Coping and Recovery in Whiplash-associated Disorders

Transcription

1 ORIGINAL ARTICLE Coping and Recovery in Whiplash-associated Disorders Early use of Passive Coping Strategies is Associated With Slower Recovery of Neck Pain and Pain-related Disability Linda J. Carroll, PhD,* Robert Ferrari, MD, MSc (Med),wz J. David Cassidy, PhD, DrMedSc,y8z and Pierre Coˆte, DC, PhDz# Objective: Coping is shown to affect outcomes in chronic pain patients; however, few studies have examined the role of coping in the course of recovery in whiplash-associated disorders (WAD). The purpose of this study was to determine the predictive value of coping style for 2 key aspects of WAD recovery, reductions in neck pain, and in disability. Methods: A population-based prospective cohort study design was used to study 2986 adults with traffic-related WAD. Participants were assessed at baseline, 6 weeks, and 4, 8, and 12 months postinjury. Coping was measured at 6 weeks using the Pain Management Inventory, and neck pain recovery was assessed at each subsequent follow-up, using a 100 mm visual analogue scale (VAS). Disability was assessed at each follow-up using the Pain Disability Index (PDI). Pain recovery was defined as a VAS score of 0 to 10; disability recovery was defined as a PDI score of 0 to 4. Data analysis used multivariable Cox proportional hazards models. Results: Those using high versus low levels of passive coping at 6 weeks postinjury experienced 28% slower pain recovery and 43% slower disability recovery. Adjusted hazard rate ratios for pain recovery and disability recovery were 0.72 (95% CI, ) and 0.57 (95% CI, ), respectively. Active coping was not associated with recovery of neck pain or disability. Conclusions: Passive coping style predicts neck pain and selfassessed disability recovery. It may be beneficial to assess and improve coping style early in WAD. Key Words: whiplash, pain, disability, coping (Clin J Pain 2014;30:1 8) Coping style has been described as the pattern of behavioral or cognitive efforts to manage external or internal stressors. 1 This style is influenced by both stable, Received for publication July 12, 2012; revised December 27, 2012; accepted January 7, From the *Department of Public Health Sciences, Alberta Centre for Injury Prevention and Research; Departments of wmedicine; zrheumatic Diseases, University of Alberta, Edmonton, AB; 8Division of Outcomes and Population Health, Toronto Western Research Institute; zdivision of Epidemiology, Dalla Lana School of Public Health, University of Toronto; #Faculty of Health Sciences and UOIT-CMCC Centre for the Study of Disability Prevention and Rehabilitation, University of Ontario Institute of Technology (UOIT), Toronto, ON, Canada; and yinstitute of Sport Science and Clinical Biomechanics, Faculty of Health, University of Southern Denmark, Odense, Denmark. The authors declare no conflict of interest. Supported by Saskatchewan Government Insurance Corporation, Regina, Saskatchewan, Canada. Reprints: Linda J. Carroll, PhD, 4075 Research Transition Facility, Street, Edmonton, AB, Canada T6G 2E1 ( lcarroll@ualberta.ca). Copyright r 2013 by Lippincott Williams & Wilkins person-based factors (such as sociodemographic factors and personality), but also may depend on the nature of an acute stressor event. 2 Although the nature of the stressor may influence coping style, there is a large body of evidence indicating that coping style may impact an individual s health outcomes. Coping with pain is a particularly important example, with coping style having been shown to influence disability in a wide variety of painful disorders However, most of the pain coping research have studied populations with chronic pain. There is some evidence that outcome of acute injury may equally be influenced by coping style. Specifically, regarding whiplash-associated disorders (WAD), pain coping strategies assessed within a few days of the injury have been found to be associated with concurrent disability and to predict whether or not WAD patients complied with referral to active rehabilitation. 14 However, the longer term implications of very early coping styles is still unclear, and Kivioja et al 15 found no association between coping measured within days of the injury and WAD recovery 1 year later. Yet, when Buitenhuis et al 16 assessed coping a little later in the recovery process (at 2 to 18 wk after the injury), coping predicted claim duration, which might be considered a proxy for recovery. In examining more direct indices of WAD recovery, a small study by Olsson and colleagues showed an association between adaptive pain coping (measured by the West Haven-Yale Multidimensional Pain Inventory at 1 month after injury) and recovery of neck pain; whereas a larger, population-based study by Carroll and colleagues showed that use of passive pain coping strategies (measured by the Vanderbilt Pain Management Inventory [VPMI]) 6 weeks after the injury predicted slower self-reported global recovery. 17,18 Adaptive coping, as defined by the West Haven-Yale questionnaire reflects a high level of self-perceived control over pain, whereas the passive coping scale of the Vanderbilt PMI reflects a feeling of helplessness to deal with the pain and relying on others. 3,19 In contrast, when Kasch et al 20 measured coping using a general health behavior questionnaire (the Millon Behavioral Health Inventory) rather than a questionnaire specific to pain coping, there was no association between coping style and WAD recovery. Thus, the role of pain coping in WAD recovery has not yet been clearly established and requires further study. During the course of recovery from WAD, persons typically report neck pain as the most common symptom for which they seek treatment. In clinical scenarios, neck pain is a key aspect of the morbidity associated with WAD, and a frequent reason for reported disability. Both neck pain and pain-related disability are relevant outcomes as Clin J Pain Volume 30, Number 1, January

2 Carroll et al Clin J Pain Volume 30, Number 1, January 2014 they are often reported during the course of recovery, and are a significant source of societal costs associated with WAD. 21,22 Thus, neck pain severity and neck pain disability are not only important clinical markers of a person s progress in recovery, but also targets for early and costeffective intervention. Knowing which persons will report higher levels of pain and more neck pain disability during their course of recovery is of value if those persons are to be targeted for specific secondary prevention. The purpose of this study was to determine the predictive association between coping style, as measured at 6 weeks postinjury and both (1) neck pain recovery and (2) recovery of disability during the first year after onset of WAD. MATERIALS AND METHODS Design and Study Population Data for this study come from a population-based, prospective cohort study conducted in Saskatchewan, a Canadian province of approximately 1 million at that time, with a single motor vehicle insurer, Saskatchewan Government Insurance. 23 Coping data from this study have not previously been reported. Included were all adult Saskatchewan residents, 18 years or older, who submitted a claim to Saskatchewan Government Insurance for a traffic injury (a personal injury claim) that occurred over an 18-month inception period from 1994 to This included those presenting to insured health care providers for treatment of injuries, including physicians, physical therapists, chiropractors, and massage therapists. We included a subcohort of persons who had a WAD, and who consented to participate in the follow-up. WAD was operationally defined as self-reported neck pain resulting from having been in a motor vehicle collision. To exclude those more seriously injured, we excluded those who were hospitalized for >2 days. Measures and Sources of Data All data are self-reported, apart from insurance company administrative data on whether or not the claimant was considered to be at fault for the collision. All persons making a claim for a traffic injury completed an administrative claim form, which also served as our baseline questionnaire. This questionnaire included information about sociodemographic characteristics, collision-related factors, prior self-reported health, and postcollision symptoms, including depressive symptomatology, pain extent (percentage of body in pain), and intensity (measured separately for neck pain, headache pain, and other pain). Descriptions of these measures are reported later in this section. At 6 weeks postcollision, consenting participants completed a mailed questionnaire that repeated questions about their current pain intensity and extent. This followup questionnaire also assessed pain-related disability and pain coping for the first time (these 2 measures described below). These follow-up questionnaires were repeated at 4, 8, and 12 months after the collision. During the course of the study period, the insurance system changed from a tort system (with limited benefits regardless of fault, but in which those not at fault for the collision could sue for pain and suffering) to a no fault system (with expanded benefits regardless of fault, but limited scope for legal action and provided no payment for pain and suffering). The first 6 months of the inception period was under the tort system, and the remaining 12 months of inception was no fault. Pain Coping Our predictor of interest, pain coping, was assessed at the time of the first follow-up (6 wk postcollision) using the VPMI, an 18-item Likert-type scale, which asks participants to report how often they use particular coping strategies when their pain is at a moderate or greater level of intensity. 3 This questionnaire yields 2 independently scored subscales reflecting types of pain coping strategies, labeled by the questionnaire developers as passive and active. The VPMI is valid and reliable. 3,21,22,24 The 2- factor structure of the scale has been validated, and the internal consistency of both subscales has been confirmed. 3,21,22,24 The 2 subscales are relatively orthogonal: The original validation study of people with chronic pain reported a correlation between these 2 subscales of 0.29, and in a different cohort of persons with subacute WAD, the correlation between scales was ,17 Active coping refers to use of coping strategies that involve attempts to control the pain or to function in spite of the pain, and the subscale includes strategies such as engaging in physical exercise or physical therapy and staying busy or active. Scores from this 7-item subscale range from 7 to 35. Passive coping refers to coping strategies that involve withdrawing from activities due to pain or relying on others for pain management, and includes items such as restricting or canceling social activities due to pain and wishing for better pain medication. Scores on that 11-item subscale range from 11 to 55. Passive and active coping strategies should not be considered mutually exclusive; that is, frequent use of the types of strategies included in the passive coping scale does not preclude frequent use of the types of strategies included in the active coping scale. 25 There are no cut-points validated in the literature; thus, we conducted our analyses using a tertile split for each subscale. Outcome 1: Neck Pain Recovery Neck pain intensity was measured on a 100 mm visual analogue scale (VAS) at baseline and at each follow-up. The VAS is a valid, reliable, and commonly used way of assessing pain intensity. 26,27 To develop a pain outcome criterion, a neck pain intensity score of 0 to 10 mm was considered to be recovered. Although this does not necessarily reflect a complete cessation of pain, a score of r30 on the pain VAS is considered to be mild, 28 and thus a score of 10 is at the lower end of the mild classification. We judged this to be a conservative cut-off point for pain recovery. Outcome 2: Recovery of Pain Disability As a measure of pain-related disability, we used the Pain Disability Index (PDI), which is a 7-item self-report scale designed to measure the extent to which pain interferes with various areas of activity: family/home responsibility, recreation, social activity, occupation, sexual behavior, selfcare, and life-support activity It was designed as a brief global measure of self-perceived disability, and total scores range from 0 to 70, with higher scores reflecting greater perceived disability. The PDI has been demonstrated to be a valid and reliable measure of perceived disability There is no cut-off suggested for the PDI to reflect recovery, and we arbitrarily chose a cut-off score of 0 to 4 to reflect recovery as this cut-off has been used in previous research. 35 As this is a very conservative index of recovery, we also conducted a 2 r 2013 Lippincott Williams & Wilkins

3 Clin J Pain Volume 30, Number 1, January 2014 Coping and Recovery in WAD sensitivity analysis using scores of 0 to 10 to reflect disability recovery. Potential Confounders Factors considered to potentially confound the association between coping and pain and disability recovery were selected on the basis of theory and empirical findings from other studies. Factors considered as potential confounders were: sociodemographic factors (age, sex, marital status, education level, and income), collision factors (designated to have been at fault for the collision, direction of the impact, head position at the time of the collision, seatbelt use, and fixed or adjustable head restraint), proxy factors for collision or injury severity (whether or not the vehicle was drivable after the collision, admission to hospital, and time lost from work due to the collision), whether they had consulted with a lawyer, whether they claimed under the tort or no fault system, self-rated precollision health (excellent, very good, good, fair, and poor), postcollision depressive symptomatology (as measured by the Centre for Epidemiological Studies Depression Scale, CES-D), 36 percentage of body in pain (as assessed by a pain drawing), 37 and pain intensity (as assessed on a 100 mm VAS). Headache and pain intensity in other parts of the body were considered as potential confounders for all analyses. Baseline neck pain intensity was considered a potential confounder for those analyses where disability recovery was the outcome. Several factors (claiming under tort or no fault, at fault for the collision, and retaining a lawyer to assist with the claim and depressive symptomatology) were also examined to assess their potential role as effect modifiers. Analysis Descriptive Analyses We reported the characteristics of the cohort, stratified by low, medium, and high tertiles of active and passive coping. We reported the correlation between passive and active coping scores (using total scores) using Pearson product-moment correlation coefficients. We also reported the Pearson product-moment correlation coefficients between both passive and active coping scores (using total scores) and baseline pain intensity, pain intensity at 6 weeks; depression scores (as assessed using the total score of the CES-D) at baseline and at 6 weeks; and pain disability (as assessed using the total score of the PDI) at 6 weeks postcollision. Recovery Analyses In 4 separate sets of analyses, we report the independent association between each of our 2 predictor variables, passive and active coping and our 2 outcome variables, recovery of neck pain and recovery of disability. To do this, we built separate multivariable Cox proportional hazards models for each type of coping and each outcome variable, to yield 4 final models reflecting time to recovery. Estimates of association are reported in hazard rate ratios (HRRs), which are interpreted in a similar way to relative risks or odds ratios (although they are not identical to these 2 estimates of association), and are used in time to event (in this case, time to recovery) studies. 38 As mentioned above, neck pain recovery was operationally defined as recovery to neck pain intensity score of 0 to 10 mm on the 100 mm VAS (failure to recover was defined as a score of 11 to 100 mm). For these analyses we excluded those with neck pain intensity scores of r10 mm at the first follow-up (as they had already reached the criterion for pain recovery). Likewise, for the analyses examining time to disability recovery, we excluded those with PDI scores of r4 at the first followup. For the sensitivity analysis, in which disability recovery was defined as a score of 0 to 10 on the PDI, we excluded those with PDI scores of r10 at the first follow-up. Data were censored at 365 days for those with no disability recovery but complete follow-up through the follow-up period. For those with no disability recovery, but missing data for subsequent follow-up points, the censor date was midway between the last disability score provided and the first missing follow-up point. For each model, we identified the crude (unadjusted) association between the coping score in tertiles and the pain recovery outcome. We then built a series of bivariate models that included the coping score and each of the potential confounders (listed above) individually. Where the estimate of the association between coping and the outcome was changed by Z10% by the addition of a potential confounder, that factor was considered to be a confounder. 39 The final models adjusted for these confounders. In addition, for the pain recovery analysis, we adjusted for pain intensity at 6 weeks (concurrent with the assessment of coping), in consideration of the idea that those with lower pain scores may be more likely to meet the pain recovery criterion than those with high pain scores. Likewise, for the disability recovery analysis, we adjusted for PDI score at the 6-week follow-up (the first point at which PDI was administered). Those analyses with passive coping as the predictor variable were also adjusted for active coping scores, and those with active coping as the predictor were adjusted for passive coping scores. No effect modification was identified. Analyses were performed using SPSS 19 and STATA-SE, version ,41 Identifying confounders of the coping/recovery association is distinct from identifying predictors or recovery. A factor that confounds the coping/recovery association is, by definition, a predictor of recovery, but must also be associated with the exposure, coping; must be a variable that is extraneous to the coping/recovery association but must affect the association between coping and recovery; and must not be a mediator of the coping/recovery association. 42 Thus, the reported models reflect the independent association between coping and recovery, but predictors of recovery that do not affect the association between coping and recovery are not included in the models. RESULTS Of the 7462 persons with traffic-related whiplash injuries at baseline, 2847 responded at follow-up 1 (6 wk postcollision). Of these, 2428 indicated that they had moderate or greater levels of pain (and so were eligible to complete the coping questionnaire), 2425 completed the passive coping scale, and 2420 completed the active coping scale. The mean passive coping score was 29.4 (SD 7.9; median, 20) and the mean active coping score was 20.4 (SD 4.8; median, 20). After categorizing into tertiles, the cut-off scores for low, medium, and high passive coping were 26 and 33; the cut-off scores for low, medium, and high active coping were 18 and 22. The Pearson correlation between the passive and active subscale scores was The coping questionnaire was completed a median of 49 days postinjury. From the above sample, we further excluded those who had reached the relevant recovery criterion at the r 2013 Lippincott Williams & Wilkins 3

4 Carroll et al Clin J Pain Volume 30, Number 1, January 2014 TABLE 1. Characteristics of Sample Passive Coping at 6 wk* Active Coping at 6 wk* Factors Low Medium High Low Medium High Age: mean (SD) 35.5 (14.7) 37.6 (14.2) 37.5 (14.4) 35.2 (14.3) 35.8 (13.8) 39.7 (14.7) Sex: female (N [%]) 503 (30.9) 559 (34.3) 566 (34.8) 403 (24.8) 520 (32.0) 702 (43.2) Education (N [%]) rgrade 8 32 (32.7) 28 (28.6) 38 (38.6) 44 (45.4) 21 (21.6) 32 (33.0) > Grade 8 but <High school 105 (25.1) 121 (28.9) 193 (46.1) 132 (31.5) 146 (34.8) 141 (33.7) High school 218 (33.0) 203(30.7) 240 (36.3) 194 (29.4) 225 (34.1) 241 (36.5) Postsecondary 317 (33.0) 352 (36.6) 292 (30.1) 234 (24.2) 310 (32.3) 415 (43.3) University graduate 107 (37.4) 115 (40.2) 64 (22.4) 45 (15.8) 95 (33.3) 145 (50.9) Prior health (N [%]) Excellent 344 (37.0) 317 (34.1) 269 (28.9) 259 (27.9) 292 (31.5) 376 (40.6) Very good 278 (32.1) 289 (33.4) 299 (34.5) 215 (24.9) 300 (34.7) 350 (40.6) Good 141 (27.5) 175 (34.1) 197 (38.4) 131 (25.6) 178 (34.8) 203 (39.6) Fair or poor 16 (16.3) 37 (32.2) 62 (53.9) 43 (37.4) 27 (23.5) 45 (39.1) Neck pain (mean [SD])w Baseline 50.4 (23.4) 56.9 (23.4) 64.7 (21.9) 61.2 (22.9) 57.1 (23.8) 55.3 (23.6) 6 wk 34.3 (23.1) 43.5 (24.5) 56.8 (24.1) 49.0 (26.7) 43.4 (24.9) 43.8 (25.3) Pain disability: mean (SD)z 14.0 (11.8) 25.3 (14.0) 35.8 (14.2) 28.9 (16.8) 25.4 (15.1) 22.8 (15.9) Depression (mean [SD])y Baseline 13.4 (9.8) 18.0 (11.5) 23.0 (12.3) 20.6 (12.5) 18.6 (11.9) 16.3 (11.3) 6 wk 9.9 (8.5) 16.2 (10.7) 23.9 (11.6) 20.1 (12.6) 17.1 (11.3) 14.3 (11.1) Admitted to hospital (N [%])8 33 (31.1) 24 (22.6) 49 (46.2) 29 (27.6) 38 (36.2) 38 (36.2) Off work due to collision (N [%])z8 333 (28.9) 410 (35.6) 410 (35.6) 304 (26.4) 411 (35.7) 437 (37.9) Hired a lawyer (N [%])# Baseline 26 (16.1) 49 (30.4) 86 (53.4) 50 (31.3) 60 (37.5) 50 (31.3) 6 wk 50 (17.8) 91 (32.4) 140 (49.8) 88 (31.3) 90 (32.0) 103 (36.7) *Coping measured using the Vanderbilt Pain Management Inventory. Cut-points for low, medium, and high passive coping were 26 and 33. Cut-points for low, medium, and high active coping were 18 and 22. wneck pain measured at baseline and at 6 weeks postcollision on a 100 mm visual analogue scale. zpain disability measured at 6 weeks postcrash on the Pain Disability Index. ydepressive symptoms measured at baseline and at 6 weeks postcollision on the Centre for Epidemiological Studies Depression Scale. 8Admitted to hospital overnight or longer after the crash. zmissed work because of the injury. #Had hired a lawyer to help them with their claim at baseline and at 6 weeks postcollision. first follow-up; that is, for the outcome of pain recovery, we excluded those who had already recovered to pain intensity scores of r10 mm on the pain VAS; and for the outcome of disability recovery, we excluded those who had scores of r4 at that first follow-up (for the sensitivity analysis which utilized a recovery score of 0 to 10 on the PDI, we excluded those with scores of r10 at the first follow-up). Numbers included in each analysis are reported in Tables 2 5. Description of the Sample A description of the sample, stratified by low, medium, and high passive and active coping scores at 6 weeks is presented in Table 1. In addition, the correlations between passive coping scores at 6 weeks and pain intensity measured at baseline and 6 weeks were and 0.397, respectively. Similarly, the correlations between passive coping and depression measured at baseline and at 6 weeks TABLE 2. Association of Passive Coping, Measured at 6 Weeks Postcollision, With Time to Pain Recovery* Passive Coping Crude HRR (95% CI)w Adjusted HRR (95% CI)z Low passive score Medium passive score 0.65 ( ) 0.82 ( ) High passive score 0.41 ( ) 0.72 ( ) Crude and adjusted hazard rate ratios (HRR) and their 95% confidence intervals (95% CI). *Pain recovery defined as self-reported pain intensity of 0 to 10 mm on the 100 mm visual analogue scale. wn = Of these, 898 experienced pain recovery during the course of follow-up. zn = Of these, 791 experienced pain recovery. Adjusted for the following confounders: baseline headache intensity and baseline depression score. Also adjusted for active coping score (at 6 wk) and neck pain intensity at 6 weeks postcollision. Log likelihood = ; LR w 2 = TABLE 3. Association of Passive Coping, Measured at 6 Weeks Postcollision, With Time to Recovery of Disability* Passive Coping Crude HRR (95% CI)w Adjusted HRR (95% CI)z Low passive score Medium passive score 0.50 ( ) 0.74 ( ) High passive score 0.21 ( ) 0.57 ( ) Crude and adjusted hazard rate ratios (HRR) and their 95% confidence intervals (95% CI). *Disability recovery defined as score of 0 to 4 on the 70-point Pain Disability Index (PDI). wn = Of these, 414 experienced disability recovery. zn = Of these, 379 experienced disability recovery. Adjusted for confounders: percentage of body in pain, baseline neck pain and baseline other pain, and baseline depression. Also adjusted for active coping (at 6 wk) and PDI score at 6 weeks. Log likelihood = ; LR w 2 = r 2013 Lippincott Williams & Wilkins

5 Clin J Pain Volume 30, Number 1, January 2014 Coping and Recovery in WAD TABLE 4. Association of Active Coping, Measured at 6 Weeks Postcollision, With Time to Pain Recovery* Active Coping Crude HRR (95% CI)w Adjusted HRR (95% CI)z Low active score Medium active score 0.85 ( ) 0.98 ( ) High active score 1.03 ( ) 0.94 ( ) Crude and adjusted hazard rate ratios (HRR) and their 95% confidence intervals (95% CI). *Pain recovery defined as self-reported pain intensity of 0 to 10 mm on the 100 mm visual analogue scale. wn = Of these, 895 experienced pain recovery. zn = Of these, 873 experienced pain recovery. No baseline confounders. Adjusted for passive coping (at 6 wk) and neck pain intensity at 6 weeks. Log likelihood = ; LR w 2 = were and 0.515, respectively, and the correlation between passive coping and pain disability scores (both measured at 6 wk) was Corresponding correlations between active coping and these measures were lower, and are as follows: pain intensity at baseline and at 6 weeks ( and 0.077, respectively); depression at baseline and at 6 weeks ( and 0.228, respectively); and pain disability ( 0.180). All correlations were statistically significant at P < Passive Coping and Recovery Using low passive coping as the reference category, the adjusted association between passive coping and pain recovery suggests that medium passive copers are 18% less likely at any follow-up time point to have achieved pain recovery, whereas highly passive copers are 28% less likely to have recovered (adjusted HRR = 0.82; 95% confidence intervals [CI], and adjusted HRR = 0.72; 95% CI, , respectively). Crude and adjusted HRRs are shown in Table 2. Again, using low passive coping as the reference category, the adjusted association between passive coping and disability recovery suggests that medium passive copers are 25% less likely at any follow-up time point to have achieved disability recovery, whereas highly passive copers are 44% less likely (adjusted HRR = 0.75; 95% CI, and adjusted HRR = 0.56; 95% CI, , respectively). Crude and adjusted HRR are shown TABLE 5. Association of Active Coping, measured at 6 Weeks Postcollision, With Time to Recovery of Disability* Active Coping Crude HRR (95% CI)w Adjusted HRR (95% CI)z Low active score Medium active score 1.18 ( ) 0.93 ( ) High active score 1.25 ( ) 0.89 ( ) Crude and adjusted hazard rate ratios (HRR) and their 95% confidence intervals (95% CI). *Disability recovery defined as score of 0 to 4 on the 70-point Pain Disability Index (PDI). wn = Of these, 412 experienced disability recovery. zn = Of these, 382 experienced disability recovery. Adjusted for confounder baseline depression. Also adjusted for passive coping (at 6 wk) and PDI score at 6 weeks. Log likelihood = ; LR w 2 = in Table 3. Findings were relatively similar for the analysis using a cut-off PDI score of 10 rather than 4, with adjusted HRRs of 0.84 (95% CI, ) and 0.51 (95% CI, ) for medium and high passive coping scores, respectively. Active Coping and Recovery Active coping was not associated with either pain recovery or disability recovery in either the crude or the adjusted analyses (Tables 4 and 5). When disability recovery was defined by the less conservative cut-score of 10 instead of 4, medium active coping scores showed an association with disability recovery in the crude, but not the adjusted analysis. DISCUSSION This population-based cohort study of individuals with WAD shows that those reporting a passive coping style at 6 weeks after onset of WAD exhibit a slower rate of recovery from neck pain and neck pain disability as compared with those with WAD who are less likely to engage a passive coping style. For example, at 6 weeks postcollision, those who report they are likely to restrict social activities due to pain, focus on pain, or wish for better pain medications, took longer to recover from neck pain or disability, even after adjusting for a number of confounders, including baseline pain and disability. These findings add to the small, but growing body of evidence that suggests that when assessed within the first weeks after an injury, coping is associated with subsequent recovery, as measured in a variety of ways. Interestingly, although compensation-related factors (claiming under the tort vs. no fault system, being at fault for the crash, and retaining a lawyer to assist with the claim) were all predictors of recovery outcomes, none of these factors were either confounders or effect modifiers of the relationships between coping and recovery. Instead, the analyses identified postinjury symptoms (pain and depressive symptomatology) as the most relevant confounders. In addition, unlike a previous study that found depression to be an effect modifier of the association between passive coping and self-reported global recovery, 17 in the present cohort, depression did not serve this role. This study has some strengths. It was a large study, which permitted us to identify and adjust for a large number of relevant health-related and injury-related factors. This adds to our confidence that the reported estimates of the association between coping and recovery are largely unconfounded by other factors. We used valid and reliable measures, and were able to examine 2 clinically relevant indices of recovery, neck pain and a more global measure of disability. This study also has limitations. We did not have access to medical records that might have confirmed a diagnosis of WAD and we cannot determine WAD classification, as proposed by the Quebec Task Force on Whiplash-Associated Disorders. 43 However, determination of WAD grade I requires only pain and/or neck stiffness and/or tenderness, and no physical signs. As pain is a subjective phenomenon, and has been shown to be validly assessed using self-report, 26 there seem to be similarities between a patient reporting neck pain after a traffic crash to his or her physician, and a report of postcollision neck pain on a questionnaire. In addition, we did not assess coping before 6 weeks after the injury, and do not know to what r 2013 Lippincott Williams & Wilkins 5

6 Carroll et al Clin J Pain Volume 30, Number 1, January 2014 extent coping strategies were modified or encouraged by contact with the health care system. The association (adverse or otherwise) between coping style and treatment style is unknown. We did not have information on how health care was administered or on the specific characteristics of health care that was provided. However, the type of health care provider(s) was not a confounding variable, that is, adjusting for type of health care provider did not affect the association between passive coping and recovery. In addition, because the questionnaire asks about coping strategies for moderate or greater levels of pain, we cannot necessarily extend these findings to those with mild pain only; although it is possible that for many individuals, even mild pain is sufficiently distressing to require efforts to cope. Finally, the number of respondents who provided data regarding their coping style was relatively low, on the order of 40% of the original cohort. This is primarily due to attrition between baseline and the first follow-up, where coping style was assessed. It is possible that the association between coping and recovery was different for participants than for nonparticipants. However, the estimates were relatively robust to confounding, which argues against that idea. In addition, previous analyses comparing those who did and did not participate in follow-up revealed only minor differences, for example, the average baseline neck pain in respondents was 59 compared to 62 (on the 100 mm VAS) in nonrespondents, and 11% of participants and 9% of nonparticipants reported university education. 44 Over a period of 12 months follow-up after onset of WAD, considering those individuals who had not already reached definitions of recovery (neck pain VAS score of r10), approximately 40% of the remaining cohort had achieved recovery from neck pain. Over this same period of 12 months, approximately 20% of participants reported recovery from neck pain disability when recovery was defined as a score of 0 to 4 on the PDI and approximately 35% recovered where recovery was defined less conservatively, at a score of 0 to 10. Obviously, rates of both pain and disability recovery are highly dependent on the criteria used to define recovery. In addition, the recovery rates (from neck pain or disability) in this study do not reflect the overall proportion of recovery, as we excluded individual who had already met a definition for recovery at 6 weeks follow-up. However, the rates of recovery for those who have not yet recovered by 6 weeks have an important association with coping style. The time to recovery was slower for highly passive copers by almost 30% than it was for those using few passive coping strategies. Likewise, the time to recovery of activities of daily living was >40% slower for those using high versus low levels of passive coping. This study confirms that, in patients who have not already recovered at 6 weeks postcollision, that is, those who thus are at highest risk for chronic pain and disability, the rate of recovery from neck pain and neck pain disability is strongly influenced by coping style. This influence was independent of demographic, socioeconomic, health, compensation, and injury-related characteristics. This suggests that coping styles that emerge within a few weeks of a whiplash injury should be monitored and that frequent use of those types of coping strategies captured by the passive coping subscale should be considered a red flag for slowed recovery. The mechanism by which the endorsement of a passive coping style influences recovery from neck pain and disability is not well studied. There is some evidence from a small clinical study carried out in a primary care practice that the coping style that patients endorse early after injury predicts short-term medication use and compliance with active treatment programs in the weeks subsequent to the injury, and it may be that these behaviors have longer term consequences for recovery 14 ; nor is it clearly understood why individuals with whiplash injuries use the coping strategies they use. As can be seen from Table 1 and in the Results section, those who use more passive coping at 6 weeks after the injury had greater initial neck pain intensity and also greater neck pain at 6 weeks, have greater depressive symptomatology, both initially and at 6 weeks, were more likely to have been admitted to hospital after the crash, were more likely to have hired a lawyer to help them with their claim, and report having had only fair or poor health before the injury. However, one cannot make causal inferences from these data. It seems likely that having a more severe injury leads to more passive coping, but it would also seem that coping is not simply a reflection of the severity of the injury. The current study shows that even after adjusting for pain intensity, depression, and pain disability, passive coping predicts slower recovery. If coping style were nothing more than a consequence of injury severity, adjusting for these factors should have explained away any association between coping and recovery. Despite the strong associations between passive coping and pain intensity, depression, and disability, even after adjusting for these indices, passive coping score predicted both pain recovery and disability recovery. In the case of active coping, the associations with pain severity, depression, and disability were not strong, and even before adjusting for these indices of severity, there was little or no association between that active coping and recovery. This also suggests that active coping is not a reflection of neck condition severity. There is some evidence to suggest that having a passive coping style in general has a negative influence on health and health outcomes even before the injury, and that those who cope passively also react more negatively to health conditions. A recent study of WAD patients in the first 3 weeks after their injury suggests that daily fluctuations in physical and mental well-being are associated with fluctuations in the coping process. On good days, there were corresponding feelings of high self-efficacy and effective coping, whereas on bad days, stressors (including pain) were viewed as a threat and participants also engaged in more catastrophizing thinking. 45 Catastrophizing itself has been linked with recovery in WAD and in other pain conditions, but there is no consensus on whether it is actually a pain coping strategy or a different type of psychological variable. Sullivan et al, 46,47 who developed a questionnaire measuring catastrophizing, refers to it as an exaggerated negative orientation to actual or anticipated pain comprising elements of rumination, magnification, and helplessness. Current research in the area suggests that catastrophizing is associated with both type and effectiveness of coping strategies used. 48 Finally, one should note that active coping style had no effect on recovery from neck pain or from neck pain disability. That passive coping style is the more relevant coping style in pain outcomes has been previously demonstrated. 17,49 Thus, identifying individuals at an early stage who are likely to engage passive coping strategies seems to be the key, and flagging these individuals at risk for slowed recovery may provide a potential opportunity to 6 r 2013 Lippincott Williams & Wilkins

7 Clin J Pain Volume 30, Number 1, January 2014 Coping and Recovery in WAD modify their coping style (ie, encouraging compliance with exercise therapy, and less reliance on prescription medications may be mitigating a passive coping style). Although intervention studies have provided some evidence that coping strategies can be modified in treatment settings, 50,51 further treatment trials in these at-risk individuals with WAD are required. REFERENCES 1. Lazarus RS, Folkman S. Stress, Appraisal, and Coping. New York: Springer Publishing Company; Holahan CJ, Moos RH, Schaefer JA. Coping, stress resistence, and growth: conceptualizing adaptive functioning. In: Zeidner M, Endler NS, eds. Handbook of Coping Theory, Research, Applications. New York: John Wiley & Sons Inc.; 1996: Brown GK, Nicassio PM. Development of a questionnaire for the assessment of active and passive coping strategies in chronic pain patients. Pain. 1987;31: Gil KM, Abrams MR, Phillips G, et al. Sickle cell disease pain: 2. Predicting health care use and activity level at 9-month follow-up. J Consult Clin Psychol. 1992;60: Jensen IB, Nygren Å, Gamberale F, et al. Coping with longterm musculoskeletal pain and its consequences: is gender a factor? Pain. 1994;57: Jensen MP, Turner JA, Romano JM. Coping with chronic pain: a critical review of the literature. Pain. 1991;47: Jensen MP, Turner JA, Romano JM. Changes in beliefs, catastrophizing, and coping are associated with improvement in multidisciplinary pain treatment. J Consult Clin Psychol. 2001;69: Keefe FJ, Affleck G, Lefebvre JC, et al. Pain coping strategies and coping efficacy in rheumatoid arthritis: a daily process analysis. Pain. 1997;69: Poelsson M, Gerdle B. Coping in patients with chronic whiplash-associated disorders: a descriptive study. J Rehabil Med. 2004;36: Rosenstiel AK, Keefe FJ. The use of coping strategies in chronic low back pain patients: relationship to patient characteristics and current adjustment. Pain. 1983;17: Söderlund A, Lindberg P. Whiplash-associated disorders predicting disability from a process-oriented perspective of coping. Clin Rehabil. 2003;17: Spinhoven P, Ter Kuile MM, Kole-Snijders AM, et al. Catastrophizing and internal pain control as mediators of outcome in the multidisciplinary treatment of chronic low back pain. Eur J Pain. 2004;8: van-lankveld W, Naring G, van- t Pad B, et al. Behavioral coping and physical functioning: the effect of adjusting the level of activity on observed dexterity. J Rheumatol. 1999;26: Ferrari R, Louw D. Coping style as a predictor of compliance with referral to active rehabilitation in whiplash patients. Clin Rheumatol. 2011;20: Kivioja J, Jensen I, Lindgren U. Early coping strategies do not influence the prognosis after whiplash injuries. Injury. 2005;36: Buitenhuis J, Spanjer J, Fidler V. Recovery from acute whiplash. The role of coping styles. Spine. 2003;28: Carroll LJ, Cassidy JD, Coˆ te P. The role of pain coping strategies in prognosis after whiplash injury: passive coping predicts slowed recovery. Pain. 2006;124: Olsson I, Bunketorp O, Carlsson SG, et al. Prediction of outcome in whiplash-associated disorders using West Haven- Yale Multidimensional Pain Inventory. Clin J Pain. 2002;18: Kerns RD, Turk DC, Rudy TE. The West Haven-Yale Multidimensional Pain Inventory. Pain. 1985;23: Kasch H, Bach FW, Jensen TS. Handicap after acute whiplash injury: a 1-year prospective study of risk factors. Neurology. 2001;56: Smith CA, Wallston KA, Dwyer KA. On babies and bathwater: disease impact and negative affectivity in the selfreports of persons with rheumatoid arthritis. Health Psychol. 1995;14: Snow-Turek AL, Norris MP, Tan G. Active and passive coping strategies in chronic pain patients. Pain. 1996;64: Cassidy JD, Carroll LJ, Coˆ té P, et al. Effect of eliminating compensation for pain and suffering on the outcome of insurance claims for whiplash injury. N Engl J Med. 2000;342: Mercado AC, Carroll LJ, Cassidy JD, et al. Coping with neck and low back pain in the general population. Health Psychol. 2000;19: Carroll LJ, Mercado AC, Cassidy JD, et al. A populationbased study of factors associated with combinations of active and passive coping with neck and low back pain. J Rehabil Med. 2002;34: Jensen MP, Karoly P. Self-report scales and procedures for assessing pain in adults. In: Turk DC, Melzack R, eds. Handbook of Pain Assessment. 2nd ed. New York: The Guilford Press; 2001: Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain. 1986;27: Collins SL, Moore RA, McQuay HJ. The visual analogue pain intensity scale: what is moderate pain in millimeters? Pain. 1997;72: Pollard CA. Preliminary validity study of the Pain Disability Index. Percept Mot Skills. 1984;59: Tait RC, Chibnall JT, Krause SJ. The Pain Disability Index: psychometric properties. Pain. 1990;40: Tait RC, Pollard CA, Margolis RB, et al. The Pain Disability Index: psychometric and validity data. Arch Phys Med Rehabil. 1987;68: Gronblad M, Hypli M, Wennerstrand P, et al. Intercorrelation and test-retest reliability of the Pain Disability Index (PDI) and the Oswestry Disability Questionnaire (ODQ) and their correlation with pain intensity in low back pain patients. Clin J Pain. 1993;9: Gronblad M, Lukinmaa A, Konttinen YT. Chronic low-back pain: intercorrelation of repeated measures for pain and disability. Scand J Rehabil Med. 1990;22: Jerome A, Gross RT. Pain Disability Index: construct and discriminant validity. Arch Phys Med Rehabil. 1991;72: Holm LW, Carroll LJ, Cassidy JD, et al. Expectations for recovery important in the prognosis of whiplash injuries. PLoS Med. 2008;5:e Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1: Margolis RB, Tait RC, Krause SJ. A rating system for use with patient pain drawings. Pain. 1986;24: Hosmer DW, Lemeshow S. Applied Survival Analysis. New York: John Wiley & Sons; Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; SPSS Inc. SPSS for Windows, Version 19. Chicago: SPSS Inc.; StataCorp. Stata Statistical Software: Release 12. College Station, TX: Stata Corporation; Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia: Lippincott-Raven; Spitzer WO, Skovron ML, Salmi LR, et al. Scientific monograph of the Quebec Task Force on whiplash-associated disorders: redefining whiplash and its management. Spine. 1995;20:1S 73S. 44. Coˆ té P, Hogg-Johnson S, Cassidy JD, et al. The association between neck pain intensity, physical functioning, depressive symptomatology and time-to-claim-closure after whiplash. J Clin Epidemiol. 2001;54: r 2013 Lippincott Williams & Wilkins 7

8 Carroll et al Clin J Pain Volume 30, Number 1, January Bring A, Bring J, Söderlund A, et al. Coping patterns and their relation to daily activity, worries, depressed mood, and pain intensity in acute whiplash-associated disorders. Int J Behav Med. 2012, In Press. DOI: /s y. 46. Sullivan MJ, Bishop S, Pivik J. The Pain Catastrophizing Scale: development and validation. Psychol Assess. 1995;7: Sullivan MJ, Thorn B, Haythornthwaite J, et al. Theoretical perspectives on the relation between catastrophizing and pain. Clin J Pain. 2001;17: Sullivan MJ, Adams H, Martel M, et al. Catastrophizing and perceived injustice. Spine. 2011;36:S244 S Mercado AC, Carroll LJ, Cassidy JD, et al. Passive coping is a risk factor for disabling neck or low back pain. Pain. 2005;117: Jensen IB, Nygren A, Lundin A. Cognitive-behavioural treatment for workers with chronic spinal pain: a matched and controlled cohort study in Sweden. Occup Environ Med. 1994;51: Sharpe L, Sensky T, Timberlake N, et al. A blind, randomized controlled trial of cognitive behavioral intervention for patients with recent onset rheumatoid arthritis: preventing psychological and physical morbidity. Pain. 2001;89: r 2013 Lippincott Williams & Wilkins