Presentation M O4

M O4 (Oral Presentation):
Predicting time on benefits for injured workers with back pain, what additional information should be collected?

Presented by: Ivan Andreas Steenstra


Steenstra IA1, Franche RL1, Tolusso D1, Davilmar A1, Lee H1, Furlan AD1,2, Amick B1, Hogg-Johnson S1

  1. Institute for Work & Health
  2. Department of Medicine, Faculty of Medicine, University of Toronto



Some injured workers with work-related, compensated back pain (BP) experience a troubling course in return to work (RTW). In an earlier paper we showed that a prognostic model for time on benefits in the first episode based on readily available data had fair predictive accuracy. Further improvement of the earlier developed prediction tool aimed at identifying those workers at higher risk, was explored by adding data collected in a sub-sample of workers that also participated in the Readiness for RTW (R-RTW) cohort study..


This is a cohort study of workers with compensated LBP in 2005 in Ontario. Follow up was two years. Data collected by the Workplace Safety and Insurance Board (WSIB) and from the R-RTW study were combined. Exclusion criteria were: no-lost-time claims, > 30 days between accident and claim filing and > 65 years. We examined the added prognostic value of patient reported prognostic factors available in the first four weeks since injury through proportional hazard models. First the model fit was determined based on coefficients derived in the larger dataset to prevent over-fitting of the model in a small data-set. Next, variables collected in the R-RTW study were added to the model. Improvement of model fit was determined by comparing Area Under the Curve statistics. Outcome measure was time on benefits in the first episode.


Of 6,657 workers that had a claim, 1,442 were still on full benefits at four weeks. 113 of these workers were also part of the prospective cohort. Model fit of the established rule in the smaller data-set was comparable to the fit established in the larger data-set: AUC= 0.76 at 180 days. Adding pain score at baseline improved the rule (AUC=0.80, 95% CI= 0.68, 0.91 at 180 days, AUC= 0.88, 95% CI= 0.74, 1.00 at 360 days).


Factors currently collected by the WSIB show some ability to predict prolonged claims, collecting additional data on pain grade improves the model from classifying workers at higher risk from ‘fairly well’ to ‘good'. The quality of routinely collected data has improved over years. Prognostic factors, reported by multiple stakeholders can help identify those at high risk of extended duration on disability benefits in need of additional support.

Schedule Details

Monday September 29
10:45 - 12:15 Morning Concurrent Sessions (M O1 - M O12 and Seminar 1)
Session: Predictors of Work Disability
Room B