Descriptive statistics for the 238 women with fibromyalgia are shown in and . Subjects' average age was 62.6 years (SD = 10.8), and approximately three-quarters had either attended or graduated from college. Traumatic experiences and major life stressors were strikingly common. Moreover, subjects' average depression scores (15.7) were just under an established screening criterion for clinical depression (≥16), and their average BMIs (29.2) were just under the obesity criterion (≥30) (39.1% met the screening criterion for clinical depression and 27.3% for obesity). Slightly less than a fifth (18.1%) reported sleep apnea, and 45% reported having been treated for fibromyalgia in the last 12 months.
Women with a physician-given fibromyalgia diagnosis: descriptive statistics (continuous variables) and their relationships to pain ratings
Women with a physician-given fibromyalgia diagnosis: descriptive statistics (categorical variables) and their relationships to pain ratings
Subjects' perceived pain ratings are shown in . In response to the question, “During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)?” slightly less than half responded quite a bit (n = 75, 31.8%) or extremely (n = 38, 16.1%), and, a fifth, moderately (n = 50, 21.2%). The mean score (±SD) was 3.21 ± 1.26 on the 5-point scale.
Pain responses of women with a physician-given fibromyalgia diagnosis (n = 236).
The bivariate relationships between predictor variables and the pain criterion are shown in and . Greater pain interference was associated with lower education, higher BMI, sleep apnea, treated for fibromyalgia in the last 12 months, having been in a bad accident, the experience of physical assault/abuse, and higher depression and hostility scores.
The multivariate (hierarchical regression) results are shown in . Personal/health characteristics were entered at stage 1. Lower education, higher BMI, sleep apnea, and fibromyalgia status (treated in the last 12 months) were significant predictors of greater perceived pain interference. Entered at stage 2 were the trauma and major life stress variables; significant (positive association) were having been in a bad accident and physical abuse/assault experience/s. At stage 3, the psychological variables—depression and hostility—were entered, and both were significant; higher scores were associated with greater pain interference. Age was significant (positive) only at stages 2 and 3 (narrowly missing statistical significance at stage 1, P = 0.05). Adjusted R2 for the final model was 0.308. In the final model, physical assault/abuse experience/s was the strongest pain restriction predictor. The next strongest, respectively, were age, lower education, and BMI. Not significant at any stage were emotional abuse/neglect, sexual assault/abuse, and major life stressor.
Hierarchical regression, pain predictors (fibromyalgia subjects, n = 195)
Finally, average perceived pain ratings for the five women with all of the significant characteristics were 4.40 ± 0.55. Average scores for the seven without them were 1.43 ± 0.54. Women in the first group met all of the following criteria: age ≥ 62.6 (sample mean), education = high school or less, BMI ≥ 30 (obese criterion), fibromyalgia status = yes, experienced physical assault/abuse or bad accident = yes, depression ≥ 16 (clinical depression criterion), and hostility ≥ 2.1 (sample mean). Women in the second group met none of these criteria.
The osteoarthritis and fibromyalgia groups differed in important ways. For the continuous variables, subjects in the osteoarthritis group (means and SDs for the osteoarthritis and the fibromyalgia groups are given in parentheses): had lower pain-related restriction scores (2.67 + 1.18 vs 3.21 ± 1.26), were older (70.4 ± 11.0 vs 62.6 ± 10.8), had slightly lower BMIs (27.9 ± 7.0 vs 29.2 ± 8.0), and had lower depression scores (10.7 ± 8.6 vs 15.7 ± 10.4). Hostility scores were similar (2.05 ± 0.5 vs 2.1 ± 0.5). For the categorical variables (percentages by group are given in parentheses), individuals in the osteoarthritis group: had fewer subjects in the lowest education category (20.4% vs 29.1%) and more in the highest education category (34.2% vs 26.6%), had fewer with sleep apnea (7.0% vs 18.1%), had fewer reporting bad accidents (34.6% vs 42.9%), had less emotional abuse/neglect (32.9% vs 45.2%), had less physical assault/abuse (38.2%. vs 53.7%), less sexual assault/abuse (28.8% vs 46.6%), and fewer major life stressors (63.6% vs 71.2%). With the exception of education and hostility, all differences were statistically significant (and all but BMI differences were substantial).
A total of 744 women answered “yes” to the osteoarthritis question (and “no” to the fibromyalgia question) and had complete information on all of the independent variables (and, thus, were included in the regression analysis). Five variables were significant in the final model (staged as in the fibromyalgia model; standardized beta coefficients in the final model are shown in parentheses): age (0.196), BMI (0.243), treated in the last 12 months (0.172), major life stressors (0.078), and depression (0.213). Adjusted R2 for the final model was 0.192.