Estimating time to reinjury among Washington State injured workers by degree of permanent impairment: Using state wage data to adjust for time at risk

• Published in: American journal of industrial medicine 2021-01, Vol.64 (1), p.13-25
• Main Authors: Sears, Jeanne M., Schulman, Beryl A., Fulton‐Kehoe, Deborah, Hogg‐Johnson, Sheilah
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Aged; Calendars; Confidence intervals; disability evaluation; Estimates; Female; Hazard assessment; Health risks; Humans; Impairment; Injuries; Male; Middle Aged; occupational injuries; Occupational Injuries - epidemiology; permanent impairment; permanent partial disability; Rehabilitation; Rehabilitation, Vocational - statistics & numerical data; Reinjuries - epidemiology; reinjury; Retrospective Studies; return to work; Return to Work - statistics & numerical data; Risk; Risk Assessment; Salaries and Fringe Benefits - statistics & numerical data; Staff participation; Statistical analysis; Statistics as Topic - methods; Time; Time Factors; unemployment; vocational rehabilitation; Wages & salaries; Washington - epidemiology; Work Capacity Evaluation; Workers; Workers' Compensation - statistics & numerical data; workers’ compensation; Young Adult
• ISSN: 0271-3586; 1097-0274
• DOI: 10.1002/ajim.23200

Background Many injured workers are reinjured, but reinjury risk is challenging to quantify. Because many injured workers face delayed return‐to‐work, or return to part‐time or intermittent jobs, a calendar timescale may overestimate actual work‐time at risk, yielding underestimated reinjury rates. Objectives included determining: (1) reinjury risk by degree of permanent impairment and other factors, and (2) how choice of timescale affects reinjury estimates. Methods This retrospective cohort study included Washington State workers’ compensation (WC) claims for 43,114 injured workers, linked to state wage files (2003–2018). Three timescales were used to define at‐risk denominators: (1) calendar quarters; (2) quarters with any wages; and (3) full‐time equivalent (FTE) quarters, defined as cumulative work hours ÷ 520. Associations between reinjury outcomes and worker, injury, job, and WC vocational rehabilitation program participation characteristics were assessed using Cox proportional hazards regression. Results Overall reinjury rates were 5.9 per 100 worker‐years using a calendar timescale (95% confidence interval [CI]: 5.8–6.0), 10.0 using any‐wage quarters (95% CI: 9.9–10.2), and 12.5 using FTE quarters (95% CI: 12.3–12.7). Reinjury rates were highest in the first two quarters after initial injury, remaining elevated for about 4 years. Using FTE quarters, workers with ≥10% whole body impairment had a 34% higher risk of reinjury relative to workers with no permanent partial disability award (95% CI: 1.25–1.44); no difference was detected using calendar time. Conclusions Timescale substantially affects reinjury estimates and comparisons between groups with differential return‐to‐work patterns. Linking wage data to WC claims facilitates measurement of long‐term employment, yielding more accurate reinjury estimates.