To propose an incidence definition of back injury for epidemiologic studies using health care contacts.
Medical services, hospitalizations, and workers' compensation data were linked for a longitudinal database of health care contacts among a cohort of heavy-industry workers for trajectory, group-based analysis.
During follow-up, 25.8% of workers had no health care contacts for back injury. Among workers with at least one contact, four trajectories were identified: one with a high probability of back injury during follow-up and three with episodic trajectories of increasing and decreasing probability of back injury.
Workers with no back injury history could be followed for incidence in cohort studies or as controls in case-control designs. Episodic groups could be followed for new episodes, providing they were free of health care contacts for back injury for at least 3 years.
From the School of Population and Public Health (Dr Koehoorn, Dr Teschke); School of Environmental Health, College for Interdisciplinary Studies (Dr Koehoorn, Ms Village, Dr Trask, Dr Teschke); and Centre for Health Services and Policy Research, Faculty of Medicine (Dr Koehoorn, Mr Xu), University of British Columbia, Vancouver, British Columbia, Canada.
This project was governed by a Data Access Agreement between the Researchers and the Ministry of Health-Government of British Columbia and WorkSafeBC (03–032). The project was approved by the Behavioural Research Ethics Board, University of British Columbia (B03-0644).
Address correspondence to: Mieke Koehoorn, School of Population and Public Health, University of British Columbia, 2206 East Mall, Vancouver, British Columbia, Canada V6T 1Z3; E-mail: firstname.lastname@example.org.