J Rheumatol 2002;29:1034-40)
SHARLA J. KING, JEAN WESSEL, YAGESH BHAMBHANI, DALE SHOLTER, and WALTER MAKSYMOWYCH
From the Health Science Council Office; the Faculty of Rehabilitation
Medicine; and the Division of Rheumatology, University of Alberta, Edmonton, Alberta; and the School of Rehabilitation Science, McMaster University, Hamilton, Ontario, Canada.
Funded by grants from the Medical Services Incorporated Foundation and from the Health Services Research and Innovation Fund, Alberta Health, administered by Alberta Heritage Foundation for Medical Research.
S.J. King, PhD, Academic Coordinator, Health Science Council Office, University of Alberta; J. Wessel, PhD, Professor, School of Rehabilitation Science, McMaster University; Y. Bhambhani, PhD, Professor, Faculty of Rehabilitation Medicine; D. Sholter, MD, FRCPC, Division of Rheumatology; W. Maksymowych, MD, FRCPC, Associate Professor, Division of Rheumatology, University of Alberta.
Address reprint requests to Dr. S.J. King, Health Science Council, 300 Campus Tower, 8625-112 Street, University of Alberta, Edmonton, AB, T6G 1K8.
Submitted June 22, 2001; revision accepted November 26, 2001.
Objective. To determine which sociodemographic, psychological, and behavioral characteristics of persons with fibromyalgia (FM) will predict a positive response to treatment; and to determine if subjects classified according to the Multidimensional Pain Inventory (MPI) responded differently to the interventions.
Methods. One hundred twenty-eight women with FM underwent baseline testing and were randomized into one of 3 intervention groups or a control group. After the 12 week program, the subjects were reexamined on the same pretest measures. Stepwise regression analyses were conducted to determine the variables that could significantly predict the change in the dependent variables. A reliability of change index was calculated to determine the proportion of responders and nonresponders.
Results. The stepwise regression revealed significant predictor variables for change for all dependent variables except the Fibromyalgia Impact Questionnaire; however, the percentage of the variance in the change scores explained by the independent variables ranged from 4 to 15%. Results from the reliability of change index indicated that no MPI subgroup responded more than another group on any measure.
Conclusion. Select sociodemographic and psychosocial variables and type of intervention were not strong predictors of improvement in a variety of measures after a treatment program. The low percentage of explained variance may be due to the heterogeneity of FM. Additionally, the low percentage of responders suggests that current forms of treatment are not effective for a large portion of the FM population. (J Rheumatol 2002;29:1034-40)