Healthy volunteer effect in industrial workers

Volunteers for epidemiological research, have lower mortality rates than non-volunteers, thereby producing a bias referred to as the "healthy volunteer effect" (HVE). Occupationally active persons have been similarly shown to have a reduced mortality relatively to the general population (the "healthy worker effect"). To determine whether a HVE exists in occupationally active persons, we followed for 8 years a cohort of Israeli male industrial employees, of whom 71.6% agreed to participate in 1985 in screening examinations for cardiovascular disease. We calculated standardized mortality ratios (SMRs) of the entire cohort relative to the general population, and compared the mortality among participants with that of the non-participants. Over 8 years follow up, SMRs were 78% for the entire cohort, 71% for participants and 99% for non-participants. Participants were older than non-participants and worked more commonly in smaller factories. A proportional hazard model indicated that after adjusting for these variables, the all cause mortality hazard ratio among participants and non-participants was 0.69 (95% CI = 0.51-0.94). During the first and last two years of the 8-year follow-up there were 39.6 and 30.0 age-adjusted deaths per 10,000 person-years among participants, and 58.6 and 51.5 respectively among non-participants. We conclude that the HVE occurs in occupationally active persons, and that it may persist for up to 8 years follow-up. J CLIN EPIDEMIOL 52;8:731-735, 1999. (C) 1999 Elsevier Science Inc.