# 4996
A pair of articles this morning published today in The Lancet (hat tip Arkanoid Legent for the link) that may help illuminate the real impact of Malaria in India and that likely have broader implications as well.
It’s a subject we’ve covered before; our difficulty in accurately estimating the number of victims from different diseases.
In this case, the disease is Malaria, which the World Health Organization has estimated claims about 15,000 lives each year in India.
Using the somewhat controversial technique of `verbal autopsies’, a new multi-year study has come up with a startlingly higher estimate.
The authors suggest that the `real’ number is likely an order of magnitude higher – estimating 205 000 malaria deaths per year with plausible lower and upper estimates ranging from 125 000 to 277 000.
Millions of people die each year around the world without ever receiving modern medical care. Living in mostly rural areas of developing countries, they are often buried quickly with no cause of death identified.
In many cases, if estimates of disease mortality are conducted at all in developing countries, those estimates are based upon limited urban sampling and then extrapolated to the country’s whole population.
Verbal autopsies are a technique used by the World Health Organization, and others, to try to determine the cause of death based on interviews with family members, friends and neighbors.
While not as precise as a conventional autopsy, this method has been used and refined over the years (see A standard verbal autopsy method for investigating causes of death in infants and children) with input from such prestigious entities as The Johns Hopkins School of Hygiene and Public Health and The London School of Hygiene and Tropical Medicine.
From this report:
A verbal autopsy is a method of finding out the cause of a death based on an interview with next of kin or other caregivers. In order for verbal autopsies to be comparable, they need to be based on similar interviews, and the cause of death needs to be arrived at in the same way in all cases.
In recent years, verbal autopsies have been used more widely to provide information on cause of death in areas where civil registration and death certification systems are weak, and where most people die at home without having had contact with the health system. This type of interview is often the only way to find out about the cause of death.
Which brings us to today’s study appearing in the early Online edition of the Lancet. You’ll need to be registered (it’s free & easy) to access the full text.
Adult and child malaria mortality in India: a nationally representative mortality survey
Neeraj Dhingra, Prabhat Jha, Vinod P Sharma, Alan A Cohen, Raju M Jotkar, Peter S Rodriguez, Diego G Bassani, Wilson Suraweera, Ramanan Laxminarayan, Richard Peto
Preview |Summary | Full Text | PDF
Despite uncertainty as to which unattended febrile deaths are from malaria, even the lower bound greatly exceeds the WHO estimate of only 15 000 malaria deaths per year in India (5000 early childhood, 10 000 thereafter). This low estimate should be reconsidered, as should the low WHO estimate of adult malaria deaths worldwide.
From the abstract, a couple of excerpts that describe the methods and their results.
Methods
Full-time non-medical field workers interviewed families or other respondents about each of 122 000 deaths during 2001—03 in 6671 randomly selected areas of India, obtaining a half-page narrative plus answers to specific questions about the severity and course of any fevers. Each field report was sent to two of 130 trained physicians, who independently coded underlying causes, with discrepancies resolved either via anonymous reconciliation or adjudication.
Findings
Of all coded deaths at ages 1 month to 70 years, 2681 (3·6%) of 75 342 were attributed to malaria. Of these, 2419 (90%) were in rural areas and 2311 (86%) were not in any health-care facility.
<SNIP>
The adjudicated results show 205 000 malaria deaths per year in India before age 70 years (55 000 in early childhood, 30 000 at ages 5—14 years, 120 000 at ages 15—69 years); 1·8% cumulative probability of death from malaria before age 70 years. Plausible lower and upper bounds (on the basis of only the initial coding) were 125 000—277 000.
Perhaps the most startling statistic was that 86% of these supposed malarial deaths occurred outside of any formal health care facility.
In an accompanying commentary (below), its authors state that this report – and growing evidence from other studies – should give the WHO and others `pause for thought’ and later, that `evidence is increasing that the scale of the burden has been greatly under-estimated’.
But the authors also caution that verbal autopsies `remain an imperfect method for the estimation of malaria mortality’, and these `unexpected findings’ require further investigation.
India's invisible malaria burden
Simon I Hay, Peter W Gething, Robert W Snow
The problem of undercounting (or in reality, under-estimating) the burden of disease extends far beyond just malaria, and spans the globe.
Here in the United States, we are only able to roughly estimate the number of influenza deaths each year, and those estimates are subject to considerable debate.
The much bandied about global fatality number from the 2009 pandemic – roughly 18,000 reported to the WHO – is used blithely (and perhaps, disingenuously) by the media to depict the impact of the outbreak even though no one in public health seriously believes that number to be accurate.
But agencies, governments, and the media thrive on numbers and will use (and often portray as reasonably accurate) whatever number they have until a `better’ number becomes available.
Which is how we saw an estimate of 1,000 barrels of oil a day being leaked into the Gulf of Mexico this summer gradually expand to 5,000 barrels - and then after several months - was eventually increased to over 40,000 barrels a day.
But of course, no one really knows the `right’ number.
And the same could be said for the right number of H5N1 infections around the world, the true count of locally acquired Dengue in the United States, or how many people are really affected by Lyme disease in this country.
As the CDC graphic below shows, only the `tip of the pyramid’ of practically any disease actually gets reported to the health department or CDC.
Even when a disease is `reportable’.
Which is why you always have to look for that pesky asterisk, and the footnote or disclaimer, when reporting on estimates and statistics.
The old joke says that `87.63% of all statistics are made up’.
While it might be more accurate to say that `x% of all statistics are based on incomplete, and possibly inaccurate data’, the point is worth making.
If it is repeated often enough, a statistic becomes `accepted fact’, even though it may be flawed by an order of magnitude.
And when economic, healthcare, or other policy decisions are based on bad, or incomplete data we usually get bad decisions.
Of course, `good numbers’ are hard to find. Sometimes, impossible.
And admittedly, my faith in the accuracy of techniques like verbal autopsies is limited. I’ve assisted in real autopsies where the cause of death was indeterminable, so interviewing relatives after-the-fact is bound to return less-than-optimal results.
But until a better method can be devised, I accept they may provide better insight (a relative term) into the burden of diseases like malaria in areas where clinical testing remains unavailable.
So, as with the previous estimate of 15,000 deaths a year from malaria in India, I take these new estimates with a large grain of salt.
While we must sometimes use numbers that we know may be less-than-accurate . . . good reporting dictates that we clearly come out and say so.
Even if it `bogs down’, or complicates the narrative of a story.
Otherwise, we are simply compounding a scientific felony.
