Relationship of pain-coping strategies and pain-specific beliefs to pain experience in children with juvenile idiopathic arthritis
To examine whether pain-specific beliefs and coping strategies of patients with juvenile idiopathic arthritis (JIA) independently predict their reported pain, while controlling for relevant demographic variables, disease activity, and parent-rated disability. To compare use of pain-coping strategies and pain-related beliefs of a selected subgroup of patients with high pain and low disease activity (high pain group) with the remaining patients.
Children with JIA (n = 56) completed the Pain Coping Questionnaire, a revised version of the Survey of Pain Attitudes, and a 3-week pain diary. The parents completed the Childhood Health Assessment Questionnaire (CHAQ). Second order principal component analyses were conducted in order to reduce the number of independent variables. Regression analyses of the dependent measure were performed. The use of coping strategies and health beliefs were compared using t-tests for independent samples. Pearson's correlation coefficients were calculated to examine the direct associations between each individual coping and belief scale, and the pain measure.
Only the CHAQ and the cognitive belief composite factor score made statistically significant contribution to the prediction of pain after controlling for other variables. Significant differences were found between the scores of high pain patients and the rest of the group for the health belief subscale of disability (mean ± SD 2.0 ± 0.6 and 1.2 ± 0.7, respectively), and for the health belief subscale of harm (mean ± SD 2.7 ± 0.6 and 1.8 ± 0.7, respectively). Significant correlations were obtained between the pain measure and the pain-coping subscale of catastrophizing, the pain belief subscales of disability, harm, solicitude (inverse), control, and medical cure.
These results support a model of pain experience in patients with JIA where psychological factors are strongly influential. It may be efficient to focus behavioral interventions on a subgroup of children where the pain experience seems to be in discordance with the disease activity.
Pain is one of the primary symptoms in children with juvenile idiopathic arthritis (JIA), and recent studies have shown that pain is more prevalent in patients with JIA than previously recognized. For example, Schanberg et al found that children with JIA reported having pain an average of 73% of days over a 2-month period, and that 31% reported pain in the severe range (1); Benestad et al found that 82% of children with JIA reported pain lasting from 30 minutes to 24 hours daily (2). Pain has been shown to be a significant predictor of impaired psychosocial function in patients with JIA, e.g., reduced participation in school and social activities (1–3), and even minimal reduction in pain has been associated with improvements in quality of life in these patients (4). Results from experimental studies have shown that patients with JIA have reduced pain tolerance and pain threshold compared with healthy controls, and that their pain tolerance is inversely associated with their clinical pain report (5–7). Such findings could suggest that the experience of recurrent pain may affect normal pain processing in these patients.
The significance and negative consequences of pain in children with JIA emphasize the need to further explore possible factors that can predict pain in this group. Several studies have shown that measures of disease activity only explain a modest proportion of the variance in the pain ratings of patients with JIA (3), suggesting that biologic variables are not the only relevant predictors of pain in this group. The amount of pain experienced by a child may be modulated by both psychological and environmental factors independently of the inflammatory disease activity. Possible modulators of pain perception in children may be their use of pain-coping strategies and health beliefs.
Lazarus and Folkman have defined coping as “constantly changing cognitive and behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person” (8). Coping can be differentiated into 2 types: problem-focused coping directed towards modifying the stressor, and emotion-focused coping directed towards modifying the emotional reactions to the stressor. Each of these 2 strategies can be either approach or avoidance oriented.
The use of pain-coping strategies by children with JIA has previously been found to be associated with both clinical and experimental pain report. Schanberg et al (9) thus found that children who rated their ability to control and decrease pain as high and who reported a reduced tendency to catastrophize also reported significantly lower pain intensity, even after controlling for relevant medical and demographic variables. Varni et al (10), using a pain coping questionnaire developed for children, found that the coping strategy of cognitive refocusing was associated with reduced pain, whereas a coping strategy termed “striving to rest and be alone” was associated with increased pain. We have previously reported an association between greater use of approach- and distraction-oriented pain-coping strategies and reduced report of arthritis-related pain (6). In this study we also found reduced tolerance and threshold to and greater intensity of experimental cold pressor pain in both patients with JIA and healthy children who responded with catastrophizing, while greater use of distraction as pain-coping strategy was reported by children who endured experimental pain longer and experienced reduced pain intensity.
Beliefs can be defined as assumptions of reality through which events are interpreted, and beliefs about a stressor such as pain are thought to influence an individual's coping responses (11). Pain-specific beliefs have previously been found to be associated with pain report as well as psychosocial and physical functioning in adult chronic pain patients (11). Pain beliefs of children with JIA have, to the best of our knowledge, not previously been investigated.
The purpose of the present study was to further explore the possible relationships between pain-coping strategies and pain report of children with JIA using daily pain diaries. Diaries are thought to improve recall of minor pain episodes and reduce possible recall bias, and have thus been proposed to be a more valid method than single or retrospective assessments (1). In addition, we wanted to examine whether pain-specific beliefs and coping strategies of patients with JIA independently predict their reported pain, while controlling for disease activity and relevant demographic variables following the method used by Turner et al (11). The objective was thus to determine whether coping predicts pain in patients with JIA after controlling for the effects of beliefs, and whether beliefs predict pain after controlling for the effects of coping.
At the clinic, patients with high pain experience despite low measurable disease activity constitute a challenge for the health provider. We have previously reported (6) that a selected group of patients with JIA with modest clinical arthritis activity, but with daily reports of pain, used less distraction pain-coping strategies than a selected group of patients with significant clinical arthritis activity, but with few complaints of pain. In the present study, we wanted to explore whether it was possible to replicate these findings by comparing the use of pain-coping strategies and pain-related beliefs in a selected subgroup of patients with JIA, e.g., patients with pain experience above the group median and disease activity below the group median (high pain group), with the remaining patients. Theorizing that the pain experience of children with JIA with high pain despite low disease activity partly can be explained by the patient's use of maladaptive coping strategies and health beliefs, we hypothesized that the high pain group would use less distraction and more catastrophizing pain coping strategies, and that they would believe themselves to be more functionally impaired and having less control over their pain.
PATIENTS AND METHODS
Over a period of 6 months, all 70 patients who met the criteria for inclusion were invited to participate. Sixty-seven children with JIA according to the Durban criteria (12) and one of their biologic parents (mother or father) agreed to participate in the study. Three patients refused to participate. The children were recruited during routine visits at the Pediatric Rheumatology Clinic of the Department of Pediatrics, Aarhus University Hospital. Inclusion criteria were a JIA diagnosis, age between 7 and 15 years, lack of comorbidity, and ability to speak fluent Danish. All patients were given written information and verbally informed about the project, and all children and their parents signed an informed consent form. Of the 67 children who participated, 56 children returned a complete 3-week pain diary with 42 pain ratings. The remaining 11 children had incomplete diaries and were excluded from the analysis. The final sample consisted of 45 (80.4%) girls and 11 (19.6%) boys (girl:boy ratio 4.1:1) with a mean ± SD age of 11.4 ± 2.1 years (range 7–15 years). The sex ratio seemed higher than expected from the total patient population in our rheumatology clinic (2.3:1), and in a Nordic population-based study (1.65:1) (13). However, the patients allocated for this study were asked nonselectively for participation. In addition, in a previous study validating the Pain Coping Questionnaire (PCQ) on healthy children, we have not observed any specific differences between the sexes (7). The mean ± SD disease duration was 63 ± 41 months (range 2–158 months). Disease duration was defined as the time from the onset of symptoms to participation in the study. The 56 patients included 7 with systemic-onset JIA, 23 with rheumatoid factor-negative polyarticular subtype, 13 with persistent oligoarticular and 11 with extended oligoarticular subtype, 1 with psoriatic arthritis, and 1 with enthesitis-related arthritis. The study was approved by the local ethics committee.
Measure of disease activity.
A composite arthritis activity score was calculated as the sum of active joint score (zero active joints = 0; 1–2 active joints = 1; 3–4 active joints = 2; >4 active joints = 3), morning stiffness (<15 minutes = 0; 15–30 minutes = 1; 30–60 minutes = 2; >60 minutes = 3), and erythrocyte sedimentation rate (<15 mm/hour = 0; 15–25 mm/hour = 1; 25–40 mm/hour = 2; >40 mm/hour = 3).
Six subscales of the Danish version of the PCQ were used. The PCQ has been preliminary validated on Danish children (14). The children indicated how often (1 = never, 2 = hardly ever, 3 = sometimes, 4 = often, 5 = very often) they used each of 29 coping strategies in response to the prompt, “When I am hurt or in pain for a few hours or days, I… .” The 6 subscales of the Danish version of the PCQ were Seeking Social Support, Positive Self Statements, Behavioral Distraction, Cognitive Distraction, Externalizing, and Catastrophizing. Chronbach's α internal consistency reliability coefficients for each of the subscales were in an acceptable range (0.74–0.85).
A revised version of the Survey of Pain Attitudes (SOPA) called SOPA children's version (SOPA-C) (15) was also used. The SOPA is originally designed to assess pain-related beliefs believed to influence the ways by which patients with adult chronic pain manage their pain (16). The SOPA consists of 8 subscales listing 52 pain-related beliefs. For this study, the subscales of Control, Disability, Harm, Emotion, Solicitude, and Medical Cure were chosen. The subscale of Medication was considered inappropriate in a child population. The items were revised to be understandable for children (negations were avoided, and some of the items were rephrased). Two items considered relevant for children's health beliefs were added; “My parents do not understand how much pain I am in” and “My siblings do not understand how much pain I am in.” The children indicated how much they agreed with statements about their pain problem on a scale from 0 (“This is very untrue for me”) to 4 (“This is very true for me”). The SOPA-C Control scale assesses the patient's belief in having some personal control over pain; the Disability scale assesses the patient's belief in being unable to function because of pain; the Harm scale assesses the patient's belief that pain signifies damage and that exercise and activity therefore should be restricted; the Emotion scale assesses the belief that emotions influence pain; the Solicitude scale assesses the belief that others should respond solicitously to pain behavior; and the Medical Cure scale assesses the belief that a medical cure exists. Chronbach's α internal consistency reliability coefficients were determined for each of the subscales. To be included in the further analysis, it was decided that Chronbach's α should be >0.60. Based on these criteria, 1 item was omitted from the subscale of Harm (“When I am in pain, it usually is a signal that damage has been done”), and 2 items were omitted from the subscale of Emotion (“Nothing but my pain really bothers me” and “My pain has nothing to do with my emotions”), in the final analysis. The reliability coefficients (Chronbach's α) of the subscales were recalculated and considered acceptable (ranging from 0.62 to 0.74).
The children completed a pain diary each day during the 3 weeks following the completion of the questionnaires. The child's pain was measured using The Faces Pain Scale (17) with the endpoints labeled “no pain” and “most pain.” The patients were instructed to measure pain experience every morning and evening in the following 3 weeks. The parents were asked to help the child remember the procedure, but to avoid influencing the child's scoring.
The Childhood Health Assessment Questionnaire (CHAQ), which measures children's functional status, was used. The CHAQ is reported to be reliable and sensitive, and has been validated in a Danish sample (18). The scale assesses performance in 8 areas, including dressing, eating, walking, and grip. Scores range between 0 and 3, with higher scores indicating greater functional impairment.
The assessment instruments were administered in the pediatric outpatient clinic. An experienced pediatric nurse, who administered all questionnaires verbally to ensure that children understood the questions, interviewed children separately from their parents.
The distribution of all data was examined for normality using the Kolmogorow-Smirnov test, and those with a significant result suggesting violation of normality were transformed to normalize them. This was required for the mean pain scores, the CHAQ, the disease severity index, and the coping strategy of catastrophizing, which were normalized by square root transformations. Outliers were inspected and 1 extreme outlier case was omitted from further analyses.
Second-order principal component analyses with varimax rotation were conducted to reduce the number of independent variables. Two analyses were performed, one for the pain coping scales and one for the health belief scales.
Two hierarchical regression analyses with pain as the dependent measure were performed; one with the coping component scores entered last (after the belief components), and one with the belief component scores entered last (after the coping component scores). In each regression model, age, sex, disease duration, disease severity, and CHAQ were entered first in order to control for the associations between these variables and the dependent variable, followed by the coping and belief scores.
The use of coping strategies and health beliefs of children with high pain (median pain diary ≥1.61) and low disease activity (median disease activity <3, high pain children) and the remaining patients were compared with t-tests for independent samples. A Bonferroni procedure was used to control for type I error in the analyses (α = 0.0041, or 0.05 divided by the 12 t-tests). Finally, Pearson's correlation coefficients were calculated to examine the associations between pain and each individual coping and belief measure.
A mean pain score based on the 42 pain ratings was calculated (the children rated their pain in the morning and in the evening over a 3-week period). The patients reported mild to moderate pain (mean ± SD 1.86 ± 8.7; range 1–5.33). Twelve children (21%) reported pain on all days of the 3-week period, and 5 children (8%) reported no pain at all during the 3-week period. Mean pain did not differ significantly between boys (mean ± SD 2.05 ± 0.9, n = 10) and girls (mean ± SD 1.82 ± 0.9, n = 46).
Second-order principal component analysis of the pain coping scales.
A 3-factor solution was found, accounting for 72% of the variance. The first component, which was labeled Distraction, reflects use of behavioral distraction, cognitive distraction, and positive self statements. The second component, labeled Catastrophizing, reflects catastrophizing and seeking social support. The third component, labeled Externalizing, loaded only on the externalizing pain coping strategy. The results are shown in Table 1.
Table 1. Component loadings for principal component analysis of coping scales*
|Positive Self Statements||0.81|| || |
|Cognitive Distraction||0.78|| || |
|Behavioral Distraction||0.72||0.41|| |
|Seeking Social Support|| ||0.79|| |
|Externalizing|| || ||0.94|
Second-order principal component analysis of the health belief scales.
A 2-component solution was found, accounting for 60% of the variance. The first component, which was labeled Cognitive Beliefs, reflects the beliefs that one is not disabled, that pain does not signify harm, that one is able to control the pain, and that a medical cure exists. The second component, labeled Emotional Beliefs, reflects the beliefs that emotions influence pain and that others should respond solicitously to pain behavior (Table 2).
Table 2. Component loadings for principal component analysis of health belief scales*
Coping and health beliefs as predictors of pain report.
The results of the hierarchical regression analysis with pain as the dependent variable are shown in Table 3. Age, sex, disease duration, and disease severity did not make significant contributions to the prediction of pain; however, disability measured by the CHAQ explained an additional 9% of the variance (P < 0.05). When the belief component scores were entered last, the coping component scores explained an additional 15% of the variance (P < 0.05), and the belief component scores then explained an additional 14% of the variance (P < 0.01). When the coping component scores were entered last, the belief component scores explained an additional 20% of the variance (P < 0.002). The coping component scores did not explain a significant additional amount of the variance in pain. Thus beliefs, but not coping strategies, made independent and statistically significant contributions to the predictions of pain after controlling for all other variables in the model. Cognitive Beliefs was the only belief component that contributed significantly to the prediction of pain after adjusting for age, sex, disease duration, disease activity, children's functional status measured by the parents (CHAQ), and all 3 coping components.
Table 3. Hierarchical regression analysis predicting pain*
|1||Demographics and disease measures||0.03||0.03||0.57|| || || |
| ||Age|| || || ||−0.02||−0.14|| |
| ||Sex|| || || ||−0.12||−0.89|| |
| ||Disease duration|| || || ||−0.04||−0.28|| |
| ||Disease activity|| || || ||0.01||0.07|| |
|3||Coping†||0.27||0.15||3.18|| || ||< 0.05|
|4||Beliefs||0.41||0.14||5.35|| || ||< 0.01|
| ||Cognitive|| || || ||−0.44||−3.02||< 0.01|
| ||Emotional|| || || ||0.16||1.19|| |
|3||Beliefs‡||0.32||0.20||7.05|| || ||< 0.002|
|4||Coping||0.41||0.09||2.26|| || || |
| ||Externalizing|| || || ||0.19||1.39|| |
| ||Distraction|| || || ||0.22||1.59|| |
| ||Catastrophizing|| || || ||0.19||1.32|| |
Comparing patients with high pain and low disease activity with the rest of the group.
A subgroup of patients, the high pain group (Group 1, n = 12) with high pain and low disease activity (pain scores equal to or above the median, and disease activity below the median) was compared with the remaining patients (Group 2, n = 44) with respect to pain-coping strategies and health beliefs. After adjustment for multiple comparisons (Bonferroni method), significant differences were found between the scores of Group 1 and Group 2 for the health belief subscale of Disability (mean ± SD 2.0 ± 0.6 and 1.2 ± 0.7, respectively), and for the health belief subscale of Harm (mean ± SD 2.7 ± 0.6 and 1.8 ± 0.7, respectively) (Table 4).
Table 4. Comparison of the scores of the Pain-Coping subscales and the Health Belief subscales*
|Pain-coping measures|| || || || |
| Behavioral distraction||2.4 ± 0.8||2.8 ± 0.9||0.120||–|
| Cognitive distraction||2.9 ± 0.9||3.4 ± 0.9||0.088||–|
| Positive self statements||2.3 ± 0.6||2.9 ± 0.9||0.022||–|
| Seeking social support||3.1 ± 1.1||3.0 ± 0.9||0.631||–|
| Externalizing||1.6 ± 0.7||1.3 ± 0.5||0.071||–|
| Catastrophizing||2.8 ± 1.1||2.1 ± 0.8||0.027||–|
|Health belief measures|| || || || |
| Control||1.8 ± 0.8||2.3 ± 0.7||0.039||–|
| Disability||2.0 ± 0.6||1.2 ± 0.7||0.001||0.02|
| Harm||2.7 ± 0.6||1.8 ± 0.7||0.002||0.05|
| Emotion||1.3 ± 0.9||0.9 ± 0.9||0.241||–|
| Solicitude||1.6 ± 1.2||1.4 ± 0.9||0.518||–|
| Medical Cure||2.6 ± 0.8||2.8 ± 0.7||0.423||–|
Associations between pain and pain-coping strategies and pain beliefs.
Significant correlations were obtained between the pain measure and the pain-coping subscale of catastrophizing (P < 0.05), the pain belief subscales of disability (P < 0.005), harm (P < 0.005), solicitude (P < 0.01), (inverse) control (P < 0.05), and medical cure (P < 0.05) (Table 5).
Table 5. Pearson's correlation coefficients between pain and the psychological variables
|Pain-coping strategies|| || |
| Externalizing||0.25|| |
| Catastrophizing||0.33||< 0.05|
| Cognitive Distraction||−0.02|| |
| Positive Self Statements||−0.19|| |
| Seeking Social Support||0.22|| |
| Behavioral Distraction||−0.16|| |
|Health beliefs|| || |
| Control||−0.30||< 0.05|
| Disability||0.49||< 0.005|
| Harm||0.44||< 0.005|
| Emotion||0.17|| |
| Solicitude||0.35||< 0.01|
| Medical Cure||−0.31||< 0.05|
Pain-related health beliefs have, to the best of our knowledge, not previously been measured in children. In studies with adult chronic pain patients, the health beliefs of disability and harm have been associated with increased physical disability and depression while the health belief of control has been inversely associated with these measures (11, 15, 19). Our results showed that cognitive health beliefs were inversely associated with pain in patients with JIA, even after controlling for age, sex, disease duration, disease activity, parent assessment of the child's health (CHAQ), and pain coping. Cognitive health beliefs reflect beliefs that one is in control over pain, that there exists a medical cure for pain, that one is able to function despite pain, and that exercise and activity should not be restricted because of pain. The coping component did not explain an additional amount of the variance in pain, after controlling for beliefs and all other variables. Previous studies, where health beliefs were not controlled for, have demonstrated that pain-coping strategies were associated with pain in JIA (6, 9, 10). The present investigation raises the possibility that coping strategies are influenced by health beliefs, and that psychosocial treatment of patients with JIA targeting only coping strategies may be less efficient than treatment aimed at modifying pain-related beliefs.
That psychological factors in this study could predict pain variations even when controlling for demographic and disease variables is in disagreement with Hagglund et al (20), who found that social-psychological factors of hopelessness, sadness, and social support did not predict pain among children with JIA. The reason for this discordance could be that pain in these patients may be more influenced by specific pain-related beliefs than by emotions not specifically related to pain.
We have previously found that a group of high pain patients with more pain than expected from their disease activity differed from a group of low pain patients with less pain than expected from their disease activity with respect to use of pain-coping strategies (6). In the previous study, only the 2 selected groups were investigated. In the present study all patients during routine clinic visits in the inclusion period were asked to participate, and the participating patients can thus be considered as representative of the total group of patients at the clinic. In the data analysis, we then identified a group of patients with high pain and low disease activity and compared this group with the remaining patients. The high pain group differed from the rest of the group by perceiving themselves as more disabled and were more likely to believe that pain signifies damage. In adult studies, the belief that one is necessarily disabled by pain has been significantly associated with both psychological and physical dysfunction, and the belief that pain signifies damage has been associated with physical dysfunction (15). A trend towards significant differences between the high pain group and the rest of the patients was found for the coping strategies of positive self statements and catastrophizing. The high pain patients tended to use less positive self statements and more catastrophizing pain-coping strategies. In the previous study, we found that the high pain group selected by the physician used less distraction, but unexpectedly we found no differences between the groups concerning use of catastrophizing. The present results are, however, in concordance with Reid et al (21), who found that externalizing and catastrophizing (emotion-focused avoidance strategies) were positively associated with pain in patients with JIA. In our previous study, the patients completed the questionnaire and rated their arthritis-related pain the same day, but in the present study as well as in the Reid study the children completed the coping questionnaire before they rated their pain in a diary. Catastrophizing may reflect a trait in contrast to distraction strategies, which may be used in response to specific types of pain. Studies with adult patients suggest that catastrophizing is related to personality characteristics such as neuroticism and low dispositional optimism, and that catastrophizing differs from other measures of coping (11).
Our findings that age, sex, disease duration, disease activity, and parent-rated child health (CHAQ) explained only 12% of the variance of the pain the scores, while a combination of coping and health beliefs explained an additional 29% of the variance of the pain scores, support a model of pain experience in JIA where psychological factors are strongly influential.
The clinical implications of our results could be that interventions specifically designed to modify maladaptive pain-related health beliefs and pain-coping strategies may be more effective in reducing pain in children. Also, our results could suggest that it could be advantageous to focus interventions on the subgroup of children where the pain experience seems to be in discordance with the disease activity.
Only few have investigated the possible effect of psychosocial interventions on the pain experience of children with JIA (22, 23). Future intervention research is needed to clarify which methods are most efficient to help the children to manage their pain, and to investigate whether it is possible to identify the subgroup of children who would benefit from interventions.