Improving the Precision of Population Health Measures and Ranks with Longitudinal and Joint Outcome Models

The University of Wisconsin Population Health Institute has published the County Health Rankings since 2010, using population-based data to highlight health outcomes, their upstream determinants, and to encourage in-depth health assessment for all United States counties. A significant methodological limitation, however, is the uncertainty of rank estimates, particularly for small counties. To address this challenge, we explore the use of longitudinal and pooled outcome data in hierarchical Bayesian models to generate county ranks of health outcomes measures with greater precision. Our models used pooled outcome data for three measure groups: (1) Poor physical and poor mental health days; (2) percent of births with low birth weight and fair or poor health prevalence; and (3) age-specific mortality rates for nine age groups. Fixed and random effects components of these models were used to generate posterior samples of rates for each measure. We also used time-series data in longitudinal random effects models for age-specific mortality. Using posterior samples from these models, we estimate ranks and Bayesian confidence intervals (credible intervals) of ranks for each measure. Credible interval widths for the ranks based on the univariate, joint outcome, and/or longitudinal models were compared to assess improvements in rank precision. We found that incorporating longitudinal or pooled outcome data generally improved rank certainty. For measures with different determinants, joint modeling neither improved nor degraded rank precision. This approach provides a simple way to use existing information to improve the precision of population health measures and ranks for small areas.