Quality of Life and the Effects of Tailored Health Coaching in Fibromyalgia Patients

Overview

Specific Aims: 1. To identify phenotypes of patients with fibromyalgia according to symptom clusters and to compare differences in quality of life (QOL) among different phenotypes. 2. To examine the effects of technology-assisted and tailored health coaching in comparison to telephone support on health status, QOL, pain catastrophizing, and self-efficacy in patients with fibromyalgia. Methods: For Aims 1, the investigators will conduct a cross-sectional study and enroll 300 patients with fibromyalgia. Symptoms will be assessed using the Numerical Rating Scale, Athens Insomnia Scale, Functional Assessment of Cancer Therapy cognition scale, Beck Depression Index-II, Beck Anxiety Index, and Fatigue Severity Scale. The World Health Organization Quality of Life-Brief Form (WHOQOL-BREF) questionnaire will be used to assess participants' overall QOL. Hierarchical cluster analyses will be used to identify fibromyalgia phenotypes according to pain, physical, and psychological variables. A series of multivariate analysis of variance analyses will be used to compare the differences in WHOQOL-BREF scores among those phenotypes. For Aims 2, the investigators conduct an assessor-blind, parallel-group, randomized controlled trial and enroll 110 participants with fibromyalgia. Participants will be randomized to a health coaching group and a control group. The tailored, interactive health-coaching program will be delivered via mobile applications during a 10-week training period. The control group will receive standard education materials and weekly telephone support. The primary outcomes are the revised Fibromyalgia Impact Questionnaire and WHOQOL-BREF scores; the secondary outcomes are pain catastrophizing score and self-efficacy, which will be examined at baseline, post-training, and the 3th month follow-up. Data will be analyzed according to the intention-to-treat principle. To determine the effectiveness of health coaching on primary and secondary outcomes, differences in outcome variables will be analyzed with mixed-effects linear regression models. The between-group differences at the two posttests examined using a mixed-model will include group x time interaction. The investigators will adjust for the baseline score on the outcome variable and for demographics and comorbidities that differ significantly between the intervention and control groups at baseline.

Background: Fibromyalgia is a chronic pain syndrome predominantly affecting adult women. It is characterized by widespread musculoskeletal pain accompanied by physical and psychological symptoms, such as sleep disturbance, cognitive impairment, fatigue, stiffness, anxiety, and depression. The complexity and heterogeneity of clinical symptoms have been observed among patients with fibromyalgia. Variability in the intensity of fibromyalgia-related symptoms differentially impact patients' degree of disability, quality of life, and pain. Several cluster studies have sought to identify clinically relevant phenotypes based on disease symptoms. However, those studies only focused on selected physical and psychological symptoms to construct the symptom clusters. Few studies included cognitive variables into the clusters and few studies investigated the difference in the quality of life (QOL) and functional status among different subgroups or phenotypes of patients with fibromyalgia. Therefore, identification of phenotypes based on a comprehensive set of symptoms and determining differential impacts of fibromyalgia phenotypes on quality of life among patients with fibromyalgia is important, both scientifically and clinically as this information may help the healthcare professionals select appropriate treatments. Recent meta-analyses showed that multicomponent treatments are effective in reducing pain and improving QOL in patients with fibromyalgia, probably owing to the complexity of pain and heterogeneity of symptoms. Pain catastrophizing and self-efficacy are powerful predictors for disability. Therefore, an effective program must include tailored, multicomponent therapy for decreasing catastrophizing and increasing self-efficacy, which in turn may improve symptoms, functional status, and quality of life. Specific aims of the study are: 1. To identify phenotypes of patients with fibromyalgia according to symptom clusters and to compare differences in QOL among different phenotypes. 2. To examine the effects of technology-assisted and tailored health coaching in comparison to telephone support on health status, QOL, pain catastrophizing, and self-efficacy in patients with fibromyalgia. Methods Specific Aim 1-Study 1 (Year 1-2): Participants will be recruited via referrals from the physicians and from the community by on-line advertisements posted on websites or social media, and advertisement (posters and flyers) posted in public places. Based on Formann's method, which states that the maximum number of clustering variables should be m, where sample size = 2\^m, the number of clustering variables was limited to 7. A recent study identified three fibromyalgia subgroups among 345 participants. Therefore the investigators will enroll 300 participants for Study 1. Specific Aim 2 - Study 2 (Year 1-3): Similarly to Study 1, participants will be recruited via referrals from the physicians and from the community by on-line advertisements posted on websites or social media, and advertisement (posters and flyers) posted in public places. Naturally, subjects who are screened for eligibility for participation in Study 1 will be offered a chance to participate in Study 2. The sample size is estimated based on the expected treatment effect size for the primary outcome. The treatment effect size of non-pharmacological treatments for improving QOL in fibromyalgia patients was 0.73 according to results from a previous meta-analysis study. Assuming a type I error of 0.05, a type II error of 0.2 and an effect size of 0.73, 44 subjects per group will achieve a 0.8 power with a 2-sided test according to the results of the power analysis. 110 participants will be enrolled to accommodate planned subgroup analyses and to allow for a 25% dropout rate. Randomization sequence generation and concealment: After obtaining written consent and completion of baseline measurements, eligible participants will be randomly assigned in a 1:1 ratio with the use of permuted blocks of ten each to an intervention group (n = 55) and a control group (n = 55). An independent research assistant who will not involve in participant recruitment, enrollment, and data collection will generate randomization sequence using computerized software. The generated random sequence will be concealed in sequentially numbered, opaque envelopes until the participant is assigned. Appropriate random allocation and allocation sequence concealment will reduce selection bias. Blinding: To minimize detection bias, assigning an independent research assistant who is masked to the group assignment for baseline and post-test data collections will ensure assessor blindness. Study procedures and data collections: The study protocol describing the flow of participants into the intervention and control groups and data collections in different study weeks is depicted in Figure 1. Briefly, subjects respond to advertisements or referred by physicians will undergo screening for eligibility. After confirming the eligibility and obtaining written informed consent, participants will complete the baseline assessment of demographic variables, medical history, lifestyles variables, primary outcomes and secondary outcomes. Participants will then be assigned to the study groups according to the pre-generated randomization sequence. Both groups will repeat the assessment of the primary and secondary outcomes immediately after the 10 weeks of training period (posttest 1) and at the 3-month follow-up (posttest 2). Statistical analyses: Differences in baseline data will be determined using the Mann-Whitney U-tests, chi-square tests, and t-tests for independent samples. Comparisons of group means at baseline, and the two post-tests will be performed for normally distributed primary and secondary outcomes using independent t test. Effect size (Cohen's d) will be calculated for each outcome variable. To determine the effectiveness of technology-assisted health coaching on primary and secondary outcomes, differences in outcome variables will be analyzed with mixed-effects linear regression models with the covariance structure unstructured. The investigators will perform statistical analyses according to the intention-to-treat principle. All missing vales will be imputed using the last value carry-forward method. The between-group differences at the two post-test assessments will be examined using a mixed-model in which group and time interaction will be included. The investigators will adjust for the baseline score on the outcome variable and for demographics and comorbidities that differ significantly between the health coaching group and control group at baseline.