Predictors and modifiers of treatment effect influencing sick leave in subacute low back pain patients.

Abstract

Modifying effects in multivariate analyses of a randomized controlled trial. To identify prognostic factors for the effect of a brief intervention ("modifiers") at a spine clinic on return to work in patients with subacute low back pain. A previous study of a brief intervention showed significant reduction of sick leave, compared with usual primary healthcare treatment. Randomized controlled trials give data only on the group as an average. Identifying prognostic factors that interact with the treatment ("modifiers") may identify specific groups requiring this or other types of treatment. A total of 457 patients who had been sick-listed 8 to 12 weeks for low back pain were randomized into an intervention group (spine clinic with medical examination, information, reassurance, encouragement to engage in physical activity, n = 237), and a control group (primary health care, n = 220). All subjects filled out questionnaires. Logistic regression and tests for interaction were used to identify prognostic factors and modifiers for return to work in the two groups, at 3 and 12 months of follow-up. At 3 months of follow-up, the strongest modifying effect on return to work was the perception of constant back strain when working and beliefs about reduced ability to work. At 12 months, gastrointestinal complaints were the strongest modifier for the effect of the intervention. The spine clinic intervention seems to have a main effect on work absenteeism via interacting with the concerns of being unable to work.