The extensive claim and medical bill payment databases of the Washington state workers' compensation system have been used to conduct epidemiologic and outcome studies of work-related conditions. Computerized administrative data must be supplemented with medical record review and structured interview of workers in outcome studies in order to adequately adjust for baseline severity and to address functional and patient satisfaction outcomes, respectively. Three examples of surgical outcome studies are described (carpal tunnel, lumbar fusion, thoracic outlet). Duration of disability prior to surgical intervention is an important predictor of duration of disability following surgery, even when other biologic markers of severity are included in multivariate modeling. Sufficient follow-up time is required to adequately assess longer-term outcomes, such as return-to-work status and the substantial effects of residual impairment even after claim settlement. Finally, well-conducted outcome studies may be linked to the development of surgical treatment guidelines in workers' compensation.