The standardized mortality ratio and life expectancy

This paper develops a theoretical relation between the standardized mortality ratio (SMR) and the expected years of life and establishes a regression equation for easy conversion between these two statistics. The mathematical expression of the derived relation is an approximation, requiring an assumption of constant age-specific mortality ratios. It underestimates the "true" value calculated based on life table technique when the age-specific mortality ratios increase with age. This equation provides a conservative method to estimate the expected years of life for cohort mortality studies and facilitates an assessment of the impact of work-related factors on the length of life of the worker. It also allows one to convert the SMR to life expectancy in smaller studies whose sole objective is to determine the SMR in a working population. A 1% decrease (or increase) in the standardized mortality ratio will result in 0.1373 years increased (or decreased) life expectancy based on white male data for the US population. Furthermore, with data from 14 large oil refinery and chemical worker cohorts of white males, the "derived" expected years of life based on the regression equation closely predicts the corresponding value calculated using a standard life table technique. This statistical equation is expected to have practical applications when used in conjunction with the SMR to provide an approximate measure of life expectancy, a term and statistic familiar to most lay people. Am J Epidemiol 1992;135:824-31.