013). Psychological risk factors for poor

outcome of spine surgery and spinal cord stimulator implant: A review of the literature

and their assessment with the MMPI-2-RF. The Clinical Neuropsychologist, 27(1), 81-

107.

Block, A.R., Ohnmeiss, D., Ben-Porath, Y.S., & Burchett, D. (2011). Presurgical psychological

screening: A new algorithm, including the MMPI-2-RF, for predicting surgery results.

The Spine Journal, 11(10),

Bradley, L.A., Prokop, C.L., Margolis, R., & Gentry, W.D. (1978). Multivariate analyses of the

MMPI profiles of low back pain patients. Journal of Behavioral Medicine, 1,

253-272.

Brown, G.K. A causal analysis of chronic pain and depression. Journal of Abnormal

Psychology, 99(2), 127-137.

Butcher, J.M., Graham, J.R., Ben-Porath, Y.S., Tellegen, A., Dahlstrom, W.G., & Kammer, B.

(2001). MMPI-2: Manual for administration and scoring (rev. ed.). Minneapolis:

University of Minnesota Press.

Crook, J., Weir, R., & Tunks, E. (1989). An epidemiological follow-up survey of persistent

pain sufferers in a group family practice and specialty pain clinic. Pain, 36(1), 49-61.

Dersh, J., Gatchel, R.J., Mayer, T., Polatin, P., & Temple, O.W. (2006). Prevalence of

Psychiatric Disorders in Patients with Chronic Disabling Occupational Spinal

Disorders. Spine, 31, 1156-1162.

Flor, H., Fydrich, T., & Turk, D.C. (1992). Efficacy of multidisciplinary pain treatment centers:

A meta-analytic review. Pain, 49: 221-230.

Gatchel, R.J., Mayer, T.G., & Eddington, A. (2006). MMPI disability profile: The least known,

most useful screen for psychopathology in chronic occupational spinal disorders. Spine,

31(25), 2973-2978.

Gatchel, R.J., Peng, Y.B., Peters, M.L., Fuchs, P.N., & Turk, D.C.. (2007). The biopsychosocial

approach to chronic pain: scientific advances and future directions. Psychological

Bulletin 133, 581.

Gough, H. G. (1946). Diagnostic patterns on the Minnesota Multiphasic Personality Inventory.

Journal of Clinical Psychology, 223-37.

Guck, T. P., Meilman, P. W., Skultety, F. M., & Poloni, L. D. (1988). Pain-patient Minnesota

Multiphasic Personality Inventory (MMPI) subgroups: evaluation of long-term treatment

outcome. Journal of Behavioral Medicine, 11(2), 159-169.

Graham, J.R. (2012). MMPI-2: Assessing personality and psychopathology (5th ed.). Oxford,

New York: Oxford University Press.

Haggard, R.A., Stowell, A.W., Bernstein, D., & Gatchel, R.J. (2008). Relationships between the

MMPI-2 and psychosocial measures in a heterogeneous population. Rehabilitation

Psychology, 53(4), 471-478.

Han, X., Geffen, S., Browning, M., Kenardy, J., & Geffen, G. (2011). Outcome evaluation of a

multidisciplinary pain management programme comparing group with individual change

measures. Clinical Psychologist, 15, 133-138.

Hanvik, L.J. (1950). MMPI profiles in patients with low-back pain. Journal of Consulting

Psychology, 15, 350-353.

Harkness, A. R., McNulty, J. L., Ben-Porath, Y. S., & Graham, J. R. (2002). MMPI-2

Personality-Psychopathology Five (PSY-5) Scales: Gaining an overview for case

conceptualization and treatment planning. Minneapolis, MN: University of Minnesota

Press.

Hathaway, S.R., & McKinley, J.C. (1943). Manual for the Minnesota Multiphasic Personality

Inventory. New York: Psychological Corporation.

Holzberg, A.D., Robinson, M.E., Geisser, M.E., & Gremillion, H.A. (1996). The effects of

depression and chronic pain on psychosocial and physical functioning. The Clinical

Journal of Pain, 12, 118-125.

Katon, W., Egan, K., & Miller, D. (1985). Chronic pain: Lifetime psychiatric diagnoses and

family history. American Journal of Psychiatry, 774. 1156-1160.

Keller, L.S., & Butcher, J.N. (1991). Assessment of chronic pain patients with the MMPI-2.

Minneapolis, Minnesota: University of Minnesota Press.

Marek, R.J., Ben-Porath, Y.S, Merrell, J., Ashton, K., & Heinberg, L.G. (2014). Predicting one

and three month postoperative somatic concerns, psychological distress, and maladaptive

eating behaviors in bariatric surgery candidates with the Minnesota Multiphasic

Personality inventory-2 Restructured Form (MMPO-2-RF). Obesity Surgery, 24, 631-

639.

Marek, R.J., Ben-Porath, Y.S., Windover, A., Tarescavage, A.M., Merrell, J., Ashton, K.,

Lavery, M. & Heinberg, L.J. (2013). Assessing psychosocial functioning of bariatric

candidates with the Minnesota Multiphasic Personality Inventory-2 Restructured Form

(MMPI-2-RF). Obesity Surgery, 23(11), 1864-1873.

McCord, D.M., & Drerup, L.C., (2011). Relative practical utility of the Minnesota Multiphasic

Personality Inventory-2 Restructured Clinical Scales versus the Clinical Scales in a

chronic pain patient sample. Journal of clinical and experimental neuropsychology,

33(1): 140-146.

McCracken, L.M., & Gross, R.T. (1998). The role of pain-related anxiety reduction in the

outcome of multidisciplinary treatment for chronic low back pain: preliminary results.

Journal of Occupational Rehabilitation, 8, 179–89.

McGill, J. C., Lawlis, G. F., Selby, D., Mooney, V., & McCoy, C. E. (1983). The relationship of

Minnesota Multiphasic Personality Inventory (MMPI) profile clusters to pain behaviors.

Journal of Behavioral Medicine, 6(1), 77-92.

Moore, J. E., Armentrout, D. P., Parker, J. C., & Kivlahan, D. R. (1986). Empirically derived

pain-patient MMPI subgroups: Prediction of treatment outcome. Journal of Behavioral

Medicine, 9(1), 51-63

Morely, S., Williams, A., & Eccleston, C. (2013). Examining the evidence about psychological

treatments for chronic pain: Time for a paradigm shift? Pain, 154(10), 1929-1931.

Nash, J.M., Williams, D.M., Nicholson, R.A., & Trask, P.C. (2006). The contribution of pain-

related anxiety to disability in headache. Journal of Behavioral Medicine, 29, 61-67.

Nordin, H., Eisemann, M., & Richter, J. (2005). MMPI-2 in chronic pain patients: An

evaluation of the Swedish version. Scandinavian Journal of Psychology, 46, 209-2016.

Pollard, C.A. (1984). Preliminary validity study of the pain disability index. Perceptual and

Motor Skills, 59, 974.

Pressman SD, Cohen S. (2005). Does positive affect influence health? Psychological Bulletin,

131(6), 925-971.

Prokop, C.K., Bradley, L.A., Margolis, R., & Gentry, W. (1980). Multivariate analysis of the

MMPI profiles of patients with multiple pain complaints. Journal of Personality

Assessment, 44, 246-252.

Roy-Byrne, Davidson, Kessler, Asmundson, Goodwin, Kubzansky, Lydiard, Massie, Katon,

Laden, & Stein. (2008). Anxiety disorders and comorbid medical illness. Focus, 6(4),

467-485.

Salomons, T.V., Moayedi, M., Weissman-Fogel, I., Goldberg, M.B., Freeman, B.V.,

Tenenbaum, H.C., & Davis, K.D. (2012). Perceived helplessness is associated with

individual differences in the central motor output system. European Journal Of

Neuroscience, 35(9), 1481-1487.

Scascighini, L., Toma, V., Dober-Spielmann, S., & Sprott, H. (2008). Multidisciplinary

treatment for chronic pain: A systemic review of interventions and outcomes.

Rheumatology 47(5); 670-678.

Slesinger, D., Archer, R.P., & Duane, W. (2002). MMPI-2 characteristics in a chronic pain

population. Assessment, 9(4), 406-414.

Steptoe, A., Wardle, J., & Marmot, M. (2005). Positive affect and health-related neuroendocrine,

cardiovascular, and inflammatory processes. Proceedings of the National Academy of

Sciences of the United States of America, 102(18), 6508-6512.

Sternbach, R. (1973). Traits of pain patients: The low back “loser”. Psychosomatics, 14: 226-

229.

Tellegen, A., & Ben-Porath, Y. S. (2008). MMPI-2-RF (Minnesota Multiphasic Personality

Inventory-2 Restructured Form): Technical manual. Minneapolis: University of

Minnesota Press.

Tellegen, A., Ben-Porath, Y. S., McNulty, J. L., Arbisi, P. A., Graham, J. R., & Kaemmer, B.

(2003). MMPI-2 Restructured Clinical (RC) Scales: Development, validation, and

interpretation. Minneapolis: University of Minnesota Press.

Thorn, B.E., Cross, T.H., & Walker, B. (2007). Meta-analyses and systematic reviews of

psychological treatment for chronic pain: Relevance to an evidence-based practice.

Health Psychology, 26(1), 10-12.

Turk, D.C., & Melzack, R. (2011). Handbook of pain assessment (3rd ed.). New York, NY:

The Guilford Press.

Turk, D.C., & Rudy, T.E. (1990). Neglected factors in chronic pain treatment outcome studies-

referral patterns, failure to enter treatment and attrition. Pain, 43, 7-25.

Turk, D.C., Rudy, T.E., & Sorkin, B.A. (1993). Neglected topics in chronic pain treatment

outcome studies: determination of success. Pain, 53, 3-16.

Turk, D.C., & Wilson, H.D. (2010). Fear of pain as a prognostic factor in chronic pain:

Conceptual models, assessment, and treatment implications. Current Pain and Headache

Reports 14: 88-95.

Vlaeyen, J.W., Kole-Snijders, A.M., Rotteveel, A.M., Ruesink, R., & Heuts, P.H. (1995). The

role of fear of movement/(re)injury in pain disability. Journal of Occupational

Rehabilitation, 5(4), 235-252.

Von Korff, Crane, Lane, Miglioretti, Simon, Saunders, Stang, Brandenburg, & Kessler (2005).

Chronic spinal pain and physical-mental comorbidity in the United States: Results from

the national comorbidity survey replication. Pain, 113(3). 331-339.

Wright, A., & Gatchel, R.J. (2002). Occupational rehabilitation: Interdisciplinary

management of work-related musculoskeletal pain and disability. In D.C. Turk &

R.J. Gatchel, Psychological Approaches to Pain Management: A Practitioner’s

Handbook (2nd Edition). New York: Guilford.

Table 1. ANOVA: PDI Items and Total Score at Each Time Point. T1 T2 T3 T4 N M SD M SD M SD M SD F Statistic η2 F(2.01, 193.13)=64.51, *Family/Home Life 97 6.48T2, T3, T4 2.17 2.68T1, T3, T4 1.99 3.60T1, T2 2.91 3.58T1, T2 2.54 .40 p< .01 F(2.61, 250.57)=71.84, Recreation 97 7.21T2, T3, T4 2.13 2.81T1, T3, T4 1.99 4.05T1, T2 3.12 3.87T1, T2 2.85 .43 p< .01 F(2.48, 238.28)=72.95, Socialization 97 6.39T2, T3, T4 2.42 2.02T1, T3, T4 1.98 3.27 T1, T2 3.01 3.42T1, T2 2.95 .43 p< .01 F(2.87, 272.27)=63.90, Work/Occupational 96 7.48T2, T3, T4 2.41 3.46T1 2.76 4.14T1 3.51 4.29T1 3.34 .40 p< .01 F(2.70, 250.93)=43.93, Sexual 94 6.85T2, T3, T4 3.12 2.74T1, T3, T4 2.79 4.05T1, T2 3.83 3.71T1, T4 3.55 .32 p< .01 F(2.60, 249.95)=49.34, Self-Care 97 3.87T2, T3, T4 2.71 1.00 T1, T3, T4 1.54 1.88T1, T2 2.57 1.63T1, T2 2.39 .34 p< .01 F(2.55, 244.28)=40.59, Life-Support 97 4.04T2, T3, T4 2.89 1.08T1, T4 1.80 1.72T1 2.64 1.87T1, T2 2.60 .30 p< .01 F(2.11,194.30)=106.80, Total PDI 93 42.85T2, T3, T4 12.55 15.81T1, T3, T4 10.68 23.04T1, T2 18.12 22.52T1, T2 16.37 .54 p< .01 Notes. T1=Intake; T2=Discharge; T3=6 Months Post-Discharge; T4=12 Months Post-Discharge. *Family/Home Life results were interpreted by way of Greenhouse Geisser, all other results were interpreted by Huynh- Feldt. Subscripts of means denote significant mean differences.

Table 2. MMPI-2-RF and Pain Disability Index: Correlations T1 T2 T3 T4 N=232 N=208 N=101 N=232 EID .24** .14* .16 .08 RCd .22** .13 .12 .05 RC1 .22** .07 .23* .19** RC2 .29** .19** .25* .12 RC7 .16* .09 .14 .10 MLS .25** .12 .17 .11 GIC .06 -.09 -.07 .01 HPC .10 .07 .22* .12 NUC .31** .06 .33** .24** COG .21** .21** .32** .22** SUI .14* -.07 .22* .02 HLP .14* .19** .25* .14* SFD .16* .09 .11 .08 NFC .10 .12 .20* .14* STW .18** .04 .14 .09 AXY .22** .08 .21* .17* ANP .06 .07 .09 .07 BRF .17** .17* .23* .18** MSF -.05 .03 .02 .12 FML .12 .03 -.02 -.01 SAV .21** .08 .20* .03 DSF .19** .13 .23* .12 DISC-r .07 .04 .08 .07 NEGE-r .19** .02 .13 .05 INTR-r .25** .16* .19 .05 Notes. T1=Intake; T2=Discharge; T3=6 Months Post- Discharge; T4=12 Months Post-Discharge. *p<.05; **p<.01.

Table 3. Final Estimation of Fixed Effects: H-O, RC, and Somatic/Cognitive Specific Problems Scales. Predictor Level Coefficient SE t-ratio Df p-value Intercept EID 1 .22 .09 2.54 230 .01

RCd 1 -.42 .24 -1.74 227 .08 RC1 1 .58 .18 3.20 227 .00 RC2 1 1.14 .34 3.33 227 .00 RC7 1 .10 .23 .42 227 .67

MLS 1 .28 .61 .46 226 .64 GIC 1 -.99 .55 -1.80 226 .07 HPC 1 .30 .49 .62 226 .54 NUC 1 1.17 .38 3.09 226 .00 COG 1 .90 .37 2.46 226 .02 Time Time*EID 2 -.00 .00 -.81 230 .42

Time*RCd 2 -.00 .00 -1.54 227 .13 Time*RC1 2 .00 .00 1.01 227 .31 Time*RC2 2 .00 .00 .53 227 .59 Time*RC7 2 .00 .00 .78 227 .44

Time*MLS 2 -.00 .00 -1.41 226 .16 Time*GIC 2 -.00 .00 -.27 226 .79 Time*HPC 2 .00 .00 .50 226 .62 Time*NUC 2 .00 .00 .66 226 .51 Time*COG 2 .00 .00 .98 226 .33

Table 4. Final Estimation of Fixed Effects: Internalizing and Interpersonal Specific Problems Scales. Predictor Level Coefficient SE t-ratio Df p-value Intercept SUI 1 -.58 .81 -.72 222 .47 HLP 1 1.67 .70 2.40 222 .02 SFD 1 .13 .74 .18 222 .86 NFC 1 -.08 .45 -.17 222 .86 STW 1 -.23 .66 -.36 222 .72 AXY 1 1.71 .95 1.80 222 .07 ANP 1 -.32 .45 -.71 222 .48 BRF 1 1.62 .84 1.92 222 .06 MSF 1 -.50 .42 -1.20 222 .23

FLM 1 -.06 .37 -.17 226 .87 IPP 1 .05 .34 .15 226 .88 SAV 1 .17 .36 .47 226 .64 SHY 1 .25 .41 .61 226 .54 DSF 1 1.93 .85 2.27 226 .02 Time Time*SUI 2 .00 .00 .46 222 .65 Time*HLP 2 -.00 .00 -.71 222 .48 Time*SFD 2 -.00 .00 -.76 222 .45 Time*NFC 2 .00 .00 1.37 222 .17 Time*STW 2 -.00 .00 -.87 222 .39 Time*AXY 2 -.00 .00 -.08 222 .93 Time*ANP 2 .00 .00 .64 222 .52 Time*BRF 2 -.00 .00 -.17 222 .86 Time*MSF 2 .00 .00 2.21 222 .03

Time*FML 2 -.00 .00 -.94 226 .35 Time*IPP 2 -.00 .00 -.85 226 .40 Time*SAV 2 -.00 .00 -.68 226 .50 Time*SHY 2 .00 .00 .09 226 .93 Time*DSF 2 .00 .00 .68 226 .50

Table 5. Final Estimation of Fixed Effects: PSY-5 Scales. Predictor Level Coefficient SE t-ratio Df p-value Intercept AGGR-r 1 -.20 .27 -.74 226 .46 PSYC-r 1 1.46 .37 3.96 226 .00 DISC-r 1 .24 .27 .89 226 .37 NEGE-r 1 -.20 .21 -.96 226 .34 INTR-r 1 .53 .21 2.55 226 .01 Time Time*AGGR-r 2 .00 .00 .01 226 .99 Time*PSYC-r 2 .00 .00 1.56 226 .12 Time*DISC-r 2 .00 .00 .07 226 .94 Time*NEGE-r 2 -.00 .00 -.90 226 .37 Time*INTR-r 2 -.00 .00 -.80 226 .43

Appendix A (Pain Disability Index)

The rating scales below are designed to measure the degree to which aspects of your life are disrupted by chronic pain. In other words, we would like to know how much pain is preventing you from doing what you would normally do or from doing it as well as you normally would. Respond to each category indicating the overall impact of pain in your life, not just when pain is at its worst.

For each of the 7 categories of life activity listed, please circle the number on the scale that describes the level of disability you typically experience. A score of 0 means no disability at all, and a score of 10 signifies that all of the activities in which you would normally be involved have been totally disrupted or prevented by your pain.

1. Family/Home Responsibilities: This category refers to activities of the home or family. It includes chores or duties performed around the house (e.g. yard work) and errands or favors for other family members (e.g. driving the children to school). No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability

2. Recreation: This disability includes hobbies, sports, and other similar leisure time activities. No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability

3. Social Activity: This category refers to activities, which involve participation with friends and acquaintances other than family members. It includes parties, theater, concerts, dining out, and other social functions. No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability

4. Occupation: This category refers to activities that are part of or directly related to one’s job. This includes non-paying jobs as well, such as that of a housewife or volunteer. No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability

5. Sexual Behavior: This category refers to the frequency and quality of one’s sex life. No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability

6. Self Care: This category includes activities, which involve personal maintenance and independent daily living (e.g. taking a shower, driving, getting dressed, etc.) No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability

7. Life-Support Activities: This category refers to basic life supporting behaviors such as eating, sleeping and breathing. No Disability 0__. 1__. 2__. 3__. 4__. 5__. 6__. 7 __. 8__. 9__. 10__. Worst Disability