The Impact of a Group-Based Multidisciplinary Rehabilitation Program on the Quality of Life in Patients With Fibromyalgia

Results From the QUALIFIBRO Study

Herlinde Jacobs, MD; Martine Bockaert, MSc; Jan Bonte, MD; Mark D'Haese, MD; Jan Degrande, MD; Line Descamps, MSc; Ulrike Detaeye, MSc; Willem Goethals, MD; Jo Janssens, MSc; Koen Matthys, MD; Lieve Muller, MSc; Hilde Van deWalle, MSc; Ilse Viane, MSc; Barbara Vlieghe, MD; Michiel Grillet, MD; Rachael Asher, MSc; Bernard Grillet, MD, PhD,; Dirk De Bacquer, PhD


J Clin Rheumatol. 2020;26(8):313-319. 

In This Article


The 94 patients with confirmed FM were 43.7 years on average and included 88 women (94%). In 54% of patients, the diagnosis was made in the 6 months prior to entering the multidisciplinary rehabilitation program according to the ACR 1990 classification (widespread pain with pain in at least 11 tender points); 45% of patients were diagnosed with at least 14 positive tender points. At the time of the baseline assessment, 97% met the ACR 2010 diagnostic criteria, being either a widespread pain index 7 or greater combined with symptom severity scale of 5 or greater or a widespread pain index of 3 to 6 combined with a symptom severity scale of 9 or greater.

Table 1 illustrates the distributions of scores resulting from completing the different questionnaires. Levels of anxiety and depression were considered high (≥11) in, respectively, 58% and 36%, whereas 38% of patients could be classified as showing signs of severe burnout (BMS-10 score ≥4.5). Mean norm-based scores for all SF-36 dimensions were markedly less than 50, indicating substantially lower HRQoL levels than the general population. Mean (SD) PCS and MCS scores were 33.1 (6.3) and 35.6 (9.2), respectively. Baseline PCS scores in the 94 FM patients were negatively correlated with HADS depression scores (Spearman ρ = −0.24, p = 0.02), whereas lower MCS scores were significantly associated with higher levels of anxiety (ρ = −0.41, p < 0.001) and depression (ρ = −0.30, p = 0.004) and higher burnout scores (ρ = −0.37, p = 0.003).

By the end of the 8-week rehabilitation program, 2 patients had left the study prematurely. Of the 92 FM patients who finished the program successfully, 89 patients (97%) completed the SF-36 questionnaire at 6 months and 82 patients at 1 year (89%). Health-related quality of life scores at 8 weeks in these dropouts were not significantly different from those of the 82 patients with a complete 1-year assessment. Table 2 gives the estimated changes in SF-36 scores from the start of the program to 8 weeks, 6 months, and 1 year for the 8 subscales. Clearly, HRQoL was significantly improved in all domains by the end of the program, at 8 weeks. Moreover, improved HRQoL levels were maintained at 6 months and at 1 year for all subscales illustrating the potential long-term benefit of the program. Even at 1 year, clinically relevant changes in the range +4 to +6 were observed for nearly all 8 subscales.

The changes in SF-36 summary scores observed at 8 weeks, 6 months, and 1 year are illustrated in Figure 1. For both the PCS and MCS scales, significant changes from baseline were observed at all points in time. At 1 year, an increase of +5 or more on the PCS scale was seen in 45% of the patients, whereas MCS scores increased with at least +5 in 52% of patients. Changes in PCS and MCS scores at 1 year were negatively related (ρ = −0.29, p = 0.008). In relation to baseline characteristics, we found that 1-year changes in PCS were significantly correlated with the patient's age (p = 0.02) and HADS depression scores (p = 0.04), with a stronger impact of the program in younger patients and those with depressive feelings before the start. Physical component summary changes were not associated with levels of anxiety and burnout. None of the baseline variables were associated with the magnitude of changes in MCS scores.

Figure 1.

Changes in SF-36 summary scores in FM patients participating in the multidisciplinary rehabilitation program.

Finally, 59% of the patients could be classified as showing a pain-persistence pattern at the start of the program; 41% showed a rather pain-avoidance pattern. The latter group of patients reported significantly lower PCS scores throughout the 1-year observation period compared with the patients who demonstrated a pain-persistence profile. There were no significant differences with respect to the MCS scale. Changes in PCS and MCS scores between baseline and 1 year were not associated with pain coping behavior.