Goldblatt P, Fox AJ, Leon D. Mortality of employed men and women. American Journal of Industrial Medicine 1991; 20 (3): 285-306.

This paper presents mortality data for a 1% sample of men and women in England and Wales who were employed at the time of the 1971 Census of Population. It provides background information on the "healthy worker effect" by age, social class (as determined by occupation), cause of death, and length of follow-up. As expected, relative mortality of those employed at census rose with subsequent follow-up. This effect was strongly age-related, apparently as a consequence of the development (and increasing predominance) of chronic diseases with age. This suggests a unified explanation for some of the variation described in the literature. Statistical modelling of the relationship between mortality and length of follow-up confirmed that the healthy worker effect did not entirely disappear as follow-up progressed in this study. We examined social background as an explanation for this persistently low mortality, but found no evidence to suggest that it was an important factor.