Psychological Assessment 2015, Vol. 27, No. 2, 433-446
© 2014 American Psychological Association 1040-3590/15/$ 12.00 http://dx.doi.org/10.1037/pas0000056
Reliability and Validity of the Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF) in Evaluations of Chronic Low Back Pain Patients Anthony M. Tarescavage
Judith Scheman
Kent State University
Cleveland Clinic Foundation, Neurological Center for Pain, Cleveland, Ohio
Yossef S. Ben-Porath Kent State University The purpose of the current study was to investigate the reliability and concurrent validity of Minnesota Multiphasic Personality Inventory (MMPI)-2-Restructured Form (2-RF) (Ben-Porath & Tellegen, 2008/ 2011) scores in a sample of 811 chronic low back pain patients (346 males, 529 females) beginning treatment in a short-term interdisciplinary pain rehabilitation program. We calculated internal consistency coefficients, mean-item correlations, and SEM for all substantive scales, as well as zero-order correlations with collateral medical record information and self-report testing. Results indicated reliability and validity for most of the MMPI-2-RF substantive scales. Implications of these findings and limitations of this study are discussed. Keywords: chronic pain, chronic low back pain, MMPI-2-RF, construct validity, applied assessment
Chronic low back pain is a significant problem. Prevalence estimates indicate that approximately 80% of the population will experience low back pain at some point (Rubin, 2007). The con dition is the second leading cause of disability in the United States (McNeil & Binette, 2001) and after the common cold, the leading reason for sick days (LaBar, 1992). Overall, about 149 million work days per year are lost because of low back pain (Guo, Tanaka, Halperin, & Cameron, 1999). The condition leads to approximately 100 to 200 billion dollars in losses per year, prin cipally because of diminished wages and productivity (Katz, 2006). The prevalence of the disorder has increased with one study reporting that the point prevalence of chronic low back pain rose from 3.9% in 1992 to 10.2% in 2006— over a 160% increase in a 14-year period (Freburger et al., 2009). Additionally, low back pain patients present with substantial rates of comorbid psychopathology. Von Korff et al. (2005) in vestigated the prevalence of mental disorders among individuals reporting chronic spinal pain in the National Comorbidity Survey
Replication sample (Kessler et al., 2004) and found that those with chronic back pain were at increased risk for mood and anxiety disorders, as well as substance use disorders. The 12-month prev alence of mood disorders in this population was 17.5% and the rate for anxiety disorders was 26.5%. These rates are meaningfully higher than the general population, which produces mood and anxiety disorder prevalence rates of 9.5% and 18.1%, respectively (Kessler, Chiu, Dernier, & Walters, 2005). The association between mental health problems and chronic pain is consistent with the bio-psychosocial perspective of pain (Gatchel, McGeary, McGeary, & Lippe, 2014; Gatchel, Peng, Peters, Fuchs, & Turk, 2007). According to this perspective, bio logical, psychological, and social factors interact to determine the experience of pain— a model that stands in stark contrast to earlier medical model notions that pain was mostly the result of physio logical pathology. Indeed, psychosocial variables have been found to be meaningful predictors of outcomes for a variety of medical procedures and conditions, especially back pain (Floogendoom, van Poppel, Bongers, Koes, & Bouter, 2000). For these reasons, the American College of Physicians and the American Pain Society recommends interdisciplinary treatment that includes an assessment of psychosocial factors that can con tribute to pain, which include depression and disability levels (Chou et al., 2007). These factors can be stronger predictors of outcome than physical examination information, as well as sever ity and duration of pain (Chou et al., 2007). One way to gather psychosocial information is through psychological testing. The Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1943) and MMPI-2 (Butcher et al., 2001) have historically been the most frequently used psychological tests in chronic pain settings (Piotrowski, 1998; Piotrowski & Lubin,
This article was published Online First December 1, 2014. Anthony M. Tarescavage, Department of Psychological Sciences, Kent State University; Judith Scheman, Cleveland Clinic Foundation, Neurolog ical Center for Pain, Cleveland, Ohio; Yossef S. Ben-Porath, Department of Psychological Sciences, Kent State University. Yossef S. Ben-Porath is a paid consultant to the MMPI-2-RF publisher, the University of Minnesota Press, and distributor, Pearson Assessments. He receives royalties on sales of MMPI-2-RF materials and research grants from the MMPI-2-RF publisher. Correspondence concerning this article should be addressed to Anthony M. Tarescavage, Department of Psychological Sciences, Kent State Uni versity, 144 Kent Hall, Kent, OH 44242. E-mail: [email protected]
433
TARESCAVAGE, SCHEMAN, AND BEN-PORATH
434
1990), but the tests began falling out of favor in this area in the mid- to late-1990s owing to psychometric, practical, and theoret ical limitations (Bradley, 1995; Keefe, Lefebvre, & Beaupre, 1995; Main & Spanswick, 1995; Sanders, 1995; Turk & Fernandez, 1995). The MMPI-2 Restructured Clinical (RC) Scales (Tellegen et al., 2003) were designed to address psychometric and concep tual problems with the Clinical Scales that were at the core of concerns about use of the inventory with chronic pain patients. Several studies comparing the psychometric functioning of the RC Scales relative to the Clinical Scales demonstrate marked improvement in reliability and validity (e.g., Arbisi, Sellbom, & Ben-Porath, 2008; Sellbom, Graham, & Schenk, 2006; Wygant et al., 2007). Relatedly, McCord and Drerup (2011) demonstrated improved interpretive utility of the RC Scales in comparison with the Clinical Scales among chronic pain patients. These authors categorized 316 chronic pain patients into depressed and nondepressed diagnostic groups. In the depressed group, mean clinical elevations were observed for the following Clinical Scales: 1 (Hypochondriasis), 2 (Depression), 3 (Hysteria), 4 (Psychopathic Deviate), 6 (Paranoia), 7 (Psychastenia), and 8 (Schizophrenia). In contrast, mean RC scale elevations were observed in the depressed group for only RCd (Demoralization), RC1 (Somatic Complaints), and RC2 (Low Positive Emotions), demonstrating substantially improved discriminant validity. However, no study published to date has investigated the psychometrics of the RC Scales among chronic pain patients undergoing conventional conservative treat ments. The MMPI-2-RF (Restructured Form) RC Scales form the core of the MMPI-2-RF (Ben-Porath & Tellegen, 2008/2011; Tellegen & Ben-Porath, 2008/2011), which shows promise for reviving MMPI research in chronic pain settings. As just noted, the test is anchored by the psychometrically improved RC Scales. The test developers used similar development strategies for two new scale sets that complement RC Scale interpretation: the Higher-Order and Specific Problems Scales of the MMPI-2-RF. Additionally, revised and improved versions of the MMPI-2 PersonalityPsychopatology-5 (PSY-5) are incorporated in the updated inven tory. The result is a 338-item broadband personalitypsychopathology measure with 42 substantive scales measuring five broad domains: Somatic/Cognitive Dysfunction, Emotional Dysfunction, Thought Dysfunction, Behavioral Dysfunction, and Interpersonal Functioning. In addition to the psychometric im provements just mentioned, because the MMPI-2-RF is consistent with hierarchical models of the structure of psychopathology (Sell bom, Ben-Porath, & Bagby, 2008), it is, in turn, relevant to the bio-psycho-social model of pain discussed earlier. The psycholog ical component of the bio-psycho-social model is primarily influ enced by mood and cognitions, and dysfunction in these areas is assessed by the MMPI-2-RFs Internalizing and Thought Dysfunc tion domains, respectively. The 40% reduction in the length of the MMPI-2-RF, although not the focus of the current investigation, may also make the test more tolerable for individuals experiencing significant levels of pain.
Current Study The purpose of the current study was to evaluate the MMPI2-RF for use in assessing chronic low back pain patients. To this end, we evaluated the reliability and concurrent validity of the
instrument in a sample of chronic low back pain patients entering multidisciplinary treatment. We also calculated MMPI-2-RF means and SDs for the sample by gender. We evaluated reliability for all substantive scales in terms of mean interitem correlations, SEM, and internal consistency. It was hypothesized that reliability coefficients in the chronic pain sample would approximate those presented for the normative sample. Specifically, we hypothesized that all substantive scales would have internal consistency reliabilities that are no more than .10 lower than the normative sample, mean interitem correlations that are no more than .05 lower than the normative sample, and SEM that are within one T score point of the normative sample. These criteria have been used in past research demonstrating the compa rability of MMPI-2-RF scale score reliability estimates in a med ical setting with normative sample estimates (Tarescavage, Wygant, Boutacoff, & Ben-Porath, 2013). Because they are longer, the Higher-Order, RC, and PSY-5 Scales were hypothesized to demonstrate relatively stronger reliability when compared with the Specific Problems Scales. To address the validity of the MMPI-2-RF in this setting, zero-order correlations were calculated between the MMPI-2-RF substantive scales and variables assessed at intake. These included information from patient medical records, psychiatric diagnoses, and physical ability test results, as well as self-report testing. It was hypothesized that MMPI-2-RF scales from the Internalizing Dys function domain would demonstrate statistically significant small to moderate associations with the Depression Anxiety Stress Scales (DASS; Lovibond & Lovibond, 1996), Pain Disability Index (PDI; Tait, Chibnall, & Krause, 1990), and mental health history variables. This hypothesis is consistent with the biopsychosocial model of pain, which postulates that mood and cog nitive factors interact with physical processes to exacerbate phys ical disability (Gatchel, 2004). Similarly, scales from the Thought Dysfunction domain were hypothesized to be correlated with the PDI because cognitive factors are also associated with pain dis ability. We also hypothesized that these MMPI-2-RF scales would be associated with antipsychotic medication use given that they measure constructs associated with psychotic disorders, and we expected that these scales would be correlated with posttraumatic stress disorder (PTSD) diagnosis based on findings in past research demonstrating that individuals with PTSD score higher on these scales because of reexperiencing and dissociative symptoms (Ar bisi, Polusny, Erbes, Thuras, & Reddy, 2011). Scores from the Externalizing domain were hypothesized to demonstrate small to moderate associations with current smoking and narcotic medica tion use, as this domain's scales measure acting out behaviors such as substance abuse. Similar findings were reported by Block, Ben-Porath, and Marek (2013) for a sample of spine surgery and spinal cord stimulator candidates. Somatic/Cognitive Dysfunction scale scores, which measure various physical complaints, were expected to have small to moderate correlations with hours resting per day, number of medications, narcotic pain medication use, duration of pain, intake pain intensity, depressive diagnosis, the PDI, physical ability tests, and the University of Alabama (UAB) Pain Behavior Scales (Richards, Nepomuceno, Riles, & Suer, 1982). Finally, the Interpersonal Functioning Scales were expected to demonstrate small to moderate associations with the PDI Family/Home, Recreation, Socialization, and Sexual disability scales, which measure pain disability particular to interpersonal situations.
MMPI-2-RF AND CHRONIC LOW BACK PAIN
Method Participants Participants included an archival sample of 875 nonconsecutive patients with chronic pain (346 males, 529 females) that completed the MMPI-2 and presented with lower back pain to a short-term multidisciplinary pain treatment program in Northeast Ohio. Ap proximately 85% of the sample completed the program. Patients who completed the program participated for an average of 21.1 days (SD = 10.0). Participants were excluded from the analyses if they produced invalid MMPI-2-RF profiles according to the test authors’ published guidelines, which included: Cannot say, CNS > Raw Score 18; Variable Response Inconsistency, VRINr > 80; True Response Inconsistency, TRIN-r > 80; Infrequent Responding, F-r = 120; and Infrequent Psychopathology Re sponses, Fp-r > 100 (Ben-Porath & Tellegen, 2008/2011). The final sample included 811 patients (318 males, 493 females) after exclusion. Consistent with standard interpretive guidelines, indi viduals demonstrating elevations on the Somatic/Cognitive over reporting validity scales were not excluded from the analyses. Overall, 6.2% of the sample produced scores >99T on Infrequent Somatic Responses (Fs) (3.5%), Symptom Validity (FBS-r) (1.8%), or the Response Bias Scale (RBS) (2.1%), indicating that they may have been overreporting somatic/cognitive problems. The majority of the sample was married (61.0%) and other marital statuses included divorced (17.9%), never married (15.9%), separated (3.1%), and widowed (2.1%). The average age was 46.7 (SD = 12.6) and the average years of education was 13.3 (SD = 3.8). The average duration of pain was 10.3 years (SD = 9.1) and the average pain intensity rating on a 0-10 scale at intake was 6.8 (SD = 1.9). Excluded individuals had significantly less education than the final sample, f(816) = 2.851, p = .004, Cohen’s d = .41, as well as significantly higher intake pain intensity ratings, t(816) = 3.427, p = .001, Cohen’s d = .45 No other significant differences were observed (ps > .150).
Measures MMPI-2-RF. As previously described in detail, the MMPI2-RF (Ben-Porath & Tellegen, 2008/2011) is a 338 item self-report broadband measure of psychopathology. It has nine Validity Scales designed to assess the test-taking approach and 42 Substan tive Scales that measure the following domains: Emotional Dys function, Thought Dysfunction, Behavioral Dysfunction, Somatic/ Cognitive Complaints, and Interpersonal Functioning. Psychometric findings presented in the test’s Technical Manual indicate adequate reliability and validity across a variety of set tings (Tellegen & Ben-Porath, 2008/2011). Pain Disability Index. The PDI (Tait et al., 1990) is a 7-item self-report measure that assesses pain-related disability in the following domains: Family/Home Responsibilities, Recreation, Social Activity, Occupation, Sexual Behavior, Self-Care, and LifeSupport Activity. It has demonstrated adequate test-retest reliabil ity and internal consistency among chronic pain patients (Tait et ah, 1990). In the current sample, the PDI items yielded a total score internal consistency of .83. Physical Ability Tests. These tests included the number of stairs climbed in 1 min, functional reach in inches, distance walked
435
in 6 min, and time to rise from seated position. These physical ability performance tests were administered at the beginning of the multidisciplinary pain treatment program and are commonly used among people with chronic low back pain (Andersson, Lin, & Smeets, 2010). Depression, Anxiety, Stress Scales. The DASS (Lovibond & Lovibond, 1996) is a 42-item self-report measure of mood prob lems. It has three scales measuring depression, anxiety, and gen eralized distress. It has demonstrated adequate internal consistency reliability in a variety of settings (Lovibond & Lovibond, 1996). UAB Pain Behavior Rating Scales. The UAB Pain Behavior Scales (Richards et ah, 1982) is a 10-item clinician rated measure that assesses the number of physical pain behaviors a patient displays. Specifically, it assesses the following behaviors: Verbal Complaints, Non-Verbal Complaints, Down-Time, Facial Gri maces, Standing Posture, Mobility, Body Language, Use of Visi ble Support Equipment, Stationary Movement, and Medication Usage. Medical Record Information. A number of variables were extracted from the patient’s medical record. These included patient demographics, diagnoses, substance use history, medications, and intake pain intensity ratings.
Procedure All chronic pain patients received a psychological evaluation at intake into the short-term chronic pain rehabilitation program. These psychological evaluations included the MMPI-2 and a va riety of self-report measures. Although the MMPI-2 was routinely administered to patients, in some cases it was not given owing to reading limitations, potential cognitive deficits, or time consider ations. In the current study, MMPI-2 items were used to calculate MMPI-2-RF scale scores, which is possible because all 338 MMPI-2-RF items are included in the MMPI-2 booklet. Past research has demonstrated the relative comparability of MMP2-RF scale scores generated from both booklets (Tellegen & Ben-Porath, 2008/2011; Van der Heijden, Egger, & Derksen, 2010). Data were collected from 1999 to 2008.
Results Descriptive Findings Means and SD for all MMPI-2-RF validity and substantive scales are presented for the full sample and by gender in Table 1. Mean differences between genders of five or more T score points (i.e., one half SD in the general population) are considered clinically meaning ful, consistent with traditional benchmarks for the interpretation of the MMPI-2 and MMPI-2-RF (Graham, 2012). Clinically meaningful gender differences were found on the following scales: Fs, FBS-r, Behavioral/Extemalizing Dysfunction (BXD), Antisocial Behavior (RC4), Gastrointestinal Complaints (GIC), Multiple Specific Fears (MSF), Juvenile Conduct Problems (JCP), Substance Abuse (SUB), Aesthetic-Literary Interests (AES), Mechanical-Physical Interests (MEC), Aggressiveness-Revised (AGGR-r), and DisconstraintRevised (DISC-r). Men scored higher than women on BXD, RC4, JCP, SUB, MEC, AGGR-r, and DISC-r, whereas women scored higher than men on Fs, FBS-r, GIC, MSF, and AES. Differences between genders were most pronounced on MEC (men scored 13T
TARESCAVAGE, SCHEMAN, AND BEN-PORATH
436
Table 1 Minnesota Multiphasic Personality Inventory-2-Restructured Form Sample Means and SDs by Gender Women (n = 493)
Men (n = 318)
Scales Validity Scales Variable Response Inconsistency (VRIN-r) True Response Inconsistency (TRIN-r) Infrequent Responses (F-r) Infrequent Psychopathology Responses (FP-r) Infrequent Somatic Responses (Fs) Symptom Validity (FBS-r) Response Bias Scale (RBS) Uncommon Virtues (L-r) Adjustment Validity (K-r) Higher-Order Emotional/Internalizing Dysfunction (EID) Thought Dysfunction (THD) Behavioral/Extemalizing Dysfunction (BXD) Restructured Clinical Demoralization (RCd) Somatic Complaints (RC1) Low Positive Emotions (RC2) Cynicism (RC3) Antisocial Behavior (RC4) Persecutory Ideation (RC6) Dysfunctional Negative Emotions (RC7) Aberrant Experiences (RC8) Hypomanic Activation (RC9) Specific Problems Malaise (MLS) Gastrointestinal Complaints (GIC) Head Pain Complaints (HPC) Neurological Complaints (NUC) Cognitive Complaints (COG) Suicidal Ideation (SUI) Helplessness (HLP) Self-Doubt (SFD) Inefficacy (NFC) StressAVorry (STW) Anxiety (AXY) Anger Proneness (ANP) Behavior Restricting Fears (BRF) Multiple Specific Fears (MSF) Juvenile Conduct Problems (JCP) Substance Abuse (SUB) Aggression (AGG) Activation (ACT) Family Problems (FML) Interpersonal Passivity (IPP) Social Avoidance (SAV) Shyness (SHY) Disaffiliativeness (DSF) Aesthetic-Literary Interests (AES) Mechanical-Physical Interests (MEC) Personality Psychopathology-5 Aggressiveness-Revised (AGGR-r) Psychoticism-Revised (PSYC-r) Disconstraint-Revised (DISC-r) Negative Emotionality/Neuroticism-Revised (NEGE-r) Introversion/Low Positive Emotionality-Revised (INTR-r)
Overall (N = 811)
M
SD
M
SD
M
SD
51 51F 70 53 62 68 64 55 48
9 9 16 11 16 13 15 11 10
52 51F 70 53 67 75 68 55 47
10 11 16 11 19 13 15 9 9
52 51F 70 53 65 72 66 55 47
10 10 16 11 18 13 15 10 9
61 51 55
12 10 11
63 54 46
11 11 9
62 53 49
12 11 10
61 69 64 49 56 53 52 52 48
12 11 13 10 11 11 11 10 10
64 73 63 48 49 54 55 54 45
11 12 12 10 9 11 11 11 8
63 72 63 49 52 54 53 53 46
11 11 12 10 11 11 11 11 9
77 61 63 70 60 57 54 57 52 55 56 54 51 46 55 54 51 46 50 49 56 51 54 42 57
8 16 9 13 14 17 14 12 11 12 15 12 9 7 13 11 11 9 11 10 11 11 14 9 10
77 67 66 71 64 56 54 59 55 57 59 53 55 54 49 48 48 49 54 53 52 51 51 47 44
8 17 10 13 14 17 12 12 12 11 14 11 12 10 10 8 8 10 12 10 11 10 11 9 6
77 64 65 71 62 57 54 58 54 56 58 53 53 51 51 50 49 48 52 51 54 51 52 45 49
8 17 10 13 14 17 13 12 12 11 14 11 11 10 12 10 10 10 12 10 11 10 12 9 10
51 52 55 56 60
10 10 11 13 13
46 54 44 58 57
8 10 7 12 12
48 53 48 57 58
9 10 10 12 12
higher than women). Both genders produced mean substantive scale clinical elevations (i.e., a score >65T) on Somatic Complaints (RC1), Malaise (MLS), and Neurological Complaints (NUC), and women additionally had a mean elevation on GIC.
Reliability Mean interitem correlations, Cronbach’s a internal consistency, and SEM estimates are provided in Table 2 for all substantive
MMPI-2-RF AND CHRONIC LOW BACK PAIN
437
Table 2 Minnesota Multiphasic Personality Inventory-2-Restructured Form Reliability Estimates in Chronic Low Back Pain (N = 811) and Normative Samples (N = 2276) Mean inter-item correlation Scale name Higher-Order Emotional/Internalizing Dysfunction (EID) Thought Dysfunction (THD) Behavioral/Extemalizing Dysfunction (BXD) Restructured Clinical Demoralization (RCd) Somatic Complaints (RC1) Low Positive Emotions (RC2) Cynicism (RC3) Antisocial Behavior (RC4) Persecutory Ideation (RC6) Dysfunctional Negative Emotions (RC7) Aberrant Experiences (RC8) Hypomanic Activation (RC9) Specific Problems Malaise (MLS) Gastrointestinal Complaints (GIC) Head Pain Complaints (HPC) Neurological Complaints (NUC) Cognitive Complaints (COG) Suicidal Ideation (SUI) Helplessness (HLP) Self-Doubt (SFD) Inefficacy (NFC) Stress/Worry (STW) Anxiety (AXY) Anger Proneness (ANP) Behavior Restricting Fears (BRF) Multiple Specific Fears (MSF) Juvenile Conduct Problems (JCP) Substance Abuse (SUB) Aggression (AGG) Activation (ACT) Family Problems (FML) Interpersonal Passivity (IPP) Social Avoidance (SAV) Shyness (SHY) Disaffiliativeness (DSF) Aesthetic-Literary Interests (AES) Mechanical-Physical Interests (MEC) Personality-Psychopathology-5 Aggressiveness-Revised (AGGR-r) Psychoticism-Revised (PSYC-r) Disconstraint-Revised (DISC-r) Negative Emotionality/Neuroticism-Revised (NEGE-r) Introversion/Low Positive Emotionality-Revised (INTR-r)
Note.
Internal consistency
SEM
LBP
NORM
LBP
NORM
LBP
NORM
Number of items
.19 .07 .14
.15 .08 .14
.90 .67 .79
.87 .69 .76
3.7 6.1 4.7
3.6 5.6 4.7
41 26 23
.28 .13 .13 .21 .13 .08 .18 .12 .10
.24 .11 .10 .21 .13 .09 .16 .11 .11
.90 .80 .71 .80 .77 .61 .84 .72 .76
.88 .76 .66 .80 .75 .65 .82 .71 .78
3.6 5.1 6.4 4.4 5.2 6.9 4.5 5.7 4.3
3.3 4.7 5.9 4.5 4.8 5.9 4.0 5.5 4.7
24 27 17 15 22 17 24 18 28
.10 .32 .22 .17 .24 .29 .24 .43 .26 .18 .20 .28 .09 .22 .29 .20 .16 .15 .20 .18 .29 .32 .18 .17 .25
.19 .27 .27 .11 .16 .16 .15 .38 .22 .16 .15 .27 .09 .25 .24 .20 .16 .15 .16 .19 .26 .31 .13 .17 .24
.47 .70 .63 .67 .76 .67 .61 .75 .76 .61 .55 .73 .46 .71 .71 .64 .63 .58 .71 .69 .80 .77 .57 .58 .75
.62 .67 .64 .55 .67 .38 .45 .70 .71 .56 .44 .72 .47 .70 .61 .62 .62 .60 .66 .70 .78 .76 .47 .55 .59
5.5 9.1 5.9 7.7 6.9 9.7 8.0 6.0 5.7 7.1 9.5 5.8 8.3 5.1 6.3 6.0 5.8 6.4 6.2 5.7 5.0 5.0 8.0 6.1 5.1
5.9 5.8 6.0 6.7 5.8 7.9 7.5 5.5 5.4 6.3 7.5 5.4 7.3 5.0 6.3 5.9 6.1 6.3 5.9 5.6 4.8 4.7 7.3 6.3 4.8
8 5 6 10 10 5 5 4 9 7 5 7 9 9 6 7 9 8 10 10 10 7 6 7 9
.13 .07 .14 .16 .16
.14 .08 .14 .14 .13
.73 .65 .77 .80 .79
.73 .69 .71 .77 .75
4.8 6.1 4.9 5.5 5.7
5.3 5.6 4.9 4.8 5.0
18 26 20 20 20
LBP = Low Back Pain Sample; NORM = Normative Sample.
scales. Also included are comparable estimates in the MMPI-2-RF normative sample (Tellegen & Ben-Porath, 2008/2011). Internal consistency estimates for the Higher-Order Scales ranged from .67 (Thought Dysfunction) to .90 (Emotional/Internalizing Dysfunction) with a median of .79. Among the Restruc tured Clinical Scales, reliability estimates ranged from .61 (Perse cutory Ideation) to .90 (Demoralization) with a median of .77. For the Specific Problems Scales, internal consistency ranged from .46 (Behavior Restricting Fears) to .80 (Social Avoidance) with a median of .67. Finally, among the Personality-Psychopatology-5 Scales, internal consistency ranged from .65 (Psychoticism-
Revised) to .80 (Negative Emotionality/Neuroticism-Revised) with a median of .77. Overall, internal consistency estimates were consistent with the normative sample, except for Malaise, which had a substantially lower estimate (a . 10) in this sample. Mean interitem correlations for the scales that provide the broadest level of measurement, the Higher-Order Scales, ranged from .07 (Thought Dysfunction) to .19 (Emotional/Internalizing Dysfunction) with a median of .14. Among the Restructured Clin ical Scale, mean interitem correlation estimates ranged from .08 (Persecutory Ideation) to .28 (Demoralization) with a median of .13. For the shortest scales, the Specific Problems Scales, mean
438
TARESCAVAGE, SCHEMAN, AND BEN-PORATH
interitem correlations ranged from .09 (Anger Proneness) to .43 (Self Doubt) with a median of .22. Finally, among the PersonalityPsychopatology-5 Scales, mean interitem correlations ranged from .07 (Psychoticism-Revised) to .16 (Negative Emotionality/ Neuroticism-Revised and Introversion/Low Positive EmotionalityRevised) with a median of .14. Overall, mean interitem correla tions were consistent with the normative sample, except for Malaise and Head Pain Complaints, which had appreciably lower (>.05) mean interitem correlations. Also provided in Table 2 are SEM for all substantive scales. For the Higher-Order Scales, SEMs (expressed in T score units) ranged from 3.7 (Emotional/Intemalizing Dysfunction) to 6.1 (Thought Dysfunction) with a median of 4.7. Among the Restructured Clin ical Scales, SEM ranged from 3.6 (Demoralization) to 6.9 (Perse cutory Ideation) with a median of 5.1. For the Specific Problems Scales, SEM ranged from 5.1 (Multiple Specific Fears) to 9.7 (Suicidal Ideation) with a median of 6.1. Finally, among the Personality-Psychopatology-5 Scales, SEM ranged from 4.8 (Aggressiveness-Revised) to 6.1 (Psychoticism-Revised) with a median of 5.5. Overall, the estimates for the substantive scales were consistent with those obtained in the normative sample, except for Gastrointestinal Complaints, Cognitive Complaints, Suicidal Ideation, and Anxiety, which had substantially higher SEM estimates ( a l T) in this sample.
Concurrent External Correlates Zero-order correlations between MMPI-2-RF substantive scales and external chart review, PDI, and DASS criteria are presented in Table 3 for the Higher-Order, Restructured Clinical, and Somatic/ Cognitive Specific Problem Scales, in Table 4 for the Internalizing and Externalizing Specific Problem Scales, and in Table 5 for the Interpersonal Specific Problem Scales, Interest Scales, and Pesonality-Psychopathology-5 Scales. We applied a family wise Bonferroni correction to reduce the risk of Type I error, which yielded a corrected a level of .001 (.05/33 comparisons for each scale). Only correlations that met this a level were interpreted. To facilitate interpretation, the findings are summarized in reference to each of the five MMPI-2-RF domains, which include: (1) Emotional Dysfunction; (2) Thought Dysfunction; (3) Behavioral Dysfunction; (4) Somatic/Cognitive Complaints; and (5) Interper sonal Functioning. The scales in the Emotional Dysfunction domain include Higher-Order Scale Emotional/Intemalizing Dysfunction (EID) and Restructured Clinical Scales Demoralization (RCd), Low Pos itive Emotions (RC2), and Dysfunctional Negative Emotions (RC7). This domain also includes the Internalizing Specific Prob lem Scales Suicidal/Death Ideation (SUI), Helplessness/Hopelessness (HLP), Self-Doubt (SFD), Inefficacy (NFC), Stress/Worry (STW), Anxiety (AXY), Anger Proneness (ANP), BehaviorRestricting Fears (BRF), and Multiple Specific Fears (MSF). Fi nally, this domain includes Personality-Psychopathology-5 Scales Negative Emotionality/Neuroticism-Revised (NEGE-r) and Low Positive Emotionality/Introversion-Revised (INTR-r). Scales in this domain demonstrated a number of convergent correlations with mental health history information. Many scales were associ ated with psychotropic medication use, including antidepressants (EID, RCd, SFD, STW, AXY, and NEGE-r) and benzodiazepines (STW, AXY, and NEGE-r), though some scales demonstrated
correlations with antipsychotic medication use as well (STW, AXY, and NEGE-r). Scales in this domain were associated with mental health diagnoses, including major depressive disorder (EID, RCd, RC2, RC7, SUI, HLP, SFD, NFC, STW, AXY, ANP, BRF, NEGE-r, and INTR-r), bipolar disorder (RC7 and AXY), anxiety disorders (STW and AXY), and PTSD (AXY). The Emo tional Dysfunction domain’s scales correlated with the PDI (EID, RCd, RC2, SUI, and INTR-r), as well as the DASS (EID, RCd, RC2, RC7, SUI, HLP, SFD, NFC, STW, AXY, ANP, BRF, and NEGE-r). Regarding the DASS correlations, most scales demon strated their strongest correlations with DASS Depression (EID, RCd, RC2, SUI, HLP, and SFD), though some scales correlated highest with DASS Anxiety (NFC, STW, AXY, and BRF) and DASS Stress (RC7, ANP, and NEGE-r). Many of the scales in this domain were associated with the number of hours resting per day (EID, RCd, RC2, SUI, HLP, SFD, and INTR-r). The MMPI-2-RFs Emotional Dysfunction’s scales were generally uncorrelated with pain and physical ability tests. The scales in the Thought Dysfunction domain include HigherOrder Scale Thought Dysfunction (THD), Restructured Clinical Scales Persecutory Ideation (RC6) and Aberrant Experiences (RC8), and Personality-Psychopathology-5 Scale PsychoticismRevised (PSYC-r). Scales in this domain were correlated with mental health diagnoses, including PTSD and bipolar disorder (THD, RC6, RC8, and PSYC-r). These scales also demonstrated associations with somatoform disorder (THD, RC8, and PSYC-r), major depressive disorder (THD, RC8, and PSYC-r), and sub stance use disorder (RC8 and PSYC-r). As expected, some of the scales correlated with the PDI self-care and life support scales (RC8 and PSYC-r). All were correlated to some extent with the DASS Scales. The MMPI-2-RFs Thought Dysfunction scales were generally uncorrelated with pain and physical ability tests. The scales in the Behavioral Dysfunction domain include Higher-Order Scale Behavioral/Externalizing Dysfunction (BXD), Restructured Clinical Scales Antisocial Behavior (RC4) and Hypomanic Activation (RC9), and the Externalizing Specific Problem Scales Juvenile Conduct Problems (JCP), Substance Abuse (SUB), Aggression (AGG), and Activation (ACT). This domain also in cludes Personality-Psychopathology-5 Scales AGGR-r and DISC-r. These scales demonstrated associations with mental health history information, particularly a history of substance use disorder (BXD, RC4, RC9, JCP, SUB, AGG, and DISC-r) and PTSD (BXD, RC4, RC9, JCP, AGG, and DISC-r). Scales in this domain were also associated with smoking (BXD, RC4, JCP, SUB, and DISC-r). Some negative associations were demonstrated with pain disability (BXD, SUB, and DISC-r) and poor physical ability test results (DISC-r). Scales in this domain were generally uncorrelated with pain, internalizing problems, and narcotic medication use. The scales in the Somatic/Cognitive Complaints domain include Restructured Clinical Scale Somatic Complaints (RC1) and the Somatic/Cognitive Specific Problem Scales, which are comprised of MLS, GIC, Head Pain Complaints (HPC), NUC, and Cognitive Complaints (COG). These scales demonstrated convergent associ ations with measures of pain, including intake pain intensity (RC1 and NUC) and the UAB (NUC). They were correlated with hours resting per day (RC1, MLS, and NUC) and the PDI scales (RC1, MLS, HPC, NUC, and COG). A number of associations were observed for medication use, including benzodiazepines (RC1 and NUC), antidepressants (RC1), and prescription narcotics (NUC).
MMPI-2-RF AND CHRONIC LOW BACK PAIN
—
i n co d 0 — 0
x 0
h 0
*
*
*
—
i n d x x — — o o
#
*
*
* *
*
o cc t
os d
m
m d - — o — —
»n o
Table 3 Minnesota Multiphasic Personality Inventory-2-Restructured Form Higher Order, Restructured Clinical, and Somatic/Cognitive Specific Problem Scale Correlations With Criteria
1
d- dl oo 0 0 0 1
*
*
I
n o N - t o c o c o n - t f - v o 0 0 - 0 0 0 0 ------------< —
C O0 -
T X O d 0 0 0 0
n ox
V, b
w *i c- *n d
t dl
t -t
On —
x dl
d - - t o o d i o o o x d — in o o o — — o o o o o
( OCOX o — o
— C 4 — — o —
T t ( N O O O X C 4 X
O OO N in — dl —
^t 0
1
0
0
o
1
0
0
0
— 0
—
*
— — O O O O
*
*
o x t N r - i n x r ' — o o o o o o
— —
I
' t M J ' N O ' O l N O i O b l O O O O — O O — O dl
d d m — C MC N
dl 0
— co o o o
x o
h d in — — o o
in o
d ( Nd l
d- os d n * - co — — — dl —
X n
o — cn n
do
1
O oo — o o o
1
-
- t on 0 0 0
T O
1
x co dl ■'t 0 0 0 0
0
1
cl T T — — O O
1
X di — o
I I
o
d o
d T — —
Xoo 0 0
1
*- r i r, — O O O O
T
X
g
X
IQ | T
d- d- d— — dl
— X d l d ~ X O 0 0 0 0 0 0
1
1
1
t o x o o x m x — o o o o o — —
1
N Od d d
1
d0 1
On ci
dl 0 1
— d i o x o N O c n o m r 0 — 0 0 0 — 0 — 0 0
O i n On d — C H O O
1
in rf — o o
1
I
Tt o
oo o o
I I
I
1
- - x o o
— t o o
I
t d X o o o
T
——
1
com — o o o
— t o o
cn — i n O' 0
1
oo i n o 0 0 0
Vi I n d
o
t o
1
d o
— OnX dl — dl
t o
o
1
x d X d * x Tj- o o o
I
I
d in — —
m ■*?
x o
1
1
-
1
— d 0 —
x 0
d o
0 1
0 1
t 0
1
d cn i n 0 0 0
1
©
1
1
i n - i n - ' t o d o o o o o — —
d d
o
o d
X 0
O' d - 0 0
OnX 0 0
x o i n V
V
o d i d -
d x cn —
in c n
cs o
X o
O'
—
O Nd o —
O o
d
t
d
d
t
—
x 0
1
oo — — —
ro
— —
d i n d ’- h ,--------- (N(N
d d
O in
X d
o
In
—
—
— Os ' t cn o o
O
—
X o
*
*
*
— in 0 0
X t d O O O 1
1
1
X
1
d
s
1
d l dl
1
1
1
1
* cn -t
1
c o in — — c i o\ o o\ in o o ——o o o o o o o
C)
* o 1 1
*
x
iq
g
* o
oo
i n o N X d d O N d — o o — — o — —
nq n
w
i— n on d 0o 0 0 1 1 1 1
»n t 0 -
-
in 0 0
in o >n t
md
— 0 —
dt
xcc x t O O O
I I
—d ' i n x x n c —o o o o o
1
cn
O
d
n c n
——
c i i n d f c n c i f c c c f c i n
o o - o x — dl — O — O
d
di
m
—x
d
d
d
o
X
d
d
O N
*-
—O
oo
d-
dl 1
C N 00
- d - r ^ r - ' d - d - x x
N Od
r-
—O
i > d-’ d^
O '-
O N im n in O d in x
d*'
t—X
X
X
X
X
X
X
>n in
o
di
cs
di
-
0
1
—C Nto O S d l — >n to 0 —0 0 —0 — 1 1
1
1
1
oo d
O
On t
d
1
1
dl OO 0 0 0
O N
1
1
' O ' O O c i T d i c i m o T f o m o o o o o o o o
On
*
1
1
1
1
1
*
—d
3 * 2
1
1
1
x
8
cn o
1
- C ' l O T t C C O m T O ' t C — 0 — 0 — 0 0 — 0 —
1
1
1
1
x r - cn o O O O O
t d X O O O
1
t d
1
i n — oo o o o
8
1
I
- C l i f M O l C h O ' O T C O 0 — o o o o o o o — 1
o
1
o c - C ' O — TfN — n o O c d O O O O O O O O
re ot m c s o o
i n r o o o o
1
i n o i n c K N T f t n o \ m 0 C 4 0 0 0 0 0 0 0 0
o o
d O' — O
1
1
11
1
1
1
1
1
1
1
dl d o o
I
1
-t 0
1
1
* 0
1
i n oo o —
1
0
1
*
1
d
O
1
O *n
1
1
t
1
O No
■'t- in
00 00
u u u y < Q >n C U- O c L s? 2 C N ^ y X •c G -o ‘•S && •G • s-1 “ *Q . y C C 0) 3 W) o P" ^a..*° .c >s
~
S'? x ) . S G
O
O
•S n rt
on on
© 2014 American Psychological Association 1040-3590/15/$ 12.00 http://dx.doi.org/10.1037/pas0000056
Reliability and Validity of the Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF) in Evaluations of Chronic Low Back Pain Patients Anthony M. Tarescavage
Judith Scheman
Kent State University
Cleveland Clinic Foundation, Neurological Center for Pain, Cleveland, Ohio
Yossef S. Ben-Porath Kent State University The purpose of the current study was to investigate the reliability and concurrent validity of Minnesota Multiphasic Personality Inventory (MMPI)-2-Restructured Form (2-RF) (Ben-Porath & Tellegen, 2008/ 2011) scores in a sample of 811 chronic low back pain patients (346 males, 529 females) beginning treatment in a short-term interdisciplinary pain rehabilitation program. We calculated internal consistency coefficients, mean-item correlations, and SEM for all substantive scales, as well as zero-order correlations with collateral medical record information and self-report testing. Results indicated reliability and validity for most of the MMPI-2-RF substantive scales. Implications of these findings and limitations of this study are discussed. Keywords: chronic pain, chronic low back pain, MMPI-2-RF, construct validity, applied assessment
Chronic low back pain is a significant problem. Prevalence estimates indicate that approximately 80% of the population will experience low back pain at some point (Rubin, 2007). The con dition is the second leading cause of disability in the United States (McNeil & Binette, 2001) and after the common cold, the leading reason for sick days (LaBar, 1992). Overall, about 149 million work days per year are lost because of low back pain (Guo, Tanaka, Halperin, & Cameron, 1999). The condition leads to approximately 100 to 200 billion dollars in losses per year, prin cipally because of diminished wages and productivity (Katz, 2006). The prevalence of the disorder has increased with one study reporting that the point prevalence of chronic low back pain rose from 3.9% in 1992 to 10.2% in 2006— over a 160% increase in a 14-year period (Freburger et al., 2009). Additionally, low back pain patients present with substantial rates of comorbid psychopathology. Von Korff et al. (2005) in vestigated the prevalence of mental disorders among individuals reporting chronic spinal pain in the National Comorbidity Survey
Replication sample (Kessler et al., 2004) and found that those with chronic back pain were at increased risk for mood and anxiety disorders, as well as substance use disorders. The 12-month prev alence of mood disorders in this population was 17.5% and the rate for anxiety disorders was 26.5%. These rates are meaningfully higher than the general population, which produces mood and anxiety disorder prevalence rates of 9.5% and 18.1%, respectively (Kessler, Chiu, Dernier, & Walters, 2005). The association between mental health problems and chronic pain is consistent with the bio-psychosocial perspective of pain (Gatchel, McGeary, McGeary, & Lippe, 2014; Gatchel, Peng, Peters, Fuchs, & Turk, 2007). According to this perspective, bio logical, psychological, and social factors interact to determine the experience of pain— a model that stands in stark contrast to earlier medical model notions that pain was mostly the result of physio logical pathology. Indeed, psychosocial variables have been found to be meaningful predictors of outcomes for a variety of medical procedures and conditions, especially back pain (Floogendoom, van Poppel, Bongers, Koes, & Bouter, 2000). For these reasons, the American College of Physicians and the American Pain Society recommends interdisciplinary treatment that includes an assessment of psychosocial factors that can con tribute to pain, which include depression and disability levels (Chou et al., 2007). These factors can be stronger predictors of outcome than physical examination information, as well as sever ity and duration of pain (Chou et al., 2007). One way to gather psychosocial information is through psychological testing. The Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1943) and MMPI-2 (Butcher et al., 2001) have historically been the most frequently used psychological tests in chronic pain settings (Piotrowski, 1998; Piotrowski & Lubin,
This article was published Online First December 1, 2014. Anthony M. Tarescavage, Department of Psychological Sciences, Kent State University; Judith Scheman, Cleveland Clinic Foundation, Neurolog ical Center for Pain, Cleveland, Ohio; Yossef S. Ben-Porath, Department of Psychological Sciences, Kent State University. Yossef S. Ben-Porath is a paid consultant to the MMPI-2-RF publisher, the University of Minnesota Press, and distributor, Pearson Assessments. He receives royalties on sales of MMPI-2-RF materials and research grants from the MMPI-2-RF publisher. Correspondence concerning this article should be addressed to Anthony M. Tarescavage, Department of Psychological Sciences, Kent State Uni versity, 144 Kent Hall, Kent, OH 44242. E-mail: [email protected]
433
TARESCAVAGE, SCHEMAN, AND BEN-PORATH
434
1990), but the tests began falling out of favor in this area in the mid- to late-1990s owing to psychometric, practical, and theoret ical limitations (Bradley, 1995; Keefe, Lefebvre, & Beaupre, 1995; Main & Spanswick, 1995; Sanders, 1995; Turk & Fernandez, 1995). The MMPI-2 Restructured Clinical (RC) Scales (Tellegen et al., 2003) were designed to address psychometric and concep tual problems with the Clinical Scales that were at the core of concerns about use of the inventory with chronic pain patients. Several studies comparing the psychometric functioning of the RC Scales relative to the Clinical Scales demonstrate marked improvement in reliability and validity (e.g., Arbisi, Sellbom, & Ben-Porath, 2008; Sellbom, Graham, & Schenk, 2006; Wygant et al., 2007). Relatedly, McCord and Drerup (2011) demonstrated improved interpretive utility of the RC Scales in comparison with the Clinical Scales among chronic pain patients. These authors categorized 316 chronic pain patients into depressed and nondepressed diagnostic groups. In the depressed group, mean clinical elevations were observed for the following Clinical Scales: 1 (Hypochondriasis), 2 (Depression), 3 (Hysteria), 4 (Psychopathic Deviate), 6 (Paranoia), 7 (Psychastenia), and 8 (Schizophrenia). In contrast, mean RC scale elevations were observed in the depressed group for only RCd (Demoralization), RC1 (Somatic Complaints), and RC2 (Low Positive Emotions), demonstrating substantially improved discriminant validity. However, no study published to date has investigated the psychometrics of the RC Scales among chronic pain patients undergoing conventional conservative treat ments. The MMPI-2-RF (Restructured Form) RC Scales form the core of the MMPI-2-RF (Ben-Porath & Tellegen, 2008/2011; Tellegen & Ben-Porath, 2008/2011), which shows promise for reviving MMPI research in chronic pain settings. As just noted, the test is anchored by the psychometrically improved RC Scales. The test developers used similar development strategies for two new scale sets that complement RC Scale interpretation: the Higher-Order and Specific Problems Scales of the MMPI-2-RF. Additionally, revised and improved versions of the MMPI-2 PersonalityPsychopatology-5 (PSY-5) are incorporated in the updated inven tory. The result is a 338-item broadband personalitypsychopathology measure with 42 substantive scales measuring five broad domains: Somatic/Cognitive Dysfunction, Emotional Dysfunction, Thought Dysfunction, Behavioral Dysfunction, and Interpersonal Functioning. In addition to the psychometric im provements just mentioned, because the MMPI-2-RF is consistent with hierarchical models of the structure of psychopathology (Sell bom, Ben-Porath, & Bagby, 2008), it is, in turn, relevant to the bio-psycho-social model of pain discussed earlier. The psycholog ical component of the bio-psycho-social model is primarily influ enced by mood and cognitions, and dysfunction in these areas is assessed by the MMPI-2-RFs Internalizing and Thought Dysfunc tion domains, respectively. The 40% reduction in the length of the MMPI-2-RF, although not the focus of the current investigation, may also make the test more tolerable for individuals experiencing significant levels of pain.
Current Study The purpose of the current study was to evaluate the MMPI2-RF for use in assessing chronic low back pain patients. To this end, we evaluated the reliability and concurrent validity of the
instrument in a sample of chronic low back pain patients entering multidisciplinary treatment. We also calculated MMPI-2-RF means and SDs for the sample by gender. We evaluated reliability for all substantive scales in terms of mean interitem correlations, SEM, and internal consistency. It was hypothesized that reliability coefficients in the chronic pain sample would approximate those presented for the normative sample. Specifically, we hypothesized that all substantive scales would have internal consistency reliabilities that are no more than .10 lower than the normative sample, mean interitem correlations that are no more than .05 lower than the normative sample, and SEM that are within one T score point of the normative sample. These criteria have been used in past research demonstrating the compa rability of MMPI-2-RF scale score reliability estimates in a med ical setting with normative sample estimates (Tarescavage, Wygant, Boutacoff, & Ben-Porath, 2013). Because they are longer, the Higher-Order, RC, and PSY-5 Scales were hypothesized to demonstrate relatively stronger reliability when compared with the Specific Problems Scales. To address the validity of the MMPI-2-RF in this setting, zero-order correlations were calculated between the MMPI-2-RF substantive scales and variables assessed at intake. These included information from patient medical records, psychiatric diagnoses, and physical ability test results, as well as self-report testing. It was hypothesized that MMPI-2-RF scales from the Internalizing Dys function domain would demonstrate statistically significant small to moderate associations with the Depression Anxiety Stress Scales (DASS; Lovibond & Lovibond, 1996), Pain Disability Index (PDI; Tait, Chibnall, & Krause, 1990), and mental health history variables. This hypothesis is consistent with the biopsychosocial model of pain, which postulates that mood and cog nitive factors interact with physical processes to exacerbate phys ical disability (Gatchel, 2004). Similarly, scales from the Thought Dysfunction domain were hypothesized to be correlated with the PDI because cognitive factors are also associated with pain dis ability. We also hypothesized that these MMPI-2-RF scales would be associated with antipsychotic medication use given that they measure constructs associated with psychotic disorders, and we expected that these scales would be correlated with posttraumatic stress disorder (PTSD) diagnosis based on findings in past research demonstrating that individuals with PTSD score higher on these scales because of reexperiencing and dissociative symptoms (Ar bisi, Polusny, Erbes, Thuras, & Reddy, 2011). Scores from the Externalizing domain were hypothesized to demonstrate small to moderate associations with current smoking and narcotic medica tion use, as this domain's scales measure acting out behaviors such as substance abuse. Similar findings were reported by Block, Ben-Porath, and Marek (2013) for a sample of spine surgery and spinal cord stimulator candidates. Somatic/Cognitive Dysfunction scale scores, which measure various physical complaints, were expected to have small to moderate correlations with hours resting per day, number of medications, narcotic pain medication use, duration of pain, intake pain intensity, depressive diagnosis, the PDI, physical ability tests, and the University of Alabama (UAB) Pain Behavior Scales (Richards, Nepomuceno, Riles, & Suer, 1982). Finally, the Interpersonal Functioning Scales were expected to demonstrate small to moderate associations with the PDI Family/Home, Recreation, Socialization, and Sexual disability scales, which measure pain disability particular to interpersonal situations.
MMPI-2-RF AND CHRONIC LOW BACK PAIN
Method Participants Participants included an archival sample of 875 nonconsecutive patients with chronic pain (346 males, 529 females) that completed the MMPI-2 and presented with lower back pain to a short-term multidisciplinary pain treatment program in Northeast Ohio. Ap proximately 85% of the sample completed the program. Patients who completed the program participated for an average of 21.1 days (SD = 10.0). Participants were excluded from the analyses if they produced invalid MMPI-2-RF profiles according to the test authors’ published guidelines, which included: Cannot say, CNS > Raw Score 18; Variable Response Inconsistency, VRINr > 80; True Response Inconsistency, TRIN-r > 80; Infrequent Responding, F-r = 120; and Infrequent Psychopathology Re sponses, Fp-r > 100 (Ben-Porath & Tellegen, 2008/2011). The final sample included 811 patients (318 males, 493 females) after exclusion. Consistent with standard interpretive guidelines, indi viduals demonstrating elevations on the Somatic/Cognitive over reporting validity scales were not excluded from the analyses. Overall, 6.2% of the sample produced scores >99T on Infrequent Somatic Responses (Fs) (3.5%), Symptom Validity (FBS-r) (1.8%), or the Response Bias Scale (RBS) (2.1%), indicating that they may have been overreporting somatic/cognitive problems. The majority of the sample was married (61.0%) and other marital statuses included divorced (17.9%), never married (15.9%), separated (3.1%), and widowed (2.1%). The average age was 46.7 (SD = 12.6) and the average years of education was 13.3 (SD = 3.8). The average duration of pain was 10.3 years (SD = 9.1) and the average pain intensity rating on a 0-10 scale at intake was 6.8 (SD = 1.9). Excluded individuals had significantly less education than the final sample, f(816) = 2.851, p = .004, Cohen’s d = .41, as well as significantly higher intake pain intensity ratings, t(816) = 3.427, p = .001, Cohen’s d = .45 No other significant differences were observed (ps > .150).
Measures MMPI-2-RF. As previously described in detail, the MMPI2-RF (Ben-Porath & Tellegen, 2008/2011) is a 338 item self-report broadband measure of psychopathology. It has nine Validity Scales designed to assess the test-taking approach and 42 Substan tive Scales that measure the following domains: Emotional Dys function, Thought Dysfunction, Behavioral Dysfunction, Somatic/ Cognitive Complaints, and Interpersonal Functioning. Psychometric findings presented in the test’s Technical Manual indicate adequate reliability and validity across a variety of set tings (Tellegen & Ben-Porath, 2008/2011). Pain Disability Index. The PDI (Tait et al., 1990) is a 7-item self-report measure that assesses pain-related disability in the following domains: Family/Home Responsibilities, Recreation, Social Activity, Occupation, Sexual Behavior, Self-Care, and LifeSupport Activity. It has demonstrated adequate test-retest reliabil ity and internal consistency among chronic pain patients (Tait et ah, 1990). In the current sample, the PDI items yielded a total score internal consistency of .83. Physical Ability Tests. These tests included the number of stairs climbed in 1 min, functional reach in inches, distance walked
435
in 6 min, and time to rise from seated position. These physical ability performance tests were administered at the beginning of the multidisciplinary pain treatment program and are commonly used among people with chronic low back pain (Andersson, Lin, & Smeets, 2010). Depression, Anxiety, Stress Scales. The DASS (Lovibond & Lovibond, 1996) is a 42-item self-report measure of mood prob lems. It has three scales measuring depression, anxiety, and gen eralized distress. It has demonstrated adequate internal consistency reliability in a variety of settings (Lovibond & Lovibond, 1996). UAB Pain Behavior Rating Scales. The UAB Pain Behavior Scales (Richards et ah, 1982) is a 10-item clinician rated measure that assesses the number of physical pain behaviors a patient displays. Specifically, it assesses the following behaviors: Verbal Complaints, Non-Verbal Complaints, Down-Time, Facial Gri maces, Standing Posture, Mobility, Body Language, Use of Visi ble Support Equipment, Stationary Movement, and Medication Usage. Medical Record Information. A number of variables were extracted from the patient’s medical record. These included patient demographics, diagnoses, substance use history, medications, and intake pain intensity ratings.
Procedure All chronic pain patients received a psychological evaluation at intake into the short-term chronic pain rehabilitation program. These psychological evaluations included the MMPI-2 and a va riety of self-report measures. Although the MMPI-2 was routinely administered to patients, in some cases it was not given owing to reading limitations, potential cognitive deficits, or time consider ations. In the current study, MMPI-2 items were used to calculate MMPI-2-RF scale scores, which is possible because all 338 MMPI-2-RF items are included in the MMPI-2 booklet. Past research has demonstrated the relative comparability of MMP2-RF scale scores generated from both booklets (Tellegen & Ben-Porath, 2008/2011; Van der Heijden, Egger, & Derksen, 2010). Data were collected from 1999 to 2008.
Results Descriptive Findings Means and SD for all MMPI-2-RF validity and substantive scales are presented for the full sample and by gender in Table 1. Mean differences between genders of five or more T score points (i.e., one half SD in the general population) are considered clinically meaning ful, consistent with traditional benchmarks for the interpretation of the MMPI-2 and MMPI-2-RF (Graham, 2012). Clinically meaningful gender differences were found on the following scales: Fs, FBS-r, Behavioral/Extemalizing Dysfunction (BXD), Antisocial Behavior (RC4), Gastrointestinal Complaints (GIC), Multiple Specific Fears (MSF), Juvenile Conduct Problems (JCP), Substance Abuse (SUB), Aesthetic-Literary Interests (AES), Mechanical-Physical Interests (MEC), Aggressiveness-Revised (AGGR-r), and DisconstraintRevised (DISC-r). Men scored higher than women on BXD, RC4, JCP, SUB, MEC, AGGR-r, and DISC-r, whereas women scored higher than men on Fs, FBS-r, GIC, MSF, and AES. Differences between genders were most pronounced on MEC (men scored 13T
TARESCAVAGE, SCHEMAN, AND BEN-PORATH
436
Table 1 Minnesota Multiphasic Personality Inventory-2-Restructured Form Sample Means and SDs by Gender Women (n = 493)
Men (n = 318)
Scales Validity Scales Variable Response Inconsistency (VRIN-r) True Response Inconsistency (TRIN-r) Infrequent Responses (F-r) Infrequent Psychopathology Responses (FP-r) Infrequent Somatic Responses (Fs) Symptom Validity (FBS-r) Response Bias Scale (RBS) Uncommon Virtues (L-r) Adjustment Validity (K-r) Higher-Order Emotional/Internalizing Dysfunction (EID) Thought Dysfunction (THD) Behavioral/Extemalizing Dysfunction (BXD) Restructured Clinical Demoralization (RCd) Somatic Complaints (RC1) Low Positive Emotions (RC2) Cynicism (RC3) Antisocial Behavior (RC4) Persecutory Ideation (RC6) Dysfunctional Negative Emotions (RC7) Aberrant Experiences (RC8) Hypomanic Activation (RC9) Specific Problems Malaise (MLS) Gastrointestinal Complaints (GIC) Head Pain Complaints (HPC) Neurological Complaints (NUC) Cognitive Complaints (COG) Suicidal Ideation (SUI) Helplessness (HLP) Self-Doubt (SFD) Inefficacy (NFC) StressAVorry (STW) Anxiety (AXY) Anger Proneness (ANP) Behavior Restricting Fears (BRF) Multiple Specific Fears (MSF) Juvenile Conduct Problems (JCP) Substance Abuse (SUB) Aggression (AGG) Activation (ACT) Family Problems (FML) Interpersonal Passivity (IPP) Social Avoidance (SAV) Shyness (SHY) Disaffiliativeness (DSF) Aesthetic-Literary Interests (AES) Mechanical-Physical Interests (MEC) Personality Psychopathology-5 Aggressiveness-Revised (AGGR-r) Psychoticism-Revised (PSYC-r) Disconstraint-Revised (DISC-r) Negative Emotionality/Neuroticism-Revised (NEGE-r) Introversion/Low Positive Emotionality-Revised (INTR-r)
Overall (N = 811)
M
SD
M
SD
M
SD
51 51F 70 53 62 68 64 55 48
9 9 16 11 16 13 15 11 10
52 51F 70 53 67 75 68 55 47
10 11 16 11 19 13 15 9 9
52 51F 70 53 65 72 66 55 47
10 10 16 11 18 13 15 10 9
61 51 55
12 10 11
63 54 46
11 11 9
62 53 49
12 11 10
61 69 64 49 56 53 52 52 48
12 11 13 10 11 11 11 10 10
64 73 63 48 49 54 55 54 45
11 12 12 10 9 11 11 11 8
63 72 63 49 52 54 53 53 46
11 11 12 10 11 11 11 11 9
77 61 63 70 60 57 54 57 52 55 56 54 51 46 55 54 51 46 50 49 56 51 54 42 57
8 16 9 13 14 17 14 12 11 12 15 12 9 7 13 11 11 9 11 10 11 11 14 9 10
77 67 66 71 64 56 54 59 55 57 59 53 55 54 49 48 48 49 54 53 52 51 51 47 44
8 17 10 13 14 17 12 12 12 11 14 11 12 10 10 8 8 10 12 10 11 10 11 9 6
77 64 65 71 62 57 54 58 54 56 58 53 53 51 51 50 49 48 52 51 54 51 52 45 49
8 17 10 13 14 17 13 12 12 11 14 11 11 10 12 10 10 10 12 10 11 10 12 9 10
51 52 55 56 60
10 10 11 13 13
46 54 44 58 57
8 10 7 12 12
48 53 48 57 58
9 10 10 12 12
higher than women). Both genders produced mean substantive scale clinical elevations (i.e., a score >65T) on Somatic Complaints (RC1), Malaise (MLS), and Neurological Complaints (NUC), and women additionally had a mean elevation on GIC.
Reliability Mean interitem correlations, Cronbach’s a internal consistency, and SEM estimates are provided in Table 2 for all substantive
MMPI-2-RF AND CHRONIC LOW BACK PAIN
437
Table 2 Minnesota Multiphasic Personality Inventory-2-Restructured Form Reliability Estimates in Chronic Low Back Pain (N = 811) and Normative Samples (N = 2276) Mean inter-item correlation Scale name Higher-Order Emotional/Internalizing Dysfunction (EID) Thought Dysfunction (THD) Behavioral/Extemalizing Dysfunction (BXD) Restructured Clinical Demoralization (RCd) Somatic Complaints (RC1) Low Positive Emotions (RC2) Cynicism (RC3) Antisocial Behavior (RC4) Persecutory Ideation (RC6) Dysfunctional Negative Emotions (RC7) Aberrant Experiences (RC8) Hypomanic Activation (RC9) Specific Problems Malaise (MLS) Gastrointestinal Complaints (GIC) Head Pain Complaints (HPC) Neurological Complaints (NUC) Cognitive Complaints (COG) Suicidal Ideation (SUI) Helplessness (HLP) Self-Doubt (SFD) Inefficacy (NFC) Stress/Worry (STW) Anxiety (AXY) Anger Proneness (ANP) Behavior Restricting Fears (BRF) Multiple Specific Fears (MSF) Juvenile Conduct Problems (JCP) Substance Abuse (SUB) Aggression (AGG) Activation (ACT) Family Problems (FML) Interpersonal Passivity (IPP) Social Avoidance (SAV) Shyness (SHY) Disaffiliativeness (DSF) Aesthetic-Literary Interests (AES) Mechanical-Physical Interests (MEC) Personality-Psychopathology-5 Aggressiveness-Revised (AGGR-r) Psychoticism-Revised (PSYC-r) Disconstraint-Revised (DISC-r) Negative Emotionality/Neuroticism-Revised (NEGE-r) Introversion/Low Positive Emotionality-Revised (INTR-r)
Note.
Internal consistency
SEM
LBP
NORM
LBP
NORM
LBP
NORM
Number of items
.19 .07 .14
.15 .08 .14
.90 .67 .79
.87 .69 .76
3.7 6.1 4.7
3.6 5.6 4.7
41 26 23
.28 .13 .13 .21 .13 .08 .18 .12 .10
.24 .11 .10 .21 .13 .09 .16 .11 .11
.90 .80 .71 .80 .77 .61 .84 .72 .76
.88 .76 .66 .80 .75 .65 .82 .71 .78
3.6 5.1 6.4 4.4 5.2 6.9 4.5 5.7 4.3
3.3 4.7 5.9 4.5 4.8 5.9 4.0 5.5 4.7
24 27 17 15 22 17 24 18 28
.10 .32 .22 .17 .24 .29 .24 .43 .26 .18 .20 .28 .09 .22 .29 .20 .16 .15 .20 .18 .29 .32 .18 .17 .25
.19 .27 .27 .11 .16 .16 .15 .38 .22 .16 .15 .27 .09 .25 .24 .20 .16 .15 .16 .19 .26 .31 .13 .17 .24
.47 .70 .63 .67 .76 .67 .61 .75 .76 .61 .55 .73 .46 .71 .71 .64 .63 .58 .71 .69 .80 .77 .57 .58 .75
.62 .67 .64 .55 .67 .38 .45 .70 .71 .56 .44 .72 .47 .70 .61 .62 .62 .60 .66 .70 .78 .76 .47 .55 .59
5.5 9.1 5.9 7.7 6.9 9.7 8.0 6.0 5.7 7.1 9.5 5.8 8.3 5.1 6.3 6.0 5.8 6.4 6.2 5.7 5.0 5.0 8.0 6.1 5.1
5.9 5.8 6.0 6.7 5.8 7.9 7.5 5.5 5.4 6.3 7.5 5.4 7.3 5.0 6.3 5.9 6.1 6.3 5.9 5.6 4.8 4.7 7.3 6.3 4.8
8 5 6 10 10 5 5 4 9 7 5 7 9 9 6 7 9 8 10 10 10 7 6 7 9
.13 .07 .14 .16 .16
.14 .08 .14 .14 .13
.73 .65 .77 .80 .79
.73 .69 .71 .77 .75
4.8 6.1 4.9 5.5 5.7
5.3 5.6 4.9 4.8 5.0
18 26 20 20 20
LBP = Low Back Pain Sample; NORM = Normative Sample.
scales. Also included are comparable estimates in the MMPI-2-RF normative sample (Tellegen & Ben-Porath, 2008/2011). Internal consistency estimates for the Higher-Order Scales ranged from .67 (Thought Dysfunction) to .90 (Emotional/Internalizing Dysfunction) with a median of .79. Among the Restruc tured Clinical Scales, reliability estimates ranged from .61 (Perse cutory Ideation) to .90 (Demoralization) with a median of .77. For the Specific Problems Scales, internal consistency ranged from .46 (Behavior Restricting Fears) to .80 (Social Avoidance) with a median of .67. Finally, among the Personality-Psychopatology-5 Scales, internal consistency ranged from .65 (Psychoticism-
Revised) to .80 (Negative Emotionality/Neuroticism-Revised) with a median of .77. Overall, internal consistency estimates were consistent with the normative sample, except for Malaise, which had a substantially lower estimate (a . 10) in this sample. Mean interitem correlations for the scales that provide the broadest level of measurement, the Higher-Order Scales, ranged from .07 (Thought Dysfunction) to .19 (Emotional/Internalizing Dysfunction) with a median of .14. Among the Restructured Clin ical Scale, mean interitem correlation estimates ranged from .08 (Persecutory Ideation) to .28 (Demoralization) with a median of .13. For the shortest scales, the Specific Problems Scales, mean
438
TARESCAVAGE, SCHEMAN, AND BEN-PORATH
interitem correlations ranged from .09 (Anger Proneness) to .43 (Self Doubt) with a median of .22. Finally, among the PersonalityPsychopatology-5 Scales, mean interitem correlations ranged from .07 (Psychoticism-Revised) to .16 (Negative Emotionality/ Neuroticism-Revised and Introversion/Low Positive EmotionalityRevised) with a median of .14. Overall, mean interitem correla tions were consistent with the normative sample, except for Malaise and Head Pain Complaints, which had appreciably lower (>.05) mean interitem correlations. Also provided in Table 2 are SEM for all substantive scales. For the Higher-Order Scales, SEMs (expressed in T score units) ranged from 3.7 (Emotional/Intemalizing Dysfunction) to 6.1 (Thought Dysfunction) with a median of 4.7. Among the Restructured Clin ical Scales, SEM ranged from 3.6 (Demoralization) to 6.9 (Perse cutory Ideation) with a median of 5.1. For the Specific Problems Scales, SEM ranged from 5.1 (Multiple Specific Fears) to 9.7 (Suicidal Ideation) with a median of 6.1. Finally, among the Personality-Psychopatology-5 Scales, SEM ranged from 4.8 (Aggressiveness-Revised) to 6.1 (Psychoticism-Revised) with a median of 5.5. Overall, the estimates for the substantive scales were consistent with those obtained in the normative sample, except for Gastrointestinal Complaints, Cognitive Complaints, Suicidal Ideation, and Anxiety, which had substantially higher SEM estimates ( a l T) in this sample.
Concurrent External Correlates Zero-order correlations between MMPI-2-RF substantive scales and external chart review, PDI, and DASS criteria are presented in Table 3 for the Higher-Order, Restructured Clinical, and Somatic/ Cognitive Specific Problem Scales, in Table 4 for the Internalizing and Externalizing Specific Problem Scales, and in Table 5 for the Interpersonal Specific Problem Scales, Interest Scales, and Pesonality-Psychopathology-5 Scales. We applied a family wise Bonferroni correction to reduce the risk of Type I error, which yielded a corrected a level of .001 (.05/33 comparisons for each scale). Only correlations that met this a level were interpreted. To facilitate interpretation, the findings are summarized in reference to each of the five MMPI-2-RF domains, which include: (1) Emotional Dysfunction; (2) Thought Dysfunction; (3) Behavioral Dysfunction; (4) Somatic/Cognitive Complaints; and (5) Interper sonal Functioning. The scales in the Emotional Dysfunction domain include Higher-Order Scale Emotional/Intemalizing Dysfunction (EID) and Restructured Clinical Scales Demoralization (RCd), Low Pos itive Emotions (RC2), and Dysfunctional Negative Emotions (RC7). This domain also includes the Internalizing Specific Prob lem Scales Suicidal/Death Ideation (SUI), Helplessness/Hopelessness (HLP), Self-Doubt (SFD), Inefficacy (NFC), Stress/Worry (STW), Anxiety (AXY), Anger Proneness (ANP), BehaviorRestricting Fears (BRF), and Multiple Specific Fears (MSF). Fi nally, this domain includes Personality-Psychopathology-5 Scales Negative Emotionality/Neuroticism-Revised (NEGE-r) and Low Positive Emotionality/Introversion-Revised (INTR-r). Scales in this domain demonstrated a number of convergent correlations with mental health history information. Many scales were associ ated with psychotropic medication use, including antidepressants (EID, RCd, SFD, STW, AXY, and NEGE-r) and benzodiazepines (STW, AXY, and NEGE-r), though some scales demonstrated
correlations with antipsychotic medication use as well (STW, AXY, and NEGE-r). Scales in this domain were associated with mental health diagnoses, including major depressive disorder (EID, RCd, RC2, RC7, SUI, HLP, SFD, NFC, STW, AXY, ANP, BRF, NEGE-r, and INTR-r), bipolar disorder (RC7 and AXY), anxiety disorders (STW and AXY), and PTSD (AXY). The Emo tional Dysfunction domain’s scales correlated with the PDI (EID, RCd, RC2, SUI, and INTR-r), as well as the DASS (EID, RCd, RC2, RC7, SUI, HLP, SFD, NFC, STW, AXY, ANP, BRF, and NEGE-r). Regarding the DASS correlations, most scales demon strated their strongest correlations with DASS Depression (EID, RCd, RC2, SUI, HLP, and SFD), though some scales correlated highest with DASS Anxiety (NFC, STW, AXY, and BRF) and DASS Stress (RC7, ANP, and NEGE-r). Many of the scales in this domain were associated with the number of hours resting per day (EID, RCd, RC2, SUI, HLP, SFD, and INTR-r). The MMPI-2-RFs Emotional Dysfunction’s scales were generally uncorrelated with pain and physical ability tests. The scales in the Thought Dysfunction domain include HigherOrder Scale Thought Dysfunction (THD), Restructured Clinical Scales Persecutory Ideation (RC6) and Aberrant Experiences (RC8), and Personality-Psychopathology-5 Scale PsychoticismRevised (PSYC-r). Scales in this domain were correlated with mental health diagnoses, including PTSD and bipolar disorder (THD, RC6, RC8, and PSYC-r). These scales also demonstrated associations with somatoform disorder (THD, RC8, and PSYC-r), major depressive disorder (THD, RC8, and PSYC-r), and sub stance use disorder (RC8 and PSYC-r). As expected, some of the scales correlated with the PDI self-care and life support scales (RC8 and PSYC-r). All were correlated to some extent with the DASS Scales. The MMPI-2-RFs Thought Dysfunction scales were generally uncorrelated with pain and physical ability tests. The scales in the Behavioral Dysfunction domain include Higher-Order Scale Behavioral/Externalizing Dysfunction (BXD), Restructured Clinical Scales Antisocial Behavior (RC4) and Hypomanic Activation (RC9), and the Externalizing Specific Problem Scales Juvenile Conduct Problems (JCP), Substance Abuse (SUB), Aggression (AGG), and Activation (ACT). This domain also in cludes Personality-Psychopathology-5 Scales AGGR-r and DISC-r. These scales demonstrated associations with mental health history information, particularly a history of substance use disorder (BXD, RC4, RC9, JCP, SUB, AGG, and DISC-r) and PTSD (BXD, RC4, RC9, JCP, AGG, and DISC-r). Scales in this domain were also associated with smoking (BXD, RC4, JCP, SUB, and DISC-r). Some negative associations were demonstrated with pain disability (BXD, SUB, and DISC-r) and poor physical ability test results (DISC-r). Scales in this domain were generally uncorrelated with pain, internalizing problems, and narcotic medication use. The scales in the Somatic/Cognitive Complaints domain include Restructured Clinical Scale Somatic Complaints (RC1) and the Somatic/Cognitive Specific Problem Scales, which are comprised of MLS, GIC, Head Pain Complaints (HPC), NUC, and Cognitive Complaints (COG). These scales demonstrated convergent associ ations with measures of pain, including intake pain intensity (RC1 and NUC) and the UAB (NUC). They were correlated with hours resting per day (RC1, MLS, and NUC) and the PDI scales (RC1, MLS, HPC, NUC, and COG). A number of associations were observed for medication use, including benzodiazepines (RC1 and NUC), antidepressants (RC1), and prescription narcotics (NUC).
MMPI-2-RF AND CHRONIC LOW BACK PAIN
—
i n co d 0 — 0
x 0
h 0
*
*
*
—
i n d x x — — o o
#
*
*
* *
*
o cc t
os d
m
m d - — o — —
»n o
Table 3 Minnesota Multiphasic Personality Inventory-2-Restructured Form Higher Order, Restructured Clinical, and Somatic/Cognitive Specific Problem Scale Correlations With Criteria
1
d- dl oo 0 0 0 1
*
*
I
n o N - t o c o c o n - t f - v o 0 0 - 0 0 0 0 ------------< —
C O0 -
T X O d 0 0 0 0
n ox
V, b
w *i c- *n d
t dl
t -t
On —
x dl
d - - t o o d i o o o x d — in o o o — — o o o o o
( OCOX o — o
— C 4 — — o —
T t ( N O O O X C 4 X
O OO N in — dl —
^t 0
1
0
0
o
1
0
0
0
— 0
—
*
— — O O O O
*
*
o x t N r - i n x r ' — o o o o o o
— —
I
' t M J ' N O ' O l N O i O b l O O O O — O O — O dl
d d m — C MC N
dl 0
— co o o o
x o
h d in — — o o
in o
d ( Nd l
d- os d n * - co — — — dl —
X n
o — cn n
do
1
O oo — o o o
1
-
- t on 0 0 0
T O
1
x co dl ■'t 0 0 0 0
0
1
cl T T — — O O
1
X di — o
I I
o
d o
d T — —
Xoo 0 0
1
*- r i r, — O O O O
T
X
g
X
IQ | T
d- d- d— — dl
— X d l d ~ X O 0 0 0 0 0 0
1
1
1
t o x o o x m x — o o o o o — —
1
N Od d d
1
d0 1
On ci
dl 0 1
— d i o x o N O c n o m r 0 — 0 0 0 — 0 — 0 0
O i n On d — C H O O
1
in rf — o o
1
I
Tt o
oo o o
I I
I
1
- - x o o
— t o o
I
t d X o o o
T
——
1
com — o o o
— t o o
cn — i n O' 0
1
oo i n o 0 0 0
Vi I n d
o
t o
1
d o
— OnX dl — dl
t o
o
1
x d X d * x Tj- o o o
I
I
d in — —
m ■*?
x o
1
1
-
1
— d 0 —
x 0
d o
0 1
0 1
t 0
1
d cn i n 0 0 0
1
©
1
1
i n - i n - ' t o d o o o o o — —
d d
o
o d
X 0
O' d - 0 0
OnX 0 0
x o i n V
V
o d i d -
d x cn —
in c n
cs o
X o
O'
—
O Nd o —
O o
d
t
d
d
t
—
x 0
1
oo — — —
ro
— —
d i n d ’- h ,--------- (N(N
d d
O in
X d
o
In
—
—
— Os ' t cn o o
O
—
X o
*
*
*
— in 0 0
X t d O O O 1
1
1
X
1
d
s
1
d l dl
1
1
1
1
* cn -t
1
c o in — — c i o\ o o\ in o o ——o o o o o o o
C)
* o 1 1
*
x
iq
g
* o
oo
i n o N X d d O N d — o o — — o — —
nq n
w
i— n on d 0o 0 0 1 1 1 1
»n t 0 -
-
in 0 0
in o >n t
md
— 0 —
dt
xcc x t O O O
I I
—d ' i n x x n c —o o o o o
1
cn
O
d
n c n
——
c i i n d f c n c i f c c c f c i n
o o - o x — dl — O — O
d
di
m
—x
d
d
d
o
X
d
d
O N
*-
—O
oo
d-
dl 1
C N 00
- d - r ^ r - ' d - d - x x
N Od
r-
—O
i > d-’ d^
O '-
O N im n in O d in x
d*'
t—X
X
X
X
X
X
X
>n in
o
di
cs
di
-
0
1
—C Nto O S d l — >n to 0 —0 0 —0 — 1 1
1
1
1
oo d
O
On t
d
1
1
dl OO 0 0 0
O N
1
1
' O ' O O c i T d i c i m o T f o m o o o o o o o o
On
*
1
1
1
1
1
*
—d
3 * 2
1
1
1
x
8
cn o
1
- C ' l O T t C C O m T O ' t C — 0 — 0 — 0 0 — 0 —
1
1
1
1
x r - cn o O O O O
t d X O O O
1
t d
1
i n — oo o o o
8
1
I
- C l i f M O l C h O ' O T C O 0 — o o o o o o o — 1
o
1
o c - C ' O — TfN — n o O c d O O O O O O O O
re ot m c s o o
i n r o o o o
1
i n o i n c K N T f t n o \ m 0 C 4 0 0 0 0 0 0 0 0
o o
d O' — O
1
1
11
1
1
1
1
1
1
1
dl d o o
I
1
-t 0
1
1
* 0
1
i n oo o —
1
0
1
*
1
d
O
1
O *n
1
1
t
1
O No
■'t- in
00 00
u u u y < Q >n C U- O c L s? 2 C N ^ y X •c G -o ‘•S && •G • s-1 “ *Q . y C C 0) 3 W) o P" ^a..*° .c >s
~
S'? x ) . S G
O
O
•S n rt
on on
