|Home | About | Journals | Submit | Contact Us | Français|
Juvenile fibromyalgia (JFM) is a chronic musculoskeletal pain condition that is associated with reduced physical function. Recent research has demonstrated that cognitive-behavioral therapy (CBT) is effective in improving daily functioning among adolescents with JFM. However, it is not known whether these improvements were accompanied by increased physical activity levels.
To analyze secondary data from a randomized clinical trial of CBT to examine if CBT was associated with improvement in objectively measured physical activity and whether actigraphy indices corresponded with self-reported functioning among adolescents with JFM.
Participants were 114 adolescents (ages 11–18) recruited from pediatric rheumatology clinics that met criteria for JFM and were enrolled in a clinical trial. Subjects were randomly (1:1) assigned to receive either CBT or fibromyalgia education (FE). Participants wore a hip-mounted accelerometer for one week as part of their baseline and post-treatment assessments.
The final sample included 68 subjects (94% female; mean age = 15.2 years) for whom complete actigraphy data was obtained. Actigraphy measures were not found to correspond with self-reported improvements in functioning. While self-reported functioning improved in the CBT condition compared to FE, no significant changes were seen in either group for activity counts, sedentary, moderate or vigorous activity. The CBT group had significantly lower peak and light activity at post-treatment.
Actigraphy monitoring provides a unique source of information about patient outcomes. CBT intervention was not associated with increased physical activity in adolescents with JFM indicating that combining CBT with interventions to increase physical activity may enhance treatment effects.
Juvenile fibromyalgia (JFM) is a chronic musculoskeletal pain condition affecting 2–6% of school-age children [1–3], diagnosed primarily in adolescent females. JFM is characterized by widespread pain, sleep difficulty, fatigue and numerous other associated symptoms . Several studies have documented high levels of physical, emotional and social impairment in adolescent JFM patients [4–6] that persist over time . Current treatment approaches for JFM include medications to manage pain and improve sleep, and recommendations for physical activity and aerobic exercise and/or cognitive-behavioral therapy to improve daily functioning and coping.
The benefits of aerobic exercise on managing fibromyalgia pain have been documented in several studies [8–11], which have shown that low to moderate intensity aerobic exercise improved pain, fatigue, sleep, and mood in adults with fibromyalgia (10–11). However, the effects of aerobic training on reducing adult fibromyalgia pain and fatigue remain inconsistent (8). Similarly for children with JFM, aerobic exercise has been shown to improve physical functioning and quality of life but did not result in incremental improvements in pain or symptom severity compared to anaerobic exercise . Current guidelines are for patients with fibromyalgia to engage in moderate exercise 3–4 times per week . Unfortunately, long-term adherence to exercise recommendations is notoriously poor  and the majority of patients with fibromyalgia, including adolescents, are very sedentary [15, 16]. There is evidence that cognitive-behavioral therapy (CBT) focused on teaching psychological pain management skills is effective in reducing functional disability in adolescents with JFM [17, 18]. Adolescents who received CBT reported greatly improved participation in school, social and daily activities at the end of treatment. Moreover, CBT was found to have lasting effects, with improvements in functioning maintained 6-months following the end of active treatment . It is currently not known whether these improvements in functioning also translated to an overall increase in actual physical activity levels. Even though CBT does not directly target increased participation in more vigorous types of physical activity such as aerobic exercise, it is conceivable that with better coping skills after CBT and significantly lower depressive symptoms, adolescents may become less sedentary and begin to engage in more physical activity .
To our knowledge, no studies so far have examined whether CBT has any effect on objectively measured physical activity in patients with JFM. Physical activity monitoring via actigraphy is a sensitive and ecologically valid measure of daily activity, and recent studies have described significantly reduced activity levels in children and adolescents with chronic pain [16, 21]. However, actigraphy remains underutilized as an objective outcome measure and there is relatively little information about how the various activity indices obtained (such as daily activity counts, peak activity, time spent in vigorous, moderate, light and sedentary activity) perform in the context of intervention studies. While self-reported measures of functioning are generally considered the gold-standard in reporting outcomes of clinical trials in chronic pain for both adults and children [22, 23], some domains of functioning, such as engagement in physical activity can be influenced by reporting biases [24, 25]. In such instances, use of objective measures can be a very useful complement to self-report instruments.
In our recently completed randomized clinical trial of CBT for adolescents with JFM, objective physical activity assessment (using hip mounted activity monitors worn for one week) was included as part of the comprehensive assessment at baseline and post-treatment. In this trial, patients were randomized to either CBT or an education/support control condition (fibromyalgia education, FE), in which the CBT group received training in active and adaptive coping strategies and the FE group received general information about fibromyalgia and guidelines in healthy habits. Neither treatment condition provided direct intervention to increase physical activity or participate in more exercise. Actigraphy was used as a secondary outcome measure in this trial and was included in order to examine how objectively measured physical activity outcomes corresponded with changes in self-reported daily functioning and depressive symptoms by participants.
The specific aims of this study were to 1) examine changes in objectively measured physical activity indices (daily activity counts, peak activity and time spent in sedentary, light, moderate and vigorous activity) within the context of a clinical trial to examine whether CBT resulted in any changes to physical activity as compared to FE and 2) to examine whether objective activity measurement before and after treatment was significantly associated with self-report of physical function. Lastly, we explored the relation between objective physical activity levels before and after treatment with self-report of depressed mood. Based upon the prior findings of improved daily functioning by self-report, we hypothesized that adolescents who received CBT would spend less time in sedentary activity and show increased daily activity counts after treatment compared to those who received FE. No changes in vigorous activity were expected since increasing aerobic exercise was not a target of the behavioral treatment per se. For the FE group, we expected that activity measures would remain stable from pre- to post-treatment given that no specific instruction was provided in the FE condition about how to increase physical activity.
Adolescents (ages 11–18) with JFM were recruited for a clinical trial of cognitive-behavioral therapy (CBT) from four pediatric rheumatology clinics in the Ohio and Kentucky region. Participants were eligible if they met the following inclusion criteria: 1) met Yunus and Masi  criteria for JFM, 2) were on stable medications for 8 weeks, 3) reported average pain intensity ≥ 4 on a 0–10 cm Visual Analog Scale based on one week of daily pain diaries, and 4) reported at least mild impairment in daily activities due to JFM symptoms (i.e., score >7 on the Functional Disability Inventory). Adolescents were excluded from enrollment if 1) they were diagnosed with any other chronic rheumatic disease, 2) had a documented developmental delay, 3) experienced current symptoms of bipolar disorder, major depressive disorder, panic disorder, or psychosis based on a standard psychiatric interview , or 4) were taking opioid medications.
Of the 114 participants enrolled in the clinical trial, 68 participants who had complete actigraphy data for both the pre- and post-treatment evaluations were included in this study (see Figure 1 for study flowchart). A complete set of actigraphy data was defined as having both a pre- and post-treatment evaluation, actigraphy monitor worn for at least 5 days, and < 5 consecutive hours of missing data per day during daytime hours. The study was approved by each hospital’s Institutional Review Board and written informed consent was obtained from patients and parents.
Actigraphy assessments were conducted at pre-treatment (Week 1) and post-treatment (Week 9) as part of a battery of outcome measures for the primary clinical trial. Participants were mailed the actigraphy monitor and provided with phone and written instructions prior to their pre- and post-treatment study visits. They were asked to wear the actigraph at all times (except while showering/bathing) for one full week prior to each of the evaluation visits. After the pre-treatment assessment, adolescents with JFM were randomly assigned to receive 8 individual weekly CBT sessions (active treatment) or 8 supportive fibromyalgia education (FE) sessions (attention-control condition) with a trained doctoral level therapist. Treatment assignment was made by a biostatistician based upon a pre-determined randomization schedule. The study employed a single-blind method whereby investigators, study physicians, study coordinator, and assessment staff remained blinded to the participants’ treatment condition throughout the trial. Only the therapists delivering the intervention and the patients themselves were aware of the treatment condition. Details of the study methodology and primary outcomes of the trial have been previously published . CBT sessions were focused on teaching the adolescents pain coping skills such as relaxation, distraction, pacing, cognitive reframing and problem solving. Parents were included in 3 of the 8 sessions and provided parental guidelines for behavior management. FE sessions were focused on providing information about JFM, its treatment, and discussion of healthy lifestyle habits. Neither of the treatment arms included specific instruction about how to increase exercise or physical activity, although all patients received recommendations from their primary rheumatologist to increase aerobic activity as part of their usual medical care.
Parents completed a demographic information form including the adolescents’ age, sex, race and ethnicity.
Small, hip-mounted, actigraphy monitors  were utilized for continuous activity monitoring over a period of one week (7 days including weekdays and the weekend) which is consistent with best practice guidelines for actigraphy research . Because our primary interest was in daytime physical activity and to ensure consistency of time intervals for data analysis across participants we selected a predetermined 16-hour monitoring period (6 am – 10 pm) which is appropriate for this age group and also aligned with published guidelines . Activity counts were expressed in 1 minute epochs for ease of description and analysis. The 5 most complete days of data were used for all analyses and a “complete” day was defined as ≥80% of monitor wearing time. Average activity counts per minute during daytime hours, peak activity, and average time (in minutes) spent in sedentary, light, moderate and vigorous activity per day during daytime hours were calculated for each participant using Actical Analysis Software 2.0 . The cut-points used for sedentary, light, moderate and vigorous activity were based on standard settings provided in the software program which uses an algorithm for average energy expenditure (kcal/min/kg) specifically for the adolescent age range (0 = sedentary; 0 – 0.01 light; 0.01–0.05 moderate; >0.05 vigorous). As in our previous actigraphy research  definitions of average activity counts and peak activity were as follows: average activity counts per day represented an average of activity counts across the 5 complete days (not necessarily consecutive) during daytime hours (6 am – 10 pm) and peak activity was defined as the highest level of physical activity per minute the study participant engaged in during a 5-minute period. Examples of sedentary activity include sleeping or sitting and watching television; light activity included playing a video game or typing on a computer; moderate activity included light chores (vacuuming or gardening), or recreation (playing catch); and vigorous activity included walking, jogging, skipping, or running.
A 15-item self-report instrument was used to assess adolescents’ perceived difficulty to perform daily activities in home, school, recreational, and social settings . Participants rated how much difficulty they had performing each of the activities on a 5-point Likert scale (0 = no trouble to 4 = impossible). Total scores on the FDI range from 0 to 60, with higher scores indicating greater disability. The FDI has been found to be well-validated with high internal consistency, moderate to high test-retest reliability, moderate cross-informant (parent-child) reliability, and good predictive validity [29–31].
A 27-item self-report measure was used to assess adolescents’ depressive symptoms. The CDI has been validated for use in children and adolescents ages 7–17 . It has strong psychometric properties and is frequently used in pediatric pain research [33–35]. Participants select one of three statements for each item. In this study, the total raw score was used as an overall indicator of severity of depressive symptoms.
Prior to conducting the primary analysis, reasons for missing data were examined and baseline characteristics of participants for whom complete actigraphy data were obtained versus those who did not have complete data were compared to ensure that there was no systematic source of bias. Descriptive statistics for the variables of interest were examined for normality. To assess whether there were significant changes in physical activity before and after treatment for each group, pre- and post-treatment means were compared for the CBT and the FE groups separately using t-tests for dependent samples (or paired-sample Wilcoxon Signed Rank test for non-normal data) for each of the activity indices (activity counts, peak activity, and average time in sedentary, light, moderate and vigorous activity), as well as for the primary self-report outcomes of functional disability (FDI) and depressive symptoms (CDI). Correlations between activity indices, FDI scores, and CDI scores were calculated for both groups at pre- and post-treatment to examine the correspondence between objective assessments of activity and self-report of functioning as well as depressed mood. Finally, correlations between activity indices before and after treatment were computed for those in the CBT and FE groups to explore the stability/reliability of actigraphy measures over time with the expectation that higher correlations would be found in the FE group which did not receive active intervention.
Of the 114 participants enrolled in the trial, 105 wore the actigraphy monitor at baseline (reasons for the 9 missing were − 1 patient had a nickel allergy, 3 equipment malfunctions, 2 did not receive the monitors by mail on time and 2 did not wear). Of the 105 participants, 81 of them also wore the monitor at post-treatment (reasons for 24 missing: 8 dropped out of the study (<15% drop-out rate), 3 did not receive the monitors by mail on time, 2 equipment malfunctions, 5 did not wear and 6 additional patients wore the actigraph but were excluded from analyses because of insufficient data from baseline actigraphy). As can be seen in Figure 1, the final sample consisted of 68 participants for whom complete data was available at both time-points. We compared the final study sample (N = 68) with the group of participants who had incomplete/missing data (N = 37) and found no significant differences in demographics or baseline characteristics of functional disability (t=1.52, p = 0.13), pain intensity (t= −0.35, p = 0.73) or depressive symptoms (t = −0.17, p = 0.86) between the groups. The majority of adolescents in the final sample were female (94%) and Caucasian (88%), with a mean age of 15.2 years (SD = 1.6) which is consistent with the demographics of the larger clinical trial sample. The average number of daytime hours of actigraphy data per day was 14.9 (out of 16; 93%), which is above the 80% mark as recommended by Ward and colleagues . Additionally, 97% of patients in the sample had at least one weekend day of complete data and 24% of the sample completed actigraphy assessment during the summer months. Of the 68 participants who completed actigraphy at pre- and post-treatment, 33 adolescents received CBT and 35 received FE.
Means and standard deviations for activity levels pre- and post-treatment for each of the treatment conditions can be found in Table 1. Indices of higher activity levels (e.g., peak activity and average time in vigorous activity) demonstrated significant skew and kurtosis (i.e., z-score > ± 1.64; ; thus, non-parametric tests (paired-sample Wilcoxon Signed Rank test) were computed for these indices. Overall, patients were very sedentary prior to treatment, with < 3% of them meeting recommended guidelines of at least 60 minutes of physical activity . For the FE group, as expected, there were no significant differences in actigraphy indices (time spent in sedentary, light, moderate or vigorous activity, activity counts or peak activity) from pre- to post-treatment. Contrary to expectations, there were no changes in the CBT group for time spent in sedentary activity or average activity counts. In fact, there was a slight (non-significant) trend for increased time in sedentary activity and lower activity counts after treatment for the CBT group. There were also significant decreases in peak activity level (z = −2.46; p = .014) and time spent in light activity (t = 2.38; p = .024) for the CBT group from pre- to post-treatment, along with trends towards lower moderate activity. Following treatment, 97% of patients continued to fall below the recommended guidelines of physical activity. With respect to self-report of impairment, disability (FDI) scores were significantly improved (lower scores) for the CBT group (t = 2.49; p = .018) and unchanged for the FE group (t = −0.92; ns).
Correlations between all activity indices and self-report of functioning (FDI scores) were low at pre- and post-treatment for both CBT and FE groups (Pearson r’s ranging from −0.30 to 0.22, p = ns; with the majority of correlations close to zero) indicating very low concordance between the physical activity measures and self-report of functioning. Similarly, correlations between all activity indices and self-report of depressive symptoms (CDI scores) were not found to be significant at pre- and post-treatment for both groups (Pearson r’s ranging from 0.02 to 0.15, p = ns).
Correlations of activity indices before and after treatment (Table 2) were found to be significant across time points for CBT and FE groups (Pearson r ranging from .47 to .67, with the exception of peak activity in the FE group, r = .22). For both groups, time spent in moderate activity and vigorous activity were the most stable measures (Pearson rs = .56 – .67). For the FE group, average activity counts before and after treatment were also moderately high (Pearson r = .66) whereas these correlations were lower for the CBT group (Pearson r = .47) reflecting the somewhat reduced activity levels in this group after treatment. Findings on peak activity were more difficult to interpret as they showed very low stability in the FE group (Pearson r = .22).
The results of this study confirm that objective monitoring of physical activity provides a unique and important source of information that is quite distinct from self-report in studies of adolescents with chronic pain. Overall, we found very low engagement in moderate-vigorous activity with over 95% of patients not meeting recommended guidelines of physical activity. Behavioral intervention trials for adolescents with JFM, including our recently completed randomized trial, have consistently found that CBT leads to significant improvements in self-report of daily physical functioning and overall well-being. However, in taking a closer look at actual physical activity levels using objective actigraphy measurement, these improvements do not appear to translate into increased engagement in physical activity. In fact, the results trended in the opposite direction. Even though they reported significant improvement in their functioning, adolescent JFM patients who received CBT had slightly lower activity levels at the end of treatment, with significant reduction in peak activity. Patients who received FE did not show any significant change in activity levels at the end of treatment. Possible reasons for why CBT may have had this paradoxical effect in teenagers with JFM are that 1) the training in pain coping skills emphasizes strategies that are mainly cognitive in nature (such as distraction, challenging negative thoughts, problem solving) and 2) the behavioral training includes relaxation, pacing and other skills that do not specifically target increased physical activity. The use of activity pacing is intended to allow greater participation in activity without causing pain flare-ups but this may have inadvertently resulted in reduction in vigorous activity by the adolescents (e.g., taking longer rest breaks to pace themselves and avoid a pain-flare) without an overall increase in behavioral activation (i.e., engaging in more activities between rest breaks), as was intended. Contrary to past studies , there was no relation between self-report of depressed mood and objective physical activity levels in this cohort of JFM patients. In sum, use of CBT strategies may have resulted in improved psychological coping in daily life but slightly reduced overall activity levels. The results support the notion that a sedentary lifestyle and activity avoidance in JFM patients is perhaps more entrenched and could benefit from additional direct intervention by using targeted behavioral activation training or physical exercise interventions to enhance the otherwise strongly positive effects of CBT. In addition, educating youth with JFM about what constitutes moderate to vigorous activity may enhance their knowledge about how to implement recommendations for increasing physical activity. It is clear that aerobic exercise  and strength training  yield improvements in pain and physical functioning for youth and adults with fibromyalgia. Although results from this study indicate that CBT alone does not result in increased engagement in physical activity, cognitive-behavioral strategies can help promote adherence and enhance motivation to physical activity and exercise programs  to help achieve recommended guidelines for physical activity. Studies combining CBT with behavioral activation and/or exercise training have been conducted in adults with fibromyalgia but have not yet been tested in adolescents with JFM [14, 40].
Aside from the main results discussed above, this study documented some of the challenges of actigraphy monitoring in a clinical trial. Despite adequate resources, our best efforts and attention to detail, we were able to obtain complete and analyzable data at both assessments for about 60% of the adolescents enrolled in the trial. There were a variety of reasons for missing actigraphy data and no apparent differences in baseline characteristics were found among those who did not and did not have complete data. Unlike brief self-report measures usually completed in person, actigraphy involves continuous data collection for 7 days at home. As a result, data can be lost at numerous points along the way. Participants must receive the monitor and wear it for the entire time according to instructions, they must remember to return the devices, and the devices must function correctly. Also, unlike questionnaires, it is rarely possible to have patients agree to complete actigraphy if they have dropped out of the trial. Despite these challenges, the results of the study confirm that it is worthwhile to obtain as much objective information as possible for several reasons. Objective measurement of activity has high ecological validity in terms of measuring daily activity levels in real life, is quite distinct from adolescents’ self-report of their functioning, and it provides a more in-depth understanding of how interventions may have differential (and sometimes unanticipated) effects.
Although the technology for activity monitoring has become more available, there remains a paucity of information regarding the psychometric properties of this instrument especially among chronic pain populations. The majority of actigraphy studies have been in adults, with only three known studies assessing actigraphy data in pediatric chronic pain populations , , . Studies have reported a variety of activity indices including average daily activity counts, peak activity, vigorous activity, sedentary activity, and metabolic equivalence, but there is little information about which variables are most reliable and meaningful. Even less is known about the reliability of activity data across multiple time points/assessments, and to our knowledge, there is no self-report activity questionnaire that has been validated for adolescents with JFM to help support the psychometric properties of activity data. In this study, we explored the consistency/stability of various activity indices from pre- to post-treatment. The best indication of the stability of the measures came from the FE group which did not receive the active intervention and was not expected to change over time. In this group, moderately high correlations (Pearson r’s >.60) were found between pre- and post-measures of average activity counts and time spent in vigorous and moderate activity. There was slightly lower agreement for time spent in light and sedentary activity, and peak activity appeared to be the least stable measure of physical activity. In the CBT group, time spent in moderate and vigorous activity were also the most stable measures suggesting that actigraphy may be a more reliable indicator at higher activity levels. More validation work with longitudinal assessment using actigraphy is needed to further evaluate the most meaningful indicators of physical activity for use in clinical trials.
This study provides a first glimpse of how the various activity indicators obtained from physical activity monitoring can add to our understanding of the effects of behavioral and self-management interventions in adolescents with JFM. Overall the findings of this study suggest that adolescents with JFM exhibit sedentary lifestyles and that overall activity levels do not improve with CBT alone. Future studies should therefore consider enhancing CBT with education about physical activity with targeted aerobic exercise or strength training to improve physical activity and increasing pain coping skills in patients with JFM to achieve more robust improvements in physical functioning.
We would like to thank members of the Juvenile Fibromyalgia Clinical Trial Research Team: Dr. Lesley Arnold (University of Cincinnati), Drs. Anne Lynch-Jordan, Daniel Lovell, Scott Powers, Judy Bean (Cincinnati Children’s Hospital), Dr. Murray Passo (Medical University of South Carolina), Dr. T. Brent Graham (Vanderbilt University School of Medicine), Dr. Philip Hashkes (Shaare Zedek Medical Center, Jerusalem), Drs. Steven Spalding, Margaret Richards and Gerard Banez (Cleveland Clinic Lerner School of Medicine). We would also like to thank study staff including Ivy Ho, PhD., Irina Parkins, PhD, Raegan Malblanc MA, Sonya Crook, RN, Megan Johnston, BA and Emily Verkamp, MSW.
Funding: National Institutes of Arthritis and Musculoskeletal and Skin Diseases grant (NIAMS R01AR050028) to first author.