Experiences in Implementing a Community based Mortality Study Comparing Postmortem and Verbal Autopsy for Measuring Tuberculosis Mortality in Kenya

Monday, 18 August 2014
Exhibit hall (Dena'ina Center)
Peter O Nyamthimba, MA , Kenya Medical Research Institute/ CDC public health collaboration, Kisumu, Kenya
Susan Gacheri, MPH , Division of Leprosy Tuberculosis and Lung Disease- Kenya, Nairobi, Kenya
Amek Nyaguara, PhD , KEMRI/CDC Research and Public Health collaboration, Kisumu, Kenya
Kevin Cain, MD , KEMRI/CDC Research and Public Health collaboration, Kisumu, Kenya
INTRODUCTION:  

Tuberculosis (TB) kills 1.5-2 million people per year globally. Reducing mortality is the primary aim of TB control programs, but the burden of TB-related mortality is generally not known. The plan is to use verbal autopsy for the survey. However, since the performance of verbal autopsy for this purpose is not known, it is necessary to first validate its utility compared to a gold standard of post-mortem examination. Such studies have rarely been conducted in resource-limited studies, and the feasibility of conducting such studies is not well established. The study objective is to compare the performance of verbal autopsy to post-mortem examination for identifying TB-related mortality. We report on experiences with implementation, including recruitment, enrollment, and acceptability.

METHODS:  

The study is a cross-sectional survey, which began following intensive community mobilization. Deaths in patients with respiratory symptoms or among patients on TB treatment at the time of death in the KEMRI/CDC Health and Demographic Surveillance Systems (HDSS) area in Siaya County are eligible for enrollment. Death notifications are issued by Community Health Workers and reported to the Study Coordinator who gets informed consent from the deceased family and transports the body to the mortuary for post-mortem examination.

RESULTS:  

Between July 2012 and February 2013 we received 155 notifications of deaths and enrolled 78 cases. Post-mortem examination, including histology, TB culture and PCR-based testing were done. 78 cases were missed and not enrolled due refusal by next of kin to allow postmortem and other bodies having been embalmed.

CONCLUSIONS:  

Autopsy-based studies were thought to be nearly impossible to conduct, but we found that they are feasible. Community acceptability is generally high. Communities are willing to participate in autopsy-based studies. Careful attention is needed to ensure that logistics allow for the smoothest possible implementation.