The infant mortality rate, life expectancy at birth, and a linear index of mortality as measures of general health status

Abstract

The infant mortality rate is not a good indicator of overall mortality or health status. Based on new empirical life tables from the UN Population Division, it can only predict life expectancy with 95% confidence to within a 14-year range. Two infant mortality rates must be nearly 80 units apart to be 95% confident that life expectancy in the two communities is different. Life expectancy itself is not an ideal general measure of mortality, because it implicitly weights deaths at different ages in an inconsistent fashion. A measure of potential years of life lost is preferable because it is ethically more consistent.

PIP: This paper re-assesses the use of the infant mortality rate and the life expectancy at birth as indexes for general mortality in developing countries, and then suggests and alternative, called the linear index of mortality or LIM. The infant mortality rate is not a good indicator of overall mortality or health status because it is sensitive to socioeconomic changes, in particular to maternal education. Based on new empirical life tables from the UN Population Division, it can only predict life expectancy with 95% confidence to within a 14-year range. Life expectancy itself is not an ideal general measure of mortality because it implicitly weights deaths at different ages in a inconsistent way. The linear index of mortality (LIM) is a measure of potential years of life lost. By weighting the number of deaths in each age group by 100 minus the age of death, a composite index of general mortality can be calculated that is linear, uniform and ethically consistent. Using UN empirical life tables for 20 developing countries, it is possible from the LIM to get an idea how much the difference between high and low mortality countries is due to differences in infant, child and adult mortality. This kind of comparison is impossible with life expectancy because it is non-linear. The relationship between the natural log of life expectancy and LIM is linear. Alternative health status measures like LIM need to be developed, in order to make our statistical perception of the world's nations consistent with our ethical conception of how health should be measured.