Presentation M O5

M O5 (Oral Presentation):
Vocational rehabilitation program evaluation: comparison group challenges and the role of unmeasured return-to-work expectations

Presented by: Jeanne M. Sears

Download PDF


Sears JM1, Rolle LR2, Schulman BA1, Wickizer TM3,1

  1. Department of Health Services, School of Public Health, University of Washington, Seattle, WA
  2. Washington State Department of Labor and Industries, Tumwater, WA
  3. Division of Health Services Management & Policy, The Ohio State University College of Public Health, Columbus, OH, USA



Workers’ compensation (WC)-based vocational rehabilitation (VR) programs are intended to facilitate return to work (RTW) for injured workers unable to return to their previous job. VR outcome evaluations have been sparse, in part due to the scarcity of suitable comparison groups. Randomization is rare in WC settings due to regulatory barriers and equity concerns. This study assessed (1) the adequacy of a commonly recommended internal comparison group, i.e., workers who were eligible for but did not receive services, and (2) RTW expectations as a potential source of bias.


In this prospective cohort study, workers found eligible for vocational retraining were consecutively surveyed prior to initiation of retraining activities (response rate=62%). Survey data, WC claims data, and state wage data were used to compare the three-year outcomes of eligible workers who did and did not receive vocational retraining services. The magnitude of confounding contributed by RTW expectations and other baseline characteristics was quantified.


Of 361 workers surveyed, 80% were in the retraining plan group and 20% in the no-plan group. Of the retraining plan group, 61% were somewhat/very certain they would RTW, compared with 23% of the no-plan group (p<0.001). RTW expectations were strongly predictive of outcomes; workers who were somewhat/very certain they would RTW were less likely to have total permanent disability (1% compared with 19% for the uncertain group; p<0.001), more likely to earn any wages (40% compared with 19% for the uncertain group; p<0.001), and more likely to earn higher wages ($1,842 compared with $782 for the uncertain group; p=0.004). Adding RTW expectations to regression models reduced retraining plan effect estimates by 23%, while education and physical capacity each changed effect estimates by <5%.


RTW expectations predicted long-term RTW outcomes and can play a strong confounding role if unmeasured. The internal comparison group approach, commonly recommended for VR evaluations, appears inadequate for WC settings. There is no simple solution to the challenge of comparison group identification using administrative data. However, RTW expectations, an easily-measured multi-dimensional construct, may be a useful addition to the VR and WC evaluation toolbox.

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