Verbal Autopsy: Counting The Dead To Save The Living

By Dr. Samuel Oti,Senior Research Officer

2012-06-18

“Most people in Africa and Asia are born and die without leaving a trace in any legal record or official statistic. Absence of reliable data for births, deaths, and causes of death are at the root of this scandal of invisibility, which renders most of the world's poor as unseen, uncountable, and hence uncounted.”

~ PW Setel

From a human perspective, death is more often than not a tragedy. And from a public health perspective, death without a trace –that is, without any documentation of the death or the cause of death– is an even bigger tragedy. This is because if we do not know who is dying among us and why they are dying, then we cannot do much to prevent those deaths that could have been avoided.

Unfortunately, as PW Setel pointed out, the lack of cause-of-death data is basically the norm in many countries in Africa and Asia, especially among the poorer communities.  For example, it is quite normal in such countries for a person resident in an urban city and who falls chronically ill, to return to his or her rural home and await death. And when death comes, that individual is buried, and no legal or official documentation of the death or the cause of the death is made anywhere. In developed countries, such a scenario is highly unlikely. Many of these countries, such as those in Scandinavia, have almost complete records of deaths dating back to a few hundreds of years. Records of deaths are typically captured in what is known as a vital registration system. In this system, all deaths (and the causes of deaths) are documented wherever they may have occurred.

Countries like Sweden and Australia are said to have complete and high quality vital registration systems because between 90 and 100 per cent of deaths within their borders are captured by their vital registration systems. A developing country like Kenya on the other hand would be said to have an incomplete and low quality vital registration system since less than 50 per cent of deaths across the country are captured. Even within countries that have incomplete systems, the proportion of deaths captured varies from place to place. For example in Kenya, it is estimated that as much as 75 per cent of deaths in the urban city of Nairobi are captured by the vital registration system. Whereas, in the far flung towns in Kenya’s North Eastern province, less than 10 per cent of deaths are captured by the same system. The reason why vital registration systems are of low quality especially in the developing world is largely because they are quite expensive to set up and maintain. Therefore many developing countries are finding it difficult to set up vital registration systems that are comparable to what obtains in most developed countries.

So what can be done to document deaths in the absence of a complete and high quality vital registration system? There is a process known as Verbal Autopsy that has been utilized to record deaths in several developing countries such as India and China over the past few decades. Verbal Autopsy (VA) is the process of interviewing any respondent –usually a family member, close friend or neighbor– who might have accurate information about the circumstances leading to a death. Typically, in areas where VA is utilized, whenever a death occurs, the community alerts a trained VA interviewer. The interviewer visits the residence of the deceased and extends condolences. After this, the interviewer schedules a return visit, anywhere from a few weeks to a couple of months after the death has occurred, to come back to the residence and conduct the VA interview.

The interviewers use standard questionnaires to conduct the interview. Generally, a standard VA questionnaire has two parts. One part is a narrative section wherein the interviewer records a verbatim account of the circumstances surrounding death as narrated by the respondent. The other part has a wide array of specific questions about symptoms and signs that the deceased may have experienced before death e.g. did the deceased have a fever? How long did the fever last? Etc.  After the interview, the completed questionnaires are sent to medical doctors who go through each of them and by consensus, assign possible causes of death.

The process of VA is usually coordinated by research institutions, non-governmental organizations and in some cases, the local authorities.  APHRC is one such organization that employs trained interviewers to conduct VA interviews in two slums in Nairobi as part of its research work. We have been collecting cause of death information in the slums for over a decade in Korogocho and Viwandani slums in Nairobi. This is largely due to the fact that almost 80 per cent of deaths in the slums do not occur in a health facility and so without VA there would be no complete record of the causes of deaths in these slums. Our data shows that about half of all adult deaths in the slums are due to AIDS and/or Tuberculosis.  Deaths from violent injuries and non-communicable diseases such as Stroke and Heart Attack are also leading causes of death among adult slum dwellers. Among children, Pneumonia, Diarrheal disease and Malnutrition account for about half of all deaths.

This information has set the stage for APHRC to develop and implement interventions that could save lives in the slums. For example, APHRC recently won a multimillion-dollar grant from Comic Relief UK to improve maternal, newborn and child health (MNCH) services in the slums. Part of this project will involve the upgrading of six health facilities to become one-stop comprehensive primary health facilities for MNCH. Also, as part of its corporate social responsibility, APHRC runs medical camps once or twice a year in both slums where an average of 2000 residents receive basic medical attention free of charge.

In conclusion, VA provides a useful alternative to generating cause of death data and should be implemented widely in places where vital registration systems are incomplete or even non-existent. Cause of death data provides crucial information to health authorities that can be used to prevent the further loss of life. Indeed, the VA process is not perfect. For example, respondents may not recall enough useful information during the interview and this will affect the doctors’ ability to correctly determine the cause of death. Notwithstanding, several validation studies across the world have shown that VA is still a very useful way of counting the dead and finding out why they are dying in order to save future lives.