Collaborators: Orvalho Augusto, Clara Calvert (LSHTM), Sam Clark (Sociology), Tsuyoshi Kunihama (Statistics, Sociology, eScience postdoc), Zehang Richard Li (Statistics PhD student), Tyler McCormick (Statistics, Sociology, Sr. Data Science Fellow), Clarissa Surek-Clark (Linguistics), Basia Zaba (LSHTM)
Global health priorities are set and progress assessed by measuring how many lives are ended or affected by various diseases. Most low and middle income countries (LMIC) do not have adequate vital registration, with the result that two-thirds of global deaths are not registered and have no cause. The primary aim of this research is to develop alternatives to vital registration that provide enough direct information to understand and monitor population health. Activities in this proposal aim to develop affordable, robust, calibrated methods to assign causes to individual deaths and characterize the distribution of deaths by cause.
Verbal autopsy (VA) is a feasible, cheap method for assessing cause of death when full autopsy and death certification are not possible. Existing VA methods do not produce consistent, reproducible, comparable cause of death assignments and can be unacceptably slow. Nevertheless LMICs (e.g. China, India, Indonesia, Brazil, Zambia, Malawi, Moc ̧ambique), encouraged by the WHO, are moving rapidly toward using VA to describe and compare the overall health of populations through time. There are three major sources of variability in VA that can prevent VA-derived causes from being accurate and comparable: (1)the VA instrument,(2) the VA interview, and (3) the method of assigning causes. (1) is addressed by WHO standard VA instruments; this project aims to understand and address (2) and (3) by pursuing the the following specific aims (SA).
- McCormick, T. H., Li, Z., Calvert, C., Crampin, A. C., Kahn, K., and Clark, S. J. (2014) Probabilistic Cause-of-death Assignment using Verbal Autopsies. Preprint pdf.