YSPH Biostatistics Seminar: “Generalized Bayes Calibration of Compositional Cause-specific Mortality Data from Verbal Autopsies"

NOTE: BIS 525 students are required to attend in person. Others are invited to attend in person, but may also attend via Zoom.

SPEAKER: Abhirup Datta, PhD, Associate Professor, Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health

TITLE: “Generalized Bayes Calibration of Compositional Cause-specific Mortality Data from Verbal Autopsies"

ABSTRACT: Verbal autopsies (VA) are extensively used to determine cause-of-death (COD) in many low and middle income countries. However, COD determination from VA can be inaccurate, misclassifying COD for a large proportion of deaths. If not accounted for, this misclassification leads to biased estimates of cause-specific mortality fractions (CSMF), a critical piece in health-policy making. The predicted COD data is also compositional (probabilistic) in nature. We present VA-calibration -- a likelihood-free generalized Bayes approach to calibrate inaccurate and uncertain compositional cause of death data from verbal autopsies. The approach works with 0’s and 1’s in the compositions and does not require any transformation of the data. We provide a complete primer on how to use the VA calibration approach with the calibratedVA R-package to correct for verbal autopsy misclassification bias in cause-specific mortality estimates. We use calibratedVA to obtain CSMF for child (1-59 months) and neonatal deaths using VA data from the Countrywide Mortality Surveillance for Action Mozambique (COMSA) project in Mozambique. As a byproduct we offer a transformation-free interpretable method for regressing compositional response on a compositional covariate – implemented in the codalm R-package.

YSPH values inclusion and access for all participants. If you have questions about accessibility or would like to request an accommodation, please contact Charmila Fernandes at harmila.fernandes@yale.edu. We will try to provide accommodations requested by October 12, 2023.