PERSISTENT PAIN is a prevalent problem in a large proportion

Transcription

1 395 Assessment of the Impact of Pain and Impairments Associated With Spinal Cord Injuries Eva G. Widerström-Noga, DDS, PhD, Robert Duncan, PhD, Ernesto Felipe-Cuervo, MS, MEd, Dennis C. Turk, PhD ABSTRACT. Widerström-Noga EG, Duncan R, Felipe- Cuervo E, Turk DC. Assessment of the impact of pain and impairments associated with spinal cord injuries. Arch Phys Med Rehabil 2002;83: Objectives: To determine the adequacy of the Multidimensional Pain Inventory (MPI) for assessing pain impact after spinal cord injury (SCI) and to determine whether the impact of pain can be separated from other consequences of SCI. Design: Postal survey. Setting: General community. Participants: Of the 159 subjects contacted who experienced chronic pain, 120 (75.5%) participated. Interventions: Subjects were mailed the original MPI and a set of additional items specific to SCI. Main Outcome Measure: The MPI. Results: Confirmatory (CFA) and exploratory factor analyses were performed for each section of the MPI. Elimination of several items, including those related to work in section 1 (pain impact), improved the goodness-of-fit index (GFI). A CFA for section 2 (response of significant other) resulted in acceptable GFI after 2 items were deleted. Decrease in activity levels (section 3) because of other consequences of injury was significantly greater after tetraplegia than after paraplegia. In contrast, pain-related reduction in activities was not associated with injury level. Although other consequences of SCI may have greater impact on activities than pain, severe pain is likely to affect activity levels significantly. Conclusion: The MPI appears to be appropriate for use in a SCI population when modified to eliminate questions related to work and to supplement the activity scale with items addressing decreased activity levels due to pain. Key Words: Disabled persons; Pain; Pain measurement; Paraplegia; Rehabilitation; Spinal cord injuries; Tetraplegia by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation From the Miami Project to Cure Paralysis (Widerström-Noga, Felipe-Cuervo), Department of Neurological Surgery (Widerström-Noga), Department of Epidemiology and Public Health (Duncan), University of Miami, Miami, FL; and Department of Anesthesiology, University of Washington School of Medicine, Seattle, WA (Turk). Accepted in revised form March 20, Supported by The Miami Project to Cure Paralysis, State of Florida, Hollfelder Foundation, and Gordon Family Foundation, and by the National Institute of Child Health and Human Development, National Center for Rehabilitation Research (grant no. HD33989). Presented as a poster at the American Pain Society s meeting in November 4, 2000, Atlanta, GA. No commercial party having a direct financial interest in the results of the research supporting this article has/will confer a benefit on the author(s) or on any organization with which the author(s) is/are associated. Reprint requests to Eva G. Widerström-Noga, DDS, PhD, The Miami Project to Cure Paralysis, University of Miami, School of Medicine, PO Box , R-48, Miami, FL 33101, ewiderst@miamiproject.med.miami.edu /02/ $35.00/0 doi: /apmr PERSISTENT PAIN is a prevalent problem in a large proportion of community-residing people with spinal cord injury (SCI). 1,2 The ability to adapt to limitations resulting from the injury is significantly related to quality of life (QOL). 3 The ramifications of pain after SCI are compounded by the fact that so many medical consequences, some directly related to the extent of the injury to the nervous system, are associated with SCI. 4 Furthermore, the refractory nature of the painful conditions after SCI 5 suggests that personal characteristics related to adaptation and coping skills are crucial for improving QOL. 6 Chronic pain, regardless of origin, affects people differently. A person s response to it depends on a variety of psychologic and behavioral factors. 7,8 Thus, a treatment intervention should be tailored not only to specific features of the pain but also to each person s psychologic characteristics and life circumstances to increase the likelihood for a positive outcome. 5,9,10 The diagnosis and treatment strategy should rely on an assessment procedure, reflecting both physical and psychosocial factors, as well as the pathophysiologic mechanisms, involved in pain Although a number of assessment instruments have been developed for chronic pain patients in general, 14,15 they may not be appropriate for use with SCI patients because of the associated limitations of other consequences of the injury that may influence the perception of pain and subsequent responses in different ways. 2 In addition, SCI is a general classification, and there may be many differences in the impact, depending on factors such as the level of the lesion (eg, paraplegic vs tetraplegic) and pain mechanisms (eg, neuropathic, nociceptive). 12,13,16 Research is needed to determine whether measures developed for other chronic pain populations can be generalized to people with SCI and its subtypes, based on level of lesion, and to determine what modifications, if any, are required. The West Haven Yale Multidimensional Pain Inventory 17 (MPI) is a comprehensive instrument designed to assess a range of self-reported behavioral and psychosocial factors associated with chronic pain syndromes. It has been used in various pain populations (eg, back pain, headache, temporomandibular disorders, fibromyalgia, cancer) The purposes of this study were (1) to test the adequacy of the MPI for use with an SCI population, (2) to determine the degree to which pain decreases activity compared with other aspects of injury, (3) to define the relationships among these factors, and (4) to compare psychosocial and behavioral responses of persons with paraplegia to those with tetraplegia. METHODS The subjects of the present study were recruited from The Miami Project to Cure Paralysis database. The study was approved by the University of Miami s institutional review board. A total of 258 persons with traumatic SCI who were more than 18 years of age and who had reported chronic pain or nonpainful sensations in a previous study, 16 were mailed the MPI and a structured questionnaire that asked about their pain

2 396 IMPACT OF PAIN AFTER SCI, Widerström-Noga and nonpainful sensations (part of another study). We received 136 responses. We attempted to contact subjects who did not respond to the initial mailing by telephone, and another 8 agreed to participate. We determined that 75 subjects had invalid telephone numbers or had moved, leaving no forwarding address or telephone number. Of the 184 subjects we were able to contact, 24 indicated that they did not experience chronic pain but had only nonpainful sensations. Thus, we located 159 subjects who experienced chronic pain, and 120 of these (75.5%) agreed to participate in the study. Multidimensional Pain Inventory The MPI 17 is a 60-item (56 scored), self-report questionnaire based on the cognitive-behavioral perspective 21 designed to assess the impact of and adaptation to chronic pain. It comprises 3 sections: section 1 (pain impact), section 2 (responses by significant others), and section 3 (common activities), and has been factor-analyzed to include 12 scales. Eight of the scales measure cognitive, affective, social, and behavioral responses: pain severity, life interference, life control, affective distress, support, negative responses from others, solicitous responses from others, and distracting responses from others. The remaining 4 scales assess the degree of participation in various types of daily activities: household activities, activities away from home, social activities, and outdoor activities. These 4 scales are commonly used to create a single general activity scale. 17 The MPI has good psychometric properties, 17,22 the factor structure has been confirmed in several studies, it has validity as an assessment tool for a number of chronic pain patients and pain conditions, 19 and it is sensitive to change. 26 Supplemental Questions for Section 3 of the MPI To define the degree to which pain and other consequences of injury reduced participation in a specific activity, we added 2 items after each item in section 3: (1) Pain has reduced my participation in this activity, and (2) Other consequences of SCI have reduced my participation in this activity. The response options ranged from 0 (not at all) to 6 (extremely). The items were reviewed with 10 people with chronic pain and SCI to ensure that the questions were properly worded and interpreted. All 10 reported a clear understanding of the questions, and no modifications in the wording were required. Sociodemographic Data, Characteristics of Injury, and Location of Pain Demographic information and injury characteristics obtained from The Miami Project database included age at time of the study, age at time of injury, time postinjury, gender, and level of injury (cervical, below cervical). If more than 1 level of injury was reported, the highest level was used in the analyses. Statistical Methods When there were missing data, the mean value for each scale was prorated based on the sum scores for the number of items completed for each scale divided by the number of completed items for each scale (ie, if 1 question of a 6-item subscale was incomplete, the sum of the scores for the 5 completed items would be divided by 5 instead of by 6 to obtain the mean for the scale). Confirmatory factor analyses (CFAs) are performed to test how well a theoretic factor structure is substantiated with a different sample. The fit is expressed in goodness-of-fit indices (GFIs) based on variances and covariances in the data set. 27 Even though a low GFI suggests that the proposed factor structure may be inadequate, Hoelter 28 has emphasized that the most important question to be asked is how well a model approximates the data, rather than whether it fits the data. That is, do the model parameter estimates describe the data, not whether these estimates are significant. The following 2 indices were applied: the GFI, and the nonnormed fit index (NNFI). The GFI is usually a value between 0 and 1 29 and has been reported to be more accurate than other stand-alone indices. 27 The NNFI has been used to compare samples with unequal numbers, is minimally affected by sample size, 27 and is equivalent to the Tucker Lewis 30 index. When the GFI falls below 0.9, this suggests that the model often can be significantly improved, and exploratory factor analyses (EFAs) can be performed to identify an improved factor structure. The adequacy of a new structure can then be reevaluated with CFA. We performed separate CFAs for each section of the MPI to estimate how well the SCI data sample approximated the original factor structure reported by Kerns et al. 17 The analysis PROC CALIS 31 (using SYSTAT a ) was based on covariance matrices from the observed items in the original factor structure for each section. Decisions regarding retention of items resulting from the EFA were based on criteria used in the original MPI study, 17 namely, that an item had to have a maximum loading of at least.35 on 1 factor and load with a difference greater than.15 on any other factor. We used the principal component analysis 32 as the method for factor extraction in the EFAs. The analyses were based on correlation matrices and pairwise relationships. Only factors with eigenvalues greater than 1.0 were accepted. The factor loadings were sorted in descending order and the oblique rotation method, Oblimin, was used because correlations between factors were expected. For pairwise comparisons and correlations, we used the Student t test and Pearson correlation. All tests were 2 tailed, and we used Bonferroni correction to adjust for multiple comparisons. 32 RESULTS Participants Demographic variables (age, age at injury, gender, injury level) were compared within the group of 258 subjects who were sent the survey (table 1). These variables were calculated and compared by using analysis of variance and the Bonferroni post hoc test for 4 subgroups: (1) subjects with pain (n 120), (2) subjects with no pain (n 24), (3) subjects who did not want to participate (n 39), and (4) subjects who could not be reached (n 75). There were no significant differences between the 4 subgroups with respect to gender or level of injury. There were also no significant differences on any of the measures between the pain sample and the 2 groups of subjects who did not respond or could not be reached. However, when comparing subjects who had pain with subjects who did not, those with pain were on average 7.6 years (P.05) older and were 6.8 years (P.05) older when injured. Even though most subjects commonly reported pain in several areas of the body, the most common locations were back (63.3%), buttocks (56.7%), and legs and feet areas (54.2%). The percentages reported for the pain group were almost identical to what were reported in a preceding study that included a larger sample. 16 The demographic data of the 120 subjects who completed the MPI were also compared with data reported in the national database. 33 For example, average age at injury was 30.7 years, which is similar to our subjects 32.4 years. The national database had a slightly higher percentage of men (82.2%); however, this percentage is higher than the average model system representation of 4:1 male-to-female ratio. Also, with

3 IMPACT OF PAIN AFTER SCI, Widerström-Noga 397 Table 1: Comparison of Demographic Variables of Sample Sent the Questionnaire (N 258) Pain (n 120) No Pain (n 24) Withdrew (n 39) Not Reached (n 75) Age, y (mean SD)* Age at injury, y (mean SD)* Time since injury, y (mean SD) Gender, n (%) Men 94 (78.3) 16 (66.7) 28 (71.8) 53 (70.7) Women 26 (21.7) 8 (33.3) 11 (28.2) 22 (29.3) Injury level, n (%) Cervical 62 (51.7) 16 (66.7) 23 (59.0) 40 (53.3) Below cervical 58 (48.3) 7 (29.2) 16 (41.0) 35 (46.7) Not reported 1 (4.2) * Difference between persons with pain (n 120) and no pain (n 24) (P.05) (analysis of variance, Bonferroni-adjusted probability). respect to level of injury, our sample is similar to the 52.9% of people with tetraplegia reported from the national database. In conclusion, our sample appears to adequately represent an average SCI population. MPI Section 1: Pain Impact The GFI for the CFA, including all items, indicated that the degree of approximation of the original factor structure could be further improved (GFI.65, NNFI.69). Because many people with SCI do not work, for reasons that may be unrelated to pain, we excluded the 2 items related to employment: Since your pain began, how much has your pain changed your ability to work? and How much has your pain changed the amount of satisfaction or enjoyment you get from work? The exclusion of these items resulted in an improvement of the fit indices (ie, GFI.70, NNFI.77). To improve the fit further, we performed an EFA that included the remaining 22 items, followed by an Oblimin rotation. 32 In this analysis, In general, how much does your pain interfere with your day-to-day activities? loaded highly on life interference (.51) and on pain severity (.51), indicating that day-to-day interference is intimately linked to the overall perception of severity of pain in this population; thus, the question was excluded. In addition, During the past week, how successful were you in coping with stressful situations of your life? loaded on life control (.43) and on affective distress (.54). Thus, it appears that unsuccessfully coping with stress is strongly associated with negative mood. According to principles for item retention, 17 we deleted this item. A second EFA was performed; it included the remaining 20 items (table 2). The 5 factors (pain severity, life interference, life control, affective distress, support) accounted for 71.4% of the total Table 2: Exploratory Factor Analysis of MPI Section 1: Impact of Pain Scales and Items Life Interference How much do you limit your activities in order to keep your pain from getting worse?.93 How much has your pain changed your ability to take part in recreational and other social activities?.91 How much has your pain changed the amount of satisfaction or enjoyment you get from family-related activities?.85 How much has your pain interfered with your ability to plan activities?.75 How much has your pain changed the amount of satisfaction or enjoyment you get from taking part in social and recreational activities?.69 How much has your pain changed or interfered your friendships with people other than your family?.62 How much has your pain changed your ability to do household chores?.59 How much has your pain changed your relationship with your spouse, family, or significant other?.44 Support How supportive or helpful is your spouse (significant other) to you in relation to your pain?.91 How attentive is your spouse (significant other) to you because of your pain?.82 How worried is your spouse (significant other) about you because of your pain?.61 Life Control How much control do you feel that you have over your pain?.84 During the past week how much control do you feel that you have had over your life?.72 During the past week how much do you feel that you have been able to deal with your problems?.53 Pain Severity On the average, how severe has your pain been during the last week?.81 Rate the level of your pain at the present moment..80 How much suffering do you experience because of your pain?.71 Affective Distress During the past week how tense or anxious have you been?.92 During the past week how irritable have you been?.87 Rate your overall mood during the past week..51 Factor Loadings

4 398 IMPACT OF PAIN AFTER SCI, Widerström-Noga variation in the data. No intercorrelations between factors were greater than Cronbach, indicating that each factor assesses a distinctive domain. We found the internal consistency (Cronbach ) for the 20 items to be acceptable: (1) pain severity (.88), (2) life interference (.91), (3) life control (.62), (4) affective distress (.75), and (5) support (.76). When testing these items in a CFA, we found the fit indices improved the GFI (.73) and NNFI (.79). Despite moderately low fit indices, the model could not be further improved by deleting items without compromising the internal consistency of the subscales. However, the replication of the factor structure in the EFA suggests that no alternative factor structure with better fit exists in this set of data. MPI Section 2: Responses by Significant Others A CFA that included all 14 items resulted in a GFI (.88) and an NNFI (.90), indicating that the degree of approximation of the original factor structure might be further improved. An EFA revealed that Ignores me loaded on both negative responses (.33) and distracting responses (.43); it was deleted. Elimination of this item resulted in an improvement in Cronbach from.84 to.95 for negative responses. The next CFA with this item deleted resulted in a modest improvement in the fit indices: GFI (.89) and NNFI (.92). An EFA including the 13 items showed that Turns on the TV to take my mind off my pain loaded on both solicitous responses (.42) and distracting responses (.40); thus, we decided to exclude it and to perform another CFA. The result was a further improvement in the GFI (.90) and the NNFI (.93). The factor loadings for the 12-item version are displayed in table 3. Even though the indices indicated a good fit, 2 items correlated more highly with factors other than those originally hypothesized. One was Tries to get me to rest, which originally correlated with the solicitous responses. However, in our analysis, the factor loading on solicitous responses was only.21, whereas it was.57 on distracting responses, indicating that this study sample viewed this item as a distracting action rather a solicitous one. The other item that loaded on another scale was Reads to me, which loaded only.19 on distracting responses (the originally hypothesized structure) and.42 on solicitous responses. This suggests that when a significant other reads to a person with SCI, the action is regarded as being more Table 3: Exploratory Factor Analysis of MPI Section 2: Responses by Significant Others Scales and Items Factor Loadings Distracting Responses Encourages me to work on a hobby..88 Tries to involve me in some activity..85 Talks to me about something else to take my mind off the pain..65 Tries to get me to rest..57 Negative Responses Gets frustrated with me..96 Gets irritated with me..95 Gets angry with me..94 Solicitous Responses Gets me pain medication..76 Gets me something to eat or drink..75 Takes over my jobs or duties..64 Asks me what he/she can do to help..50 Reads to me..42 Table 4: Exploratory Factor Analysis of MPI Section 3: Activities Scales and Items Factor Loadings Household Activities How often do you wash dishes?.92 How often do you do the laundry?.86 How often do you prepare a meal?.86 How often do you help with the house cleaning?.81 How often do you go grocery shopping?.68 Activities Away from Home How often do you go out to eat?.77 How often do you take a trip?.66 How often do you go to a movie?.66 How often do you take a ride in a car or bus?.66 How often do you go to a park or beach?.44 Social Activities How often do you play cards or other games?.64 How often do you wash the car?.63 How often do you work on the car?.62 How often do you visit friends?.57 How often do you visit relatives?.49 Outdoor work How often do you work on a needed household repair?.74 How often do you mow the lawn?.71 How often do you work in the garden?.50 solicitous than distracting. The 3 factors (solicitous responses, distracting responses, negative responses) accounted for 61.4% of the total variation of the data. The distracting responses factor was correlated with solicitous responses (r.42). No other intercorrelations among the factors exceeded.03. The Cronbach values for the factors were.95 (negative responses),.64 (solicitous responses), and.76 (distracting responses). MPI Section 3: Activities The GFI for the CFA, including all items, was GFI.80 and NNFI.81, indicating that the degree of approximation of the original factor structure could be significantly improved. Thus, an exploratory analysis that included all 18 items was performed. Based on the criteria used by Kerns et al, 17 all items were retained but the resulting factor structure was slightly different (table 4). Household activities included all items in the original factor structure, with an internal consistency of.90. Activities away from home also contained all 4 items originally hypothesized, with the exception of How often do you go to a park or beach?, which loaded.44 on this factor, and not at all on social activities, where it was hypothesized to load. Cronbach including the 5 items was.72. Thus, this item was regarded as a household activity rather than a social activity. The social activities factor contained 3 of the items in the Table 5: Student t Tests Comparing Decrease in Level of Activity Because of Pain Versus Other Consequences of Injury Pain Other Consequences Household activities * Activities away from home * Social activities * Outdoor work * * P.001 (Bonferroni-adjusted significance level).

5 IMPACT OF PAIN AFTER SCI, Widerström-Noga 399 Table 6: Correlation Matrix (Pearson pairwise comparison) Showing Correlation Coefficients and Significance Levels (Bonferroni adjusted) for Pain Severity, Level of Participation in Household Activities, Degree of Reduction in Participation in Household Activities Because of Pain, and Other Consequences of SCI PS HA RHAP RHAC PS HA.183 (NS) RHAP (NS) RHAC.292* Abbreviations: PS, pain severity; HA, household activities; RHAP, reduction of household activities because of pain; RHAC, reduction of household activities because of other consequences; NS, not significant. * P.01. P.001 (Bonferroni-adjusted significance levels). Table 8: Correlation Matrix (Pearson pairwise comparison) Showing Correlation Coefficients and Significance Levels (Bonferroni adjusted) for Pain Severity, Level of Participation in Social Activities, Degree of Reduction in Participation in Social Activities Because of Pain, and Other Consequences of SCI PS SA RSAP RSAC PS SA.105 (NS) RSAP (NS) RSAC.283* Abbreviations: SA, social activities; RSAP, reduction of social activities because of pain; RSAC, reduction of social activities because of other consequences. * P.05. P.01. P.001 (Bonferroni-adjusted significance levels). original factor structure and 2 additional items, How often do you work on the car? and How often do you wash the car? that were originally in outdoor work. The former item loaded.62 on social activities and.45 on outdoor work, whereas the latter loaded.62 on social activities and only.22 on outdoor work. In this group of people, automobile-related work was regarded more as a social activity than outdoor work. The internal consistency including these 5 items was.70. The last factor, outdoor work, included only 3 of the original 5 items. Cronbach was.65 when these 3 items were included. The 4 factors (household activities, activities away from home, social activities, outdoor work) accounted for 60.1% of the variation of the data, and intercorrelations were not statistically significant. Relation Between Level of Activity and Reduction of Activity Because of Other Consequences of Injury and Pain We compared the average scores on each of the 4 subscales (household activities, activities away from home, social activities, outdoor work) for pain and other consequences of SCI by using t tests. The ratings of perceived decrease in activity levels were significantly higher for other consequences of SCI than for pain (table 5). Because other consequences of SCI on average decreases activity more than does pain in the SCI population, relationships among pain severity, level of activity, reduction of activities because of pain (reduction in pain), and reduction of activities because of other consequences of SCI (reduction in consequences) were explored by using Pearson correlations. The analysis was performed for each of the 4 factors: household activities (table 6), activities away from home (table 7), social activities (table 8), and outdoor work (table 9). The relationships among the variables displayed similar patterns for each of the factors. For example, the level of activity was always significantly and inversely related to reduction in consequences (ie, household activities, r.45, P.001; activities away from home, r.37, P.001; social activities, r.30, P.01; outdoor work, r.30, P.01). In contrast, the relationships between level of activity and reduction in pain were weak and not statistically significant, with the exception of activities away from home, in which the correlation (r.28) was significant (P.05). Reduction in activities because of pain and reduction in activities because of other consequences of SCI were also significantly correlated, indicating that a person whose pain causes decreased activity is likely to also have decreased activity resulting from the other consequences. To address the relationship between pain severity and activity levels and reduction in activities, the pain severity score was included in the correlation matrices. We found that pain severity did not significantly correlate with activity levels. However, pain severity did significantly correlate with reduction in pain (ie, household activities, r.54, P.00; activities away from home, r.55, P.001; social activities, r.56, P.001; outdoor work, r.56, P.001). Although the relationships between pain severity and reduction in consequences were weak, they were significant, indicating that severe pain may also affect the degree to which other consequences of injury decrease activity. Differences Between Persons With Tetraplegia and Paraplegia There were no significant differences between persons with tetraplegia and paraplegia in their responses to sections 1 or 2 (including the pain severity scale) of the MPI. However, as expected, t tests comparing level of injury and average degree of participation in daily activities revealed a significant difference between cervical and below cervical injuries (table 10). Table 7: Correlation Matrix (Pearson pairwise comparison) Showing Correlation Coefficients and Significance Levels (Bonferroni adjusted) for Pain Severity, Level of Participation in Activities Away from Home, Degree of Reduction in Participation in Activities Away From Home Because of Pain, and Other Consequences of SCI PS AH RAHP RAHC Pain severity AH.239 (NS) Reduction of AH because of pain (RAHP) * Reduction of AH because of other consequences (RAHC) Abbreviation: AH, activities away from home. * P.05. P.001 (Bonferroni-adjusted significance levels).

6 400 IMPACT OF PAIN AFTER SCI, Widerström-Noga Table 9: Correlation Matrix (Pearson pairwise comparison) Showing Correlation Coefficients and Significance Levels (Bonferroni adjusted) for Pain Severity, Level of Participation in Outdoor Work, Degree of Reduction in Participation in Outdoor Work Because of Pain, and Other Consequences of SCI PS OW ROWP ROWC Pain severity OW.113 (NS) ROWP (NS) ROWC.265* Abbreviations: OW, outdoor work; ROWP, reduction of outdoor work because of pain; ROWC, reduction of outdoor work because of other consequences. * P.05 P.01. P.001 (Bonferroni-adjusted significance levels). There was also a significant difference (P.05) in decreased activity because of other consequences of SCI between the 2 levels of injury, with more decrease in activity reported by cervically injured individuals ( ) than by persons with injuries below the cervical level ( ). In contrast, no significant differences between the different levels of injury were observed in pain-related reduction of participation in activities. DISCUSSION Despite some discrepancies, reflected in moderate fit indices between the original factor structure 17 and the data collected from our sample, a somewhat revised version of the MPI appears to be useful for evaluating pain and its impact on how people with SCI adapt. The revised version of the MPI has several items deleted in sections 1 and 2, whereas section 3 contains all of the original items plus 1 question per item specifically addressing the extent to which pain reduces activity. Appendix 1 presents the MPI as it was revised after our analyses of study results. More than 50% of people with SCI are unemployed after 1 year, as reported from the Spinal Cord Injury Model Systems database. 34 In the present study, more than one third of persons with SCI were not working for unspecified reasons. Therefore, 2 items in the first section of the original MPI concerning work interference and work satisfaction do not appear to be as relevant for people with SCIs. Removal of these, and 2 cross-loading items, improved the GIF. The subsequent EFA replicated the original factor structure. The factor loadings in the 5 factors in section 1 were relatively high, with the factor structure accounting for more than 70% of the total variation. Thus, although statistically the data did not perfectly fit the original MPI factor structure, the cognitive-behavioral theory on which the 5 factors comprising section 1 was based appears to be valid for people with SCI. The internal consistency values for section 1 were similar to those observed in other pain samples. 23 Life control had the lowest value, suggesting that the reason for the lower fit in our SCI sample may be related to the issue of life control as it is associated with pain. Even though internal locus of control and pain severity after SCI are significantly related, 35 control in life is related to other consequences of the injury, as well as to chronic pain. 36 Although chronic pain can negatively affect a person s ability to cope with the consequences of SCI 37 and negative coping after SCI is associated with depression, 3 a number of factors may mediate the association between affective distress and chronic pain. Rudy et al 38 and Turk et al 39 reported that perceived interference of pain on ability to function mediated the pain-depression association. For people with SCI, activities of daily living are affected, not only by pain but also by other consequences resulting from SCI. Thus, affective distress may not be directly related to chronic pain but rather to how successfully a person is able to cope with the impact of SCI. Because it is difficult to determine why people with chronic pain and SCI decrease their activities, we addressed this issue by including 2 questions for each activity. The questions addressed the extent to which pain and other consequences of SCI, respectively, reduced participation in a particular activity. The results showed that the activity levels, as measured by the MPI, were significantly more affected by other consequences of SCI than by pain. The strong relation between decreased activity from other consequences and from pain suggests that the degree to which consequences of SCI, including pain, limit activities reflects a person s overall coping capability. Similar to the relationships observed in other studies of heterogenous pain patients 19 and after SCI, 1,40 the more severe that pain is perceived to be, the greater the probability that it will negatively affect level of activity. The significant relation between pain severity and decreased activity caused by consequences of injury other than pain parallels the observation that perceived difficulty in dealing with pain is often related to difficulty in dealing also with other aspects of SCI, such as sadness, spasticity, and unusual sensation. 2 Because activity levels overall were more affected by consequences of SCI, rather than by just pain, we explored the relationships between activity levels and level of injury (ie, tetraplegia, paraplegia). As predicted, activity levels were significantly higher in persons with paraplegia than in persons with tetraplegia. Those with tetraplegia also reported significantly less participation in various activities because of other consequences of injury. In contrast, when comparing persons with tetraplegia and those with paraplegia, no significant differences were found with regard to decreased activities from chronic pain. Thus, as seen in other chronic pain syndromes (eg, fibromyalgia 18 ) and in SCI, 40 decreased levels of participation in activities as a result of pain (ie, disability) are more related to the severity of pain rather than level of injury (ie, impairment). The analysis of the adequacy of using the MPI with persons with chronic pain after SCI shows that, although there were similarities in the responses to the MPI between the specific SCI and heterogenous pain populations, several aspects were unique to SCI. In our analyses, we used 2 different indices to Table 10: Student t Tests Comparing Average Activity Score, Reduction of Activities Because of Pain, and Reduction of Activities Because of Other Consequences of Pain for Persons With Tetraplegia Versus Paraplegia Tetrapegia Paraplegia Average activity score Decrease in activities because of pain (NS) Decrease in activities because of other consequences of SCI * * P.05. P.001 (Bonferroni-adjusted significance levels).

7 IMPACT OF PAIN AFTER SCI, Widerström-Noga 401 estimate fit (ie, GFI, NNFI). The GFI was used because it is such a common measure, which can be compared with the multiple correlation coefficient. Because the GFI may be affected by sample size, we also used the NNFI, which is generally independent of sample size. In our analyses, the 2 indices were similar in supporting the assumption that the model adequately describes the data. Thus, even though the factor structure and, consequently, the theoretic foundation were similar after SCI, it is evident that other consequences of SCI beyond pain have a significant impact on a person s life. CONCLUSION The MPI, as modified, appears to be a reasonable instrument with which to evaluate the impact of pain in people with SCI and the response of their significant others to pain. However, because activity levels are different, depending on whether a person has tetraplegia or paraplegia, and because other consequences of SCI decrease activity levels more than pain, the frequency with which a person participates in a particular activity may not necessarily reflect impairment from pain itself. Thus, because reduction in activity levels because of pain was independent of injury level, this measure appears to reflect how pain affects activity levels without being confounded by other consequences that are dependent on level of injury. Future research should evaluate whether the MPI would be improved by including questions that specifically address chronic pain and by avoiding general questions that may apply to other consequences of SCI. Additional research is needed to address the reproducibility and validity of the MPI in the revised format. As with other types of chronic pain, physical pathology alone is insufficient to determine how patients will respond to treatment. Although it is important to determine the type of pain in developing medical treatment strategy, evaluation of the relevant psychosocial factors is critical to a comprehensive approach and a successful treatment outcome. 41 For example, the ability to cope with and adapt to SCI including chronic pain is likely to be relevant for long-term therapeutic success, especially when pain after SCI may respond particularly poorly to treatment. 5 Furthermore, assessing the person and not just the pain 41 (ie, evaluating pain from a broader perspective) provides a basis for an individually tailored therapeutic approach. 9 Acknowledgments: The authors thank Alberto Martinez-Arizala, MD, for assistance in the initial phase of this project. APPENDIX 1: MULTIDIMENSIONAL PAIN INVENTORY (SCI VERSION) Today s Date: ID: When did your pain first start? Month: Year: Instructions An important part of our evaluation includes examination of pain from your perspective because you know your pain better than anyone else. The following questions are designed to help us learn more about your pain and how it affects your life. The questionnaire has three sections. Under each question is a scale to mark your answer. Read each question carefully and then circle a number on the scale under that question to indicate how that specific question applies to you. If there is a question that you think does not apply to you, please circle the number of that question. After you have completed the questionnaire, check your responses to make sure that you have answered each question. Please use the last page to add any additional information or comments that you think would be of help to us in better understanding your pain problem. Before you begin, please answer the two pre-evaluation questions below: 1. Some of the questions in this questionnaire refer to your significant other. A significant other is the person with whom you feel closest. This includes anyone that you relate to on a regular or frequent basis. It is very important that you identify someone as your significant other. Please indicate below who your significant other is (please check only one): Spouse Partner/Companion Housemate/Roommate Friend Neighbor Parent, Child or Other relative Other: 2. Do you currently live with this person? Yes No When you answer questions on the following pages about your significant other, always respond in reference to the specific person you just indicated. Section 1 This part asks questions to help us learn more about your pain and how it affects your life. Under each question is a scale to mark your answer. Read each question carefully and then circle a number on the scale under that question to indicate how that specific question applies to you. The following example may help you to better understand how you should answer these questions. Example How nervous are you when you ride in a car when the traffic is heavy? Not at all nervous Extremely nervous If you are not at all nervous when riding in a car in heavy traffic, you would want to circle the number 0. If you are very nervous when riding in a car in heavy traffic, you would then circle the number 6. Lower numbers would be used for less nervousness, and higher numbers for more nervousness.

8 402 IMPACT OF PAIN AFTER SCI, Widerström-Noga 1. Rate the level of your pain at the present moment. No pain Very intense pain 2. How much has your pain changed the amount of satisfaction or enjoyment you get from taking part in social and recreational activities? 3. How supportive or helpful is your spouse (significant other) to you in relation to your pain? Not at all supportive Extremely supportive 4. Rate your overall mood during the past week. Extremely low Extremely high 5. On the average, how severe has your pain been during the last week? Not at all severe Extremely severe 6. How much has your pain changed your ability to take part in recreational and other social activities? 7. How much do you limit your activities in order to keep your pain from getting worse? Not at all Very much 8. How much has your pain changed the amount of satisfaction or enjoyment you get from family-related activities? 9. How worried is your spouse (significant other) about you because of your pain? Not at all worried Extremely worried 10. During the past week how much control do you feel that you have had over your life? No control Extreme control 11. How much suffering do you experience because of your pain? No suffering Extreme suffering 12. How much has your pain changed your relationship with your spouse, family, or significant other? 13. How attentive is your spouse (significant other) to you because of your pain? Not at all attentive Extremely attentive 14. During the past week how much do you feel that you have been able to deal with your problems? Notatall Extremely well 15. How much control do you feel that you have over your pain? No control at all A great deal of control 16. How much has your pain changed your ability to do household chores? 17. How much has your pain interfered with your ability to plan activities? 18. During the past week how irritable have you been? Not at all irritable Extremely irritable 19. How much has your pain changed or interfered with your friendships with people other than your family? 20. During the past week how tense or anxious have you been? Not at all tense or anxious Extremely tense and anxious

9 IMPACT OF PAIN AFTER SCI, Widerström-Noga 403 Section 2 In this section, we are interested in knowing how your spouse (or significant other) responds to you when he or she knows that you are in pain. On the scale listed below each question, circle a number to indicate how often your spouse (or significant other) responds to you in that particular way when you are in pain. (All questions below are answered on the following response scale): Never Very often 1. Asks me what he/she can do to help. 2. Reads to me. 3. Gets irritated with me. 4. Takes over my jobs or duties. 5. Talks to me about something else to take my mind off the pain. 6. Gets frustrated with me. 7. Tries to get me to rest. 8. Tries to involve me in some activity. 9. Gets angry with me. 10. Gets me pain medication. 11. Encourages me to work on a hobby. 12. Gets me something to eat or drink. Section 3 Listed on the following pages are 18 daily activities. Please indicate: How often you do each of these by circling a number on the scale listed below. (All questions below are answered on the following response scale): Never Very often For each item please also circle the number indicating how pain has affected how often you participate in these activities on the following response scale. (All questions below are answered on the following response scale): Pain has reduced my participation in this activity: Not at all Extremely 1. How often do you wash dishes? 2. How often do you mow the lawn? ( Check here, if you do not have a lawn to mow). 3. How often do you go out to eat? 4. How often do you play cards or other games? 5. How often do you go grocery shopping? 6. How often do you work in the garden? ( Check here, if you do not have a garden) 7. How often do you go to a movie? 8. How often do you visit friends? 9. How often do you help with the house cleaning? 10. How often do you work on the car? ( Check here, if you do not have a car) 11. How often do you take a ride in a car or bus? 12. How often do you visit relatives? ( Check here, if you do not have relatives within 100 miles) 13. How often do you prepare a meal? 14. How often do you wash the car? ( Check here, if you do not have a car) 15. How often do you take a trip? 16. How often do you go to a park or beach? 17. How often do you do the laundry? 18. How often do you work on a needed household repair? Reprinted with permission of Robert D. Kerns, Dennis C. Turk, and Thomas E. Rudy. References 1. Rintala DH, Loubser PG, Castro J, Hart KA, Fuhrer MJ. Chronic pain in a community-based sample of men with spinal cord injury: prevalence, severity, and relationship with impairment, disability, handicap, and subjective well-being. Arch Phys Med Rehabil 1998;79: Widerström-Noga EG, Felipe-Cuervo E, Broton JG, Duncan RC, Yezierski RP. Perceived difficulty in dealing with consequences of spinal cord injury. Arch Phys Med Rehabil 1999; 80: Kemp BJ, Krause JS. Depression and life satisfaction among people aging with post-polio and spinal cord injury. Disabil Rehabil 1999;21: Levi R, Hulting C, Nash M, Seiger Å. The Stockholm spinal cord injury study 1. Medical problems in a regional SCI population. Paraplegia 1995;33: Ragnarsson KT. Management of pain in persons with spinal cord injury. J Spinal Cord Med 1997;20: Haythornthwaite JA, Benrud-Larson LM. Psychological aspects of neuropathic pain. Clin J Pain 2000;16 Suppl 2:S Turk DC. Biopsychosocial perspective on chronic pain. In: Gatchel R, Turk DC, editors. Psychological approaches to chronic pain management: a clinician s handbook. New York: Guilford Pr; p Turk DC, Flor H. Chronic pain: a biobehavioral perspective. In: Gatchel R, Turk DC, editors. Psychosocial factors in pain: critical perspectives. New York: Guilford Pr; p Turk DC. Customizing treatment for chronic pain patients: who, what, and why? Clin J Pain 1990;6: Summers JD, Rapoff MA, Varghese G, Porter K, Palmer RE. Psychosocial factors in chronic spinal cord injury pain. Pain 1991;47:183-9.

10 404 IMPACT OF PAIN AFTER SCI, Widerström-Noga 11. Yezierski RP. Pain following spinal cord injury: the clinical problem and experimental studies. Pain 1996;68: Siddall PJ, Taylor DA, Cousins MJ. Classification of pain following spinal cord injury. Spinal Cord 1997;35: Bryce TN, Ragnarsson KT. Pain after spinal cord injury. Phys Med Rehabil Clin North Am 2000;11: Turk DC, Melzack R, editors. Handbook of pain assessment. New York: Guilford Pr; Turk DC, Melzack R, editors. Handbook of pain assessment, 2nd ed. New York: Guilford Pr; Widerström-Noga EG, Felipe-Cuervo E, Yezierski RP. Relationships among clinical characteristics of chronic pain after spinal cord injury. Arch Phys Med Rehabil 2001;82: Kerns RD, Turk DC, Rudy TE. The West Haven Yale Multidimensional Pain Inventory (WHYMPI). Pain 1985;23: Turk DC, Okifuji A, Sinclair JD, Starz TW. Pain, disability, and physical functioning in subgroups of fibromyalgia patients. J Rheumatol 1996;23: Turk DC, Rudy TE. Toward an empirically derived taxonomy of chronic pain patients: integration of psychological assessment data. J Consult Clin Psychol 1988;56: Turk DC, Sist TC, Okifuji A, et al. Adaptation to metastatic cancer pain, regional/local cancer pain and non-cancer pain: role of psychological and behavioral factors. Pain 1998;74: Turk DC, Meichenbaum D, Genest M. Pain and behavioral medicine: a cognitive-behavioral perspective. New York: Guilford Pr; Bernstein IH, Jaremko ME, Hinkley BS. On the utility of the West Haven Yale Multidimensional Pain Inventory. Spine 1995;20: Bergström G, Jensen IB, Bodin L, Linton SJ, Nygren AL, Carlsson SG. Reliability and factor structure of the Multidimensional Pain Inventory Swedish language version (MPI-S). Pain 1998; 75: Lousberg R, Van Breukelen GJ, Groenman NH, Schmidt AJ, Arntz A, Winter FA. Psychometric properties of the Multidimensional Pain Inventory, Dutch language version (MPI-DLV). Behav Res Ther 1999;37: Riley JL, Zawacki TM, Robinson ME, Geisser ME. Empirical test of the factor structure of the West Haven Yale Multidimensional Pain Inventory. Clin J Pain 1999;15: Turk DC, Okifuji A, Sinclair JD, Starz TW. Differential responses by psychosocial subgroups of fibromyalgia syndrome patients to an interdisciplinary treatment. Arthritis Care Res 1998;11: Marsh HW, Balla JR, McDonald RO. Goodness-of-Fit Indexes in confirmatory factor analysis: the effect of sample size. Psychol Bull 1988;103: Hoelter JW. The analysis of covariance structures: goodness-of-fit indices. Sociol Methods Res 1983;11: Jöreskog KG, Sörbom D. LISREL V: analysis of linear structural relations by the method of maximum likelihood. Chicago: International Educational Serv; Tucker LR, Lewis C. The reliability coefficient for maximum likelihood factor analysis. Psychometrika 1973;38: Rosner B. Fundamentals of biostatistics. 3rd ed. Boston: PWS- Kent; SPSS, Inc. Systat 7.0: statistics. Chicago: SPSS; Stover SL, DeLisa JA, Whiteneck GG, editors. Spinal cord injury: clinical outcomes from the model systems. Gaithersburg (MD): Aspen; Krause JS, Kewman D, DeVivo MJ, et al. Employment after spinal cord injury: an analysis of cases from the Model Spinal Cord Injury Systems. Arch Phys Med Rehabil 1999;80: Conant LL. Psychological variables associated with pain perceptions among individuals with chronic spinal cord injury pain. J Clin Psychol Med Settings 1998;5: Macleod L, Macleod G. Control cognitions and psychological disturbance in people with contrasting physically disabling conditions. Disabil Rehabil 1998;20: Stensman R. Adjustment to traumatic spinal cord injury. A longitudinal study of self-reported quality of life. Paraplegia 1994; 32: Rudy TE, Kerns RD, Turk DC. Chronic pain and depression: toward a cognitive-behavioral mediation model. Pain 1988;35: Turk DC, Okifuji A, Scharff L. Chronic pain and depression: role of perceived impact and perceived control in different age cohorts. Pain 1995;61: Widerström-Noga EG, Felipe-Cuervo E, Yezierski RP. Chronic pain following spinal cord injury: interference with sleep and activities. Arch Phys Med Rehabil 2001;82: Turk DC. Assess the person, not just the pain. Pain Clin Updates 1993;1:1-4. Supplier a. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL