To analyze patterns of daily pain, stiffness, and fatigue related to juvenile arthritis; to examine the relationships of demographics, disease severity, and psychological adjustment to daily disease symptoms; and to examine daily disease symptoms as predictors of reduced participation in school and social activity.
For a 2-month period, 41 children with polyarticular juvenile arthritis completed daily diaries that included measures of symptoms and function. Children also underwent an initial evaluation and 4 followup evaluations that included a joint count, laboratory testing, and completion of questionnaires assessing physical and psychosocial functioning.
Children reported having pain an average of 73% of days, with the majority of children (76%) reporting pain on >60% of all days. On average, children described the intensity of their daily pain as being in the mild to moderate range; however, a significant subgroup (31%) reported pain in the severe range. Higher physician global assessment ratings, increased functional disability, and increased anxiety were significantly associated with increased daily pain and other daily symptoms. Multilevel random-effects analyses indicated that increased daily symptoms of pain, stiffness, and fatigue were significant predictors of reduced participation in school and social activities.
Physicians should consider treating pain more aggressively in children with arthritis, in order to preserve function in school and social domains, as well as physical function. Moreover, optimal pain management in children with arthritis should include therapeutic regimens addressing anxiety as well as standard pharmacologic interventions.
The course of juvenile polyarticular arthritis is often unpredictable but most frequently is fluctuating, with periods of flare and quiescence. Periods of flare are characterized by increased disease activity and worsening symptoms, such as pain, fatigue, morning stiffness, and increased functional difficulties. Although not well characterized in the pediatric literature, pain is generally considered to be a clinically significant symptom of juvenile polyarticular arthritis, by the children themselves as well as by their parents and siblings (1, 2).
Previous cross-sectional studies have shown that pain and discomfort are prevalent in juvenile polyarticular arthritis (1, 3, 4). Children usually report pain intensity in the mild to moderate range (1, 4–8). However, our research has shown considerable variability in the pain ratings of children with juvenile polyarticular arthritis, with as many as 25% of children reporting pain intensity in the higher ranges on pain measurement scales (5).
Pain and discomfort in children with arthritis can impair function by limiting leisure and physical activities, disrupting school attendance and sports participation, and impeding independent daily hygiene and self-care activities. Indeed, research suggests that children with arthritis are at increased risk for the development of psychosocial problems including depression, decreased self-esteem, decreased scholastic and social competence, and difficulty with peer relationships (9–11). Moreover, children with more severe arthritis or increased functional disability may be at increased risk for adjustment difficulties (9, 12, 13).
Several recent studies used daily diaries to examine pain and associated symptoms in children with chronic conditions such as sickle cell disease, headache, and functional abdominal pain (14–17). These studies revealed a number of advantages of daily diaries over periodic assessments. Diaries allow children to track pain and related symptoms on a day-to-day basis, thereby improving recall of minor pain episodes and reducing recall bias (18, 19). As previously demonstrated in other groups of children with chronic illness, diaries also provide information about the temporal sequencing of events, making it possible to analyze the relationship of pain to behavioral responses such as activity reduction that may place children at increased risk for adjustment difficulties (17, 20). In our previous 7-day diary study of children with juvenile rheumatic disease, we reported the successful use of a diary in children older than age 6 years, and documented daily fluctuations in disease symptoms (21).
The purpose of the present study was to analyze patterns of pain, stiffness, and fatigue related to juvenile polyarticular arthritis, using daily diaries over a 2-month period of time. In addition, we assessed the relationships of demographic variables such as age and sex, measures of disease severity, and psychological adjustment to daily disease symptoms. Finally, we examined daily disease symptoms as predictors of reduced participation in school and social activities.
PATIENTS AND METHODS
From December 1998 through October 2001, children between the ages of 8 and 18 years with a previous diagnosis of polyarticular arthritis were sequentially recruited from the Duke University Medical Center pediatric rheumatology clinic to participate in a daily diary study of arthritis pain and symptoms in juveniles. The final study sample comprised 41 children, representing 73% of patients approached. Of the 15 patients who declined participation, 7 declined because of time constraints, 4 declined because of lack of interest, 2 declined because they disliked the study procedures, and 2 refused to participate for other reasons. Specific exclusions included current use of systemic steroids or psychotropic medications, because the present study was part of a larger study that included use of hormonal measures. The presence of pain or clinically active disease was not required for participation.
Of the 41 children who participated, 24 had polyarticular juvenile rheumatoid arthritis (JRA; 5 were rheumatoid factor positive), 11 had systemic-onset JRA, 3 had psoriatic arthritis, and 3 had spondylarthropathy. The average disease duration for children in the sample was ∼4 years (range 3 months to 12 years). Eighty-eight percent of children were receiving a nonsteroidal antiinflammatory drug (NSAID), 5% were receiving sulfasalazine, 29% were receiving hydroxychloroquine, and 2% were receiving azathioprine. Twenty-seven children (66%) were taking methotrexate at an average dose of 0.34 mg/kg (range 0.13–0.82 mg/kg); of these children, 9 (33%) received the drug subcutaneously. Seven children (17%) were taking etanercept. The mean (± SD) age of the patients was 12.3 ± 2.9 years (range 8–17 years), and 59% were female. Based on self-identification, 85% of the participants (n = 35) were white, and 15% were African American (n = 6); this racial composition is representative of that of the pediatric rheumatology clinic population at Duke University Medical Center.
Eligible patients were initially contacted by phone and were enrolled during a routine clinic visit. Informed consent was obtained from parent(s) and the child, depending on the participant's age, in accordance with the Institutional Review Board at Duke University. Children underwent an initial evaluation and 4 followup evaluations at 2-week intervals during home visits by a study nurse and a research assistant. Initial and followup evaluations included a joint count, laboratory testing, and completion of a series of questionnaires assessing physical and psychosocial functioning. At the initial evaluation, a pediatric rheumatologist evaluated children in the clinic; a study nurse performed the joint counts and venipuncture at followup home visits. Blood drawn at the home visits was kept on ice until it was hand-delivered to the laboratory. In addition, following the initial evaluation, all children completed daily diaries for a 2-month period. The present study includes data collected at the initial and followup evaluations as well as from the daily diaries.
The daily diary included measures of daily symptoms and daily function. At the beginning of the recording period, children were trained by a research assistant in a standard approach to daily recording. Because a key requirement of daily studies is that participants record each day's data at the end of that day (22), children were asked to mail each day's booklet back the following morning and were provided with self-addressed stamped envelopes. During the training session, it was emphasized to both parents and children that diaries be completed without parental intervention. As incentive to increase compliance, children were reimbursed $0.50 for each completed diary and an extra $1.25 for each week in which all 7 daily recordings were returned. The maximum amount children could be reimbursed was $55.00. Reimbursement for compliance is a standard practice in daily diary research (20, 23).
The daily self-report booklet included items in which children rated daily symptoms, including pain intensity, painful locations, level of fatigue, and stiffness. Pain, fatigue, and stiffness were measured using a 100-mm visual analog scale (VAS). The pain scale from the Pediatric Pain Questionnaire (PPQ) (6) was used. The PPQ pain VAS is anchored at the end points by “no hurting, no discomfort, no pain” and “hurting a whole lot, very uncomfortable, severe pain.” In addition to rating their daily pain intensity, children were asked to color in or mark a body map showing painful locations. The pain location body map was adapted from the PPQ. Both the VAS and the body map have been shown to have adequate reliability and validity in the assessment of pain in children with rheumatic disease (1, 2, 6, 24). The 100-mm visual analog scales for fatigue and stiffness were modeled after the VAS for pain. The words “not tired” and “very tired” anchored the end points of the VAS for fatigue, and the VAS for stiffness was anchored by “no stiffness” and “very stiff.”
Daily activity reduction.
In the daily diaries, children were also asked to indicate the amount they reduced school and social activities each day. Ratings were on a 4-point Likert scale anchored by “not at all” and “a lot,” with higher values indicating greater reduction.
Baseline functional status.
Parents and children rated the child's disease-related functional status using the parent-report and the child-report versions, respectively, of the Childhood Health Assessment Questionnaire (C-HAQ) (25). Both versions of the C-HAQ yield a disability index score and a discomfort index score. The reliability and validity of this instrument in polyarticular JRA were previously established (25). The C-HAQ disability index has been used to record changes in functional disability (26).
Baseline disease severity.
Disease severity was assessed at the initial evaluation using a variety of sources, including the physician's rating, active joint count, and laboratory parameters. First, the pediatric rheumatologist provided a global assessment of the child's disease severity, using a 100-mm VAS anchored by the end points “asymptomatic” and “very severe.” Physician global assessment has been used by multiple researchers as a measure of disease activity for children and adults with rheumatic disease. In 1997, it was designated as part of the core set of outcome variables for clinical studies of juvenile arthritis (27).
The joint count was recorded, and consisted of the number of swollen and tender joints, as well as the number of joints in which range of motion was limited, as previously described (28–30). From this information, the number of active joints for each child was determined. A joint was defined as active if it met the following criteria: 1) swelling alone, or 2) a combination of joint tenderness and limitation of movement. Active joint count has been used previously as a measure of JRA disease activity (28–30).
A laboratory measure of disease activity, the erythrocyte sedimentation rate (ESR), was obtained at the initial and followup evaluations. The ESR has been shown to fluctuate with JRA disease activity (31, 32) and has been used as an outcome measure in multiple controlled trials of drug therapy in JRA (28). Analysis was performed by the Duke Clinical Laboratory.
Finally, children were classified as having either ongoing active disease or minimal disease activity, based on the measures of disease severity at the initial evaluation. Children were classified as having ongoing active disease if they met at least 1 of the following 3 criteria: physician global assessment rating >25, active joint count of ≥5, and ESR ≥30 mm/hour. Children not meeting any of these criteria at the initial evaluation were classified as having minimal disease activity.
Children who experienced a disease flare during the study period were identified using data collected at the initial and followup evaluations. Children were identified as having a flare if they demonstrated 30% worsening in at least 3 of the following 5 criteria during the study period: discomfort index of the C-HAQ, number of active joints, number of active joints with loss of motion, disability index of the C-HAQ, and ESR. In addition, there could not be >30% improvement in more than 1 of the outlined criteria, and the active joint count must have increased by at least 2. Our definition of disease flare was modified from the JRA response criteria and represents 4 of the 6 response variables (27, 33).
Baseline psychological adjustment.
The Children's Depression Inventory (CDI) (34) and the Revised Children's Manifest Anxiety Scale (RCMAS) (35) were used to assess the psychological adjustment of the child. The CDI is a 27-item self-report scale based on the Beck Depression Inventory (36), which assesses a variety of depressive symptoms including sleep disturbance, appetite loss, and dysphoria. A total scale score can be derived as a standard T score with a mean of 50 and a standard deviation of 10. The CDI is the most widely used and validated measure of childhood depression in children ages 8 to 17 years (37).
The RCMAS is a 37-item self-report scale designed to assess the level and nature of anxiety in children and adolescents ages 6 to 19 years across 4 areas: physiologic anxiety (e.g., difficulty breathing, difficulty sleeping, and abdominal pain), worry (e.g., anticipatory anxiety), social concerns/concentration (e.g., worries about the expectations of significant others), and total anxiety. Raw scores are converted to scaled scores based on children's age and sex. For the total anxiety score, the scale score is a standard T score with a mean of 50 and a standard deviation of 10; for the subscales, the scaled scores have a mean of 10 and a standard deviation of 3. Higher scores are indicative of increased anxiety. For the present study, scaled scores for the total anxiety scale, physiologic anxiety subscale, worry/oversensitivity subscale, and social concerns subscale were used in data analysis. Adequate reliability and validity data support use of this instrument in school-age children (38).
First, we examined daily pain, stiffness, fatigue, and activity reduction during the 2-month period. For these analyses, we evaluated patterns on those days on which children reported that they experienced pain (pain days). Second, we assessed the relationships of demographic variables, disease severity variables, functional status, and psychological adjustment to pain, stiffness, fatigue, and activity reduction. Finally, we analyzed daily associations between pain, stiffness, fatigue, and activity reduction using multilevel random-effects models to examine same-day within-subject relationships between symptoms and symptom response.
Daily diary completion.
All children were asked to complete the diary on a daily basis over an 8-week time period. The actual time period for study participation ranged from 34 days to 73 days, with children participating an average of 57 days. In total, there were 2,346 potential diary days, and children returned diaries on 2,095 days (89% overall completion). Completion rates for individual children ranged from 42% to 100%, with 30 children (73%) completing diaries at a rate of 89% or higher. This suggests that school-age children are capable of completing and returning daily diaries over a 2-month time period.
Psychological adjustment factors were significantly related to completion rates. Specifically, completion rates were lower in children with increased depressed mood (r = −0.55, P < 0.01) and increased anxiety as assessed by the total anxiety score and the social concerns/concentration subscale score on the RCMAS (r = −0.37, P < 0.05, and r = −0.52, P < 0.01, respectively). The completion rate was unrelated to baseline demographics, functional status, and disease severity.
Pain, stiffness, and fatigue on pain days.
A total of 2,095 diary days were recorded in this study, including 1,533 pain days (73%). Sixteen children (39%) reported pain on 100% of diary days, and 2 children (5%) reported no pain during the 2-month time period. The majority of children in the study (n = 31; 76%) reported pain on >60% of days. The mean (± SD) pain intensity rating on a pain day was 36.6 ± 24.5 (range 3–100) on the 100-mm VAS, but 31% of children reported average ratings of >40. The mean (± SD) number of painful joints reported by children on pain days was 8.1 ± 12.6. On pain days, the average stiffness rating on the 100-mm VAS was 35.6 ± 27.4, and the average level of fatigue was 37.0 ± 27.9 (Table 1 and Figure 1).
Table 1. Average daily symptom levels and activity reduction on pain days and according to disease activity grouping*
Minimal disease activity (n = 17)
Ongoing active disease (n = 24)
Values are the mean ± SD.
P < 0.05, children with minimal disease activity versus children with ongoing active disease.
Relationships with demographics, disease severity, functional status, and psychological adjustment.
Independent sample t-tests, general linear models, and Pearson's product-moment correlations were used to investigate the relationships between demographics, disease severity, functional status, psychological adjustment, and daily symptom ratings on days when children reported pain.
Demographic variables. Both child sex and age were unrelated to levels of daily pain, pain locations, stiffness, and fatigue. However, there was a trend toward significance in the relationship between sex and percentage of pain days, with girls reporting pain on a higher percentage of days compared with boys (t = −1.75, P = 0.09). Girls on average reported pain on 82% of diary days, and boys reported pain on 62% of days. Results of a power analysis indicated that increasing our sample size by 7 children would have given us sufficient statistical power (0.80) for these results to have achieved statistical significance at a value of α = 0.05.
Disease severity. At the initial evaluation, children were rated by the pediatric rheumatologist as having, on average, mild to moderate disease severity as assessed by the physician global assessment (mean ± SD 33.9 ± 28.2, range 0–100). The mean ± SD number of active joints at the initial evaluation was 10.3 ± 11.0 (range 0–43). The ESRs ranged from 1 mm/hour to 97 mm/hour, with a mean (± SD) ESR of 26.4 ± 28.6 mm/hour. Based on measures of disease severity at the initial evaluation, 24 children (59%) were classified as having ongoing active disease, and 17 children (41%) were classified as having minimal disease activity. Additionally, 8 children met our criteria for disease flare during the 2-month study period.
Baseline disease severity variables related to daily symptoms included the physician global assessment rating, active joint count, and disease classification. Specifically, higher baseline disease severity as rated by the physician and a higher number of active joints were significantly correlated with increased daily pain (r = 0.39, P < 0.05 and r = 0.36, P < 0.05) and higher levels of daily stiffness (r = 0.44, P < 0.01 and r = 0.40, P = 0.01). Additionally, the percentage of pain days reported by children classified as having ongoing active disease (85%) was higher than that reported by children with minimal disease activity (58%). On pain days, children with ongoing active disease reported higher average pain ratings (t = −2.1, P < 0.05) and increased stiffness (t = −2.7, P = 0.01) (Table 1). Daily symptoms were unrelated to the baseline ESR.
Functional status. Children were rated as having, on average, mild to moderate functional limitations. The mean (± SD) disability index scores on the child-report and parent-report versions of the C-HAQ were 0.74 ± 0.7 and 0.75 ± 0.7, respectively. Results indicated that decreased function was related to increased daily symptoms. First, increased functional disability as assessed by the disability index of the child-report and parent-report versions of the C-HAQ was significantly correlated with increased daily pain (r = 0.44, P < 0.01 and r = 0.42, P < 0.01, respectively) and increased stiffness (r = 0.40, P = 0.01 and r = 0.48, P < 0.01, respectively). Children with greater functional impairment also reported pain on a higher percentage of days (r = 0.42, P < 0.01). Increased parent-rated disability was also significantly correlated with a greater number of painful locations (r = 0.40, P = 0.01).
Psychological adjustment. The mean (± SD) T score on the CDI was 47.4 ± 8.8. Two children (5%) were considered to have clinically significant elevations on the CDI, with T scores ≥60. The mean (± SD) T score for the total scale of the RCMAS was 46.5 ± 9.8. The mean (± SD) scaled scores were 9.4 ± 3.0 for physiologic anxiety, 8.4 ± 2.9 for worry, and 8.4 ± 3.1 for social concerns. Four children (10%) were considered to have clinically significant elevations on the total anxiety scale of the RCMAS, with T scores ≥60. Five children (12%) had elevated scaled scores (score ≥ 13) on the physiologic anxiety scale, 4 (10%) had elevated scaled scores on the worry scale, and 3 children (7%) had elevated scaled scores on the social concerns scale.
Increased anxiety in a child was related to increases in daily symptoms. In particular, a higher score on the physiologic anxiety subscale of the RCMAS was related to increased daily pain ratings (r = 0.38, P < 0.05) and increased daily fatigue (r = 0.41, P < 0.01). In addition, children with higher anxiety as assessed by the total anxiety scale, physiologic anxiety subscale, worry/oversensitivity subscale, and social concerns subscale reported pain on a higher percentage of days (r = 0.43, P < 0.01, r = 0.36, P < 0.05, r = 0.39, P < 0.05, and r = 0.33, P < 0.05, respectively). Depressed mood in the patient was unrelated to daily disease symptoms.
During the study period, 56% of the children reduced their school activities, on an average of 10% of pain days (range 0–73% of pain days). Forty-one percent of the children (n = 17) did not reduce school activities while they were experiencing pain, whereas 12% (n = 5) reduced school activities on >40% of pain days. Sixty-six percent of children (n = 27) reduced their social activities at least 1 day during the study period. On pain days, children reduced their social activities an average of 20% of pain days. Thirty-two percent of children (n = 13) did not reduce their social activities at all while they were experiencing pain, and 12% (n = 5) reduced their social activities on >40% of pain days.
Relationship between activity reduction and demographics, disease severity, functional status, and psychological adjustment.
Analyses indicated that demographic variables were unrelated to activity reduction. Activity reduction was also not related to the baseline physician global assessment rating, the active joint count, disease activity classification, functional status, depression, or anxiety. Only increased ESR was significantly correlated with an increased reduction of school activities (r = 0.44, P = 0.01).
Patterns of activity reduction and daily symptoms.
We used multilevel mixed-effects models to analyze patterns of disease symptoms and activity reduction, including school and social activities. This strategy is advantageous because it controls for between-person effects when assessing the effects of within-person factors, treats participants as random factors (which increases the generalizability of the results), and controls for autocorrelation (serial dependency) between reports (39–41).
Multilevel random-effects analyses indicated that increased daily symptoms, specifically pain intensity, number of painful locations, stiffness, and fatigue, were significant predictors of reductions in school and social activities (Table 2). Children were more likely to reduce school activities on days when they experienced higher levels of pain and increased stiffness. They were more likely to reduce social activities on days when they experienced higher levels of pain, more painful locations, increased stiffness, and more fatigue (Table 2).
Table 2. Summary of multilevel random-effects analyses predicting activity reduction because of daily symptoms*
Children participating in the present study reported pain in the mild to moderate intensity range on a significant proportion of days (mean 73%). Moreover, using a conservative estimate, the incidence of flare in the present sample was higher than expected, with at least 20% of children experiencing a disease flare during the 2-month study period. These frequencies of reported pain and flare are notable, given that these children were receiving subspecialty care at a major medical center and displayed wide variation in disease activity.
The clinical significance of pain is underscored by the observed relationships between pain and activity reduction. Specifically, children were more likely to reduce their level of participation in school and social activities on days when they experienced increases in pain, stiffness, and other symptoms. Our results suggest that monitoring of daily fluctuations in pain and symptoms may be more important than use of physician global assessment, the active joint count, and laboratory parameters in identifying children at risk for activity reduction. Future research is needed to better identify children at risk for activity reduction. Important predictive factors may include parental health beliefs, parent and child coping, parent emotional state, child stressful events, child mood, and parent and child demographic variables such as level of education (21, 42–45).
Child anxiety was significantly associated with increased pain frequency, pain intensity, and fatigue. Interestingly, ∼10% of children were considered to have clinically significant elevations in anxiety at baseline. Clinically elevated depressed mood was observed in only 5% of children, and increased depressed mood did not correlate with disease symptoms. Thus, it appears that anxiety rather than depression may be a clinically significant problem for many of these children. To date, studies examining adjustments made by children with arthritis have rarely included measures of anxiety but instead focused on measures of depression and self-esteem (9–11). Moreover, rheumatologists may be more comfortable and skilled at identifying depression rather than anxiety. To optimize care of children with arthritis, researchers and clinicians may need to screen for anxiety and consider treatment of anxiety using cognitive-behavior therapy, massage therapy, or medications as an effective way to improve pain management as well as school and social functioning (46–48). In addition, longitudinal studies need to be conducted to examine patterns of change in children's adjustment and levels of anxiety.
One important limitation of the present study was the predominance of white children. Although the preponderance of white children reflected the composition of the pediatric rheumatology clinic population at our institution, future studies are needed to determine whether these results would be consistent across more racially and ethnically diverse clinical populations, especially given the well-documented ethnic and sex differences in pain perception (49, 50). In addition, future studies should address whether these findings are unique to children with arthritis, by including healthy children as well as children with other chronic diseases for purposes of comparison. Finally, because of our relatively small sample size, we did not adjust the P values to correct for the number of statistical tests conducted.
It is also important to note that the frequency of pain and disease flare reported in the present study may be higher than more current rates, given the widespread use of anti–tumor necrosis factor therapy as well as other biologic agents that became available subsequent to the beginning of study enrollment. Indeed, when we compared patients receiving etanercept plus methotrexate (n = 7) with those receiving methotrexate alone (n = 21), we observed lower mean daily pain intensity, fatigue, stiffness, and painful locations in children taking etanercept, even after controlling for disease activity. Only the results for mean daily level of fatigue reached statistical significance, due to the small number of participants in each group.
In conclusion, clinic-based assessments of pain in children with juvenile arthritis fail to capture the chronic, pervasive nature of the problem. Combined with our previous findings, the present results demonstrate that the diary method is feasible and well accepted by children with polyarticular arthritis and their parents and may provide a more complete picture of the symptoms experienced by children with this disease. Physicians should consider treating pain more aggressively in children with arthritis, using a variety of pharmacologic therapies including NSAIDs, acetaminophen, and other analgesics, and remittive agents to preserve function in school and social domains as well as physical function. Our data also suggest that standard pharmacologic management of pain in children with juvenile polyarticular arthritis might be optimized by the addition of therapeutic interventions to treat anxiety, including psychotropic medication and cognitive-behavioral therapy.
The authors thank all of the staff and families at the Pediatric Rheumatology department at Duke University Medical Center for their cooperation and participation in this research. We also thank Janet Wooten and Lisa Macharoni for assistance in data collection and management.