Butler Johnson and Baldwin 1995 — Managing Work Disability Why First Return to Work Is Not a Measure of Success

Summary

Butler, Johnson, and Baldwin (1995) challenge the standard workers' compensation (WC) metric of success — the first return-to-work rate — using longitudinal data on Ontario workers with permanent impairments injured 1974–1987. They show that while $85\%$ of injured workers initially return to work, only about half achieve stable post-injury employment. Using a multinomial logit model across four mutually exclusive employment patterns, they find that employer accommodations (reduced hours, modified equipment, light duty) are far stronger predictors of stable employment than any individual worker characteristic or benefit generosity measure.

Key Claims

• The first return-to-work rate ($\approx 85\%$) dramatically overstates disability management success; $40\%$ of those who initially returned were not employed due to injury effects at the survey date.
• Four mutually exclusive post-injury employment patterns: single absence/successful ($39\%$), single absence/unsuccessful ($29\%$), multiple absences/successful ($21\%$), multiple absences/unsuccessful ($11\%$).
• Employer accommodations dominate individual predictors: reduced hours cut the worst outcome (multiple absences/unsuccessful) by $61\%$, light-duty work by $55\%$, modified equipment by $38\%$.
• Higher education ($+28\%$), public sector employment ($+44\%$), and male sex predict single-absence successful return.
• The WC replacement rate discourages initial return to work, but has no significant effect on post-return employment stability — benefit generosity affects the entry margin, not the stability margin.
• Back injuries disproportionately produce multiple-absence patterns; especially severe for older, lower-skilled workers.
• Policy should track post-return stability, not just first return rates.

Concepts Introduced or Extended

• Disability Management and Return to Work
• Workers' Compensation and DI
• DI Replacement Rate

Entities Mentioned

• Richard J. Butler
• William G. Johnson
• Marjorie L. Baldwin

Quotes

"First return to work, the most commonly used indicator of WC program success, is not an adequate measure of the program's ability to restore injured workers to productive employment."

"The results suggest that employer accommodation behavior, not just characteristics of the worker, has a major effect on post-injury employment outcomes."

My Take

The paper's core methodological contribution — distinguishing first return from stable return — should be obvious in retrospect but was apparently not standard practice in 1995 WC research. The employer accommodation findings are striking: reduced hours alone cuts the worst outcome by $61\%$. This complements the O'Leary et al. (2012) finding that even modest workplace injuries substantially raise long-term Disability Insurance (DI) probability — if accommodations can intercept that pathway, the fiscal and human cost implications are large. The Ontario (Canada) data limit generalizability to U.S. WC systems, which vary enormously by state. The 1974–1987 injury cohort is pre–Americans with Disabilities Act (ADA) (1990); modern accommodation mandates may have changed employer behavior. The multinomial logit's independence of irrelevant alternatives (IIA) assumption is not tested.