A dispute has arisen between the WHO and the authors of a study claiming that malaria kills twice as many people as previously estimated, partly due to the under-reporting of malaria deaths in adults.
The study, published in The Lancet last week (4 February) by researchers at the Institute for Health Metrics and Evaluation (IHME) at the University of Washington, United States, concluded that there were 1.2 million malaria deaths in 2010 — almost double the WHO's estimate of 655,000.
But critics have questioned the reliability of their data, and some have accused them of using exaggerated language.
The researchers used malaria mortality data from 1980 to 2010, including birth and death records and verbal autopsies, and computer modelling to develop a database of malaria mortality by age, sex, country and year.
They estimated that malaria deaths peaked at 1.8 million in 2004 and have been declining since, thanks to efforts to control the disease, confirming WHO findings. But they differ over the extent of the decline.
The researchers said in a press release that their higher mortality figures suggest that short-term goals — such as the Roll Back Malaria Partnership's target of eliminating malaria deaths by 2015 — might be unrealistic. They estimated instead that "malaria will decrease to less than 100,000 deaths only after 2020".
They also calculated that many adults die from malaria, rather than just children as commonly believed. They estimated that more than 40 per cent of deaths from malaria involved people aged 5 years or more — almost eight times the WHO figures.
Stephen Lim, professor of global health at the IHME and an author of the study, said this is largely because the cause of death is rarely known for people aged 5 and above. It was assumed that exposure during childhood provides immunity for older people.
"When you gather all the available data, you see that there are a lot of deaths in older children and particularly in adults," he said.
However, the WHO has defended its figures.
WHO spokesman Gregory Hartl told US news organisation Voice of America that the authors used 'verbal autopsies' rather than laboratory diagnoses. "You rely on the verbal record of a friend or relative saying that X person died of fever, for example. However, we know that there are many different diseases which cause fever."
Bob Snow of the Malaria Public Health & Epidemiology Group at the Centre for Geographic Medicine in Nairobi, Kenya, one of the paper's peer reviewers, told Nature News: "Verbal autopsy is a very blunt tool — in some cases it is about as good as flipping a coin for working out the cause of death. It can be useful if somebody has died of an obvious cause … but it's not very useful for the mixed bag of symptoms malaria is associated with."
"I don't think the IHME or the WHO know how many people die of malaria worldwide — the truth is that nobody really knows. But that's not going to get headline news," he added.
Richard Horton, editor-in-chief of The Lancet, heralded the study, saying on Twitter that "a revolution is about to strike".
Kevin Marsh, chairman of the WHO's Malaria Policy Advisory Committee, who works with Snow in Nairobi, said: "[The authors] present these numbers as though they are 'real' numbers, making rather immodest statements such as 'these data show' rather than 'we believe these indicate' — it's a language issue," he says.
The authors have acknowledged that the verbal method is flawed, but Lim said that death registration systems are rarely available in malaria-endemic countries and that verbal autopsies are the "best alternative".
*free registration required to view this article
The Lancet 379, 413–431, doi: 10.1016/S0140-6736(12)60034-8 (2012)