Expressing serious doubts over the high estimates of 200,000 malaria deaths in India as reported in the latest edition of The Lancet, the World Health Organisation (WHO) on Thursday questioned the methodology adopted by the authors of the study.

The Lancet uses verbal autopsy method which is suitable only for diseases with distinctive symptoms and not for malaria. Malaria has symptoms similar to many other diseases, and cannot be correctly identified by the local population.

The use of verbal autopsy for malaria may result in many false positives. In this method, deaths due to fever from any cause are likely to be misinterpreted as malaria in areas with high incidence. In areas with low malaria incidence, the symptoms are difficult to distinguish, and would result in overestimates of malaria deaths, a statement issued by the WHO here said.

Independent studies

It said the organisation welcomed independent studies for estimating malaria deaths, provided the method used was appropriate. The limitations of verbal autopsy, and the implausibly high incidence rates implied by the malaria mortality estimates, indicate that the findings of the study cannot be accepted without further validation.

Malaria is endemic in many States of India. Maximum cases are reported from the North Eastern States, Orissa, Jharkhand, Chhattisgarh and a few districts of Maharashtra, Madhya Pradesh and Andhra Pradesh. Approximately 50 per cent of malaria cases reported in the country are due to Plasmodium Falciparum (a type of malaria which causes death), the statement said.

The WHO estimated 10,000 (deaths) - 21,000 malaria deaths in India in 2006 - based on the rate of three deaths per 1000 estimated falciparum cases. Estimates of falciparum cases are made by multiplying total estimated malaria cases by reported falciparum percentage. These estimates make necessary adjustments for under-reporting of malaria data in the countries.

This methodology was universal, it said, admitting that the present malaria estimation procedures had limitations.

Concerns raised

Pointing out that the same authors, in a study in 2005, had warned against the use of verbal autopsy for obtaining malaria death estimates, the statement said, "Given these methodological concerns raised already in the early stages of this new study, the WHO had proactively supported a project to validate the methodology of verbal autopsy used in the study. The findings of the project reconfirmed that verbal autopsy method overestimates malaria deaths.

The method was found to estimate deaths correctly only in 8 per cent fatalities."

Setbacks of the method

"This new study published in The Lancet also has bias, as deaths during 2001-03 have been assessed much later in 2005-06. Any adult member of the family or even a neighbour was interviewed during the verbal autopsy, and it is not clear how well the interviewee was familiar with the case during the period of illness prior to the death," the WHO said.

The limitations of the new study were also exposed when estimates were examined for particular States. The proposed estimate of malaria mortality in Orissa suggests, implausibly, that there are 17-50 million falciparum malaria cases annually in a population of 40 million.