Finding and Explaining Fast and Slow Mortality Transitions: A Taxonomy of Preston Curves

Robert L. Cohen, Johns Hopkins Bloomberg School of Public Health
David Bishai, Johns Hopkins University

Objective: Preston curves map life expectancy against GDP per capita. We sought to produce a taxonomy of Preston curves using historical data from 192 countries. Methods: Country-specific Preston curves were constructed using all available data since 1800. Latent growth curve analysis (LGCA) was used to validate the hypothesized relationships. Results: There are at least seven different styles of Preston curve ascent, which correlate strongly with geography but also with other factors such as governance and economic diversification. Ongoing LGCA is assessing the validity of these patterns. Discussion: This method offers a way to identify over- and under-performers along the Preston curve and target them for qualitative and quantitative analysis of factors related to performance. Besides the obvious HIV-related underperformers, we also identified a set of countries whose underperformance reflects a neglect of certain public health strategies. Positive deviants suggest promising ways for countries to leverage economic progress for health improvements.