Predictors of sick leave and improved worker productivity after 52 weeks of intensive treatment in patients with early rheumatoid arthritis

Abstract

Objective: To identify predictors of sick leave and improved worker productivity in patients with early rheumatoid arthritis (RA) treated for 52 weeks with intensive combination strategies. Methods: Patients with early RA were included in the COmbinatietherapie Bij Reumatoïde Artritis (COBRA)-light trial and followed for 52 weeks. As the COBRA-light strategy proved to be non-inferior to the COBRA strategy, all patients were pooled. Predictors for sick leave and improved worker productivity were assessed through a 3 month time-lag multivariable logistic generalized estimating equations model. Results: At baseline, 97 patients had a paid job, 59 had no job, and for six patients the work status was unknown. During the trial, 13 patients stopped working (8%) and six started working (4%). Only sick leave in the past 3 months predicted sick leave. By excluding this variable, patient global assessment and actual hours of sick leave became predictors. Increased worker productivity was predicted by higher patient global assessment levels, Sharp van der Heijde score ≥ 1, actual hours on sick leave, and higher worker productivity in the past 3 months. Conclusion: Sick leave and improved worker productivity were mainly predicted by non-disease-specific variables. Both outcomes can be predicted on a 3 month basis, using the outcome over the past 3 months for the next 3 months. By applying this model in daily practice, decisions for therapy change could be based not solely on disease activity but also taking into account a possible high risk for sick leave in the upcoming 3 months.