Back pain and social isolation: Cross-sectional validation of the friendship scale for use in studies on low back pain

Abstract

Background: Low back pain (LBP) is common, and social isolation is both a risk factor for poor recovery and a consequence. However, no studies seem to have validated social isolation measures in LBP populations. Aims: This study assessed the validity of the Friendship Scale (FS), a brief measure of social isolation. Methods: LBP participants were 100 consecutive consenting adult patients attending physiotherapy outpatient clinics, matched (1:2) by age and gender with a general population sample (GPS; n=200). FS validation was through factor analysis, internal consistency, sensitivity by known groups, and Rasch analysis. Results: There were significant differences between LPB and GPS on 5 of the 6 FS items. Social isolation on the FS was reported by 26% of the LBP cohort compared with 9% of the GPS. All FS items loaded on the principal component < 0.60, suggesting unidimensionality. Internal consistency was α=0.81. The FS was sensitive by pain severity and study cohort. Rasch analysis showed no disordered items, although 2 items were marginally misfitting. Differential item functioning by sex was observed on 1 item; there was no other observed differential item functioning. After removal of the worst fitting item (feeling alone), the remaining items fit the Rasch model. This, however, may have been a function of study samples. Conclusions: Generally, the FS performed well, and its descriptive system contains excess capacity beyond that needed in the study population; that is, those with LBP were not particularly socially isolated, and responses indicating severe social isolation were barely reported by these participants. Overall, the FS appears to be a suitable instrument for assessing social isolation among LBP patients.