Ascertaining the Cause of Death using Verbal Autopsy Data: The probabilistic InterVA.
Project team: Heribert Ramroth, Eva Lorenz, Mark Ssennono, Heiko Becher
External collaborators: Ali Sié, CRSN Nouna, Burkina Faso,
Peter Byass, Umea Universitet, Sweden
Funding: DFG (SFB 544), Project D1
Funding period: 2008-2012
In many developing countries most deaths occur outside hospitals. Therefore verbal autopsy (VA), involving following up deaths with those present at the time, has been shown to be a useful and often the only possible method for deriving cause of death statistics in such settings. The most common method for determining the probable cause of death (COD) from the VA interview is Physician Coded Verbal Autopsy (PCVA), an independent review of questionnaire data by physician(s) trained in VA coding.
A recent alternative method, the Interpret Verbal Autopsy (InterVA) model, is based on a Bayesian approach which derives posterior probabilities for causes of death given an a priori distribution at population level and a set of interview-based indicators. The InterVA model has begun to be used in developing countries, most notably in the INDEPTH network.
The aim of the current project is to compare the COD distribution using the PCVA approach versus the InterVA model, based on information from a French VA questionnaire in rural north-western Burkina Faso.
Working with verbal autopsy questionnaires: Special aspects
Project team: Heribert Ramroth , Johanna Rankin, Eva Lorenz, Heiko Becher
This project covers special aspects of the verbal autopsy (VA) questionnaires mentioned in the main project above. The question is: How informative is so called free text in the VA questionnaires. The free text part serves to describe the background situation of the time before death like general health of the deceased, health seeking behavior and other information found important to report by the relatives of the deceased. Additionally, this part is not in general coded by the fieldworker doing the interview in the standardized section of the interview. The aim of this project is to determine the impact of free text information seeking the final cause of death.