[Skip to Navigation]
Sign In
December 13, 2016

Understanding County-Level, Cause-Specific Mortality: The Great Value—and Limitations—of Small Area Data

Author Affiliations
  • 1Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts
  • 2Department of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health, Boston, Massachusetts
JAMA. 2016;316(22):2363-2365. doi:10.1001/jama.2016.12818

In this issue of JAMA, Dwyer-Lindgren and colleagues1 present advanced methods and applications of small area estimation techniques to produce county-level summary measures of cause-specific mortality rates across the United States and estimates of temporal trends in these rates. The study used validated redistribution methods to recapture mortality data that would have been lost to so-called garbage coding, the practice of assigning potentially noninformative mechanisms of mortality (eg, cardiopulmonary arrest) rather than underlying disease codes (eg, congestive heart failure) to death certificates. The authors used enhanced generalized linear mixed-effects regression models to incorporate information on geographic spatial patterns, time and age associations, and relevant population-level covariates, to achieve valid cause-specific mortality rate estimates from the National Vital Statistics System (NVSS) without pooling data across years for counties with small sample sizes.

Add or change institution