Deworming debunkedBMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.e8558 (Published 02 January 2013) Cite this as: BMJ 2013;346:e8558
At first sight, treating children in poor countries for intestinal parasites seems like a no brainer. Who’d hesitate to provide the few pennies it costs to deworm a child in India, Africa, or other areas where infestations are endemic?
So obvious are the benefits, however, that many well meaning individuals and institutions may have been guilty of overstating them. Deworming has been hailed as a panacea: a simple, cheap, and effective way of improving growth, raising brain power, and improving the educational and employment prospects of millions of children.
Its supporters are many. At the 2008 Davos economic summit, Cherie Blair, wife of the former prime minister Tony Blair, dressed up as a worm and chased a bunch of delegates pretending to be children, a gimmick inspired by the charity Deworm the World. Tony was apparently unimpressed: “Every time I mention the subject to my husband,” Cherie confided to Time magazine, “he looks very distressed and runs out of the room.”1
If you counted the international organisations that support deworming as a development rather than as simply a health initiative you would quickly run out of fingers: the World Bank, the World Health Organization, the Bill and Melinda Gates Foundation, the US National Institutes of Health, the Copenhagen Consensus, Innovations for Poverty Action, the Centre for Effective Global Action, the Abdul Latif Jameel Poverty Action Lab, and, of course, Deworm the World. Surely such a lengthy and impressive list of supporters must know it really works?
Not if you read the latest revision of the Cochrane review on the subject, published in July this year by a team from the Liverpool School of Tropical Medicine.2 Professor Paul Garner, one of the authors, says a careful reading of all the published evidence that passes the Cochrane quality test shows “quite strongly” that deworming alone has no effect on growth, cognitive ability, or school attendance. “I’d love it to work,” he says. “But to claim that it does on the basis of the evidence available is simply misleading.”
This is not the first time the Liverpool team has expressed doubts. The first version of the review was copublished with the BMJ in 1997, and the second version in 2000, both to a chorus of criticism. “Nobody likes evidence that disconfirms their beliefs,” Garner says. “The first thing they do is to rubbish the method and then they rubbish the person.” He believes attempts were made in the past at a high level to discourage him, by putting pressure on his superiors to persuade him to back off. If so, they failed. The review has since been updated twice, in 2007 and 2012.
The deworming literature is large but of variable quality. Some studies carried out in the 1990s show impressive weight gains achieved in schoolchildren after a single dose of drugs that kill intestinal helminth worms (roundworm, hookworm, and whipworm). But larger studies designed to confirm these benefits have tended to draw a blank. Several studies have remained unpublished, raising suspicions that they did not generate positive findings. Other unpublished studies are boldly cited as providing support, though because they are unpublished it is difficult to have confidence in them. One positive study that was published in the BMJ and headlined on the cover “War on worms and malnutrition: regular anthelmintic treatment improves weight gain in African children” contained a statistical error that, according to Professor Garner, invalidated its conclusions.3 This controversy means that the claims made for deworming rest uneasily on a few rather elderly studies that are constantly cited, rather like a mantra designed to keep the spirits up.
Despite this, there is little sign of flagging enthusiasm for the deworming initiative, which is propelled by an alliance of international organisations, charities, and drug companies willing to supply the drugs (mainly mebandazole and albendazole) more or less free. The Disease Control Priorities Project, supported by the US National Institutes of Health, the World Bank, WHO, the Population Reference Bureau, and the Gates Foundation, asserts that the benefits of deworming go far beyond reducing the worm burden in individuals, contributing to “higher educational attainment, labour force participation, productivity and income.”4
A bulletin from WHO on the millennium development goals is unequivocal.5 Deworming boosts the prospects of school age children earning their way out of poverty, it says. “The improvements in intellectual development and cognition that follow deworming have been shown to have a substantial impact on professional income later in life,” it further claims, citing the development of the southern states of the US after 1900 and suggesting that similar programmes in Japan in the 1950s were one of the reasons for that country’s subsequent economic boom.
The WHO bulletin dates from 2005, but little has subsequently changed. A meeting of economists, including four Nobel Prize winners, held in Copenhagen earlier this year concluded that deworming was among the top four cost effective interventions, along with improving nutrition, treating malaria, and improving childhood immunisation.6 Robert Mundell, a Nobel laureate in economics and a participant at the meeting, concluded: “Deworming is an overlooked intervention deserving of greater attention and resources. This simple, cheap investment can mean a child is healthier and spends more time in school.”
It is puzzling that he should have considered it overlooked. GlaxoSmithKline already donates 400 million tablets of albendazole a year, and Johnson & Johnson 200 million tablets of mebendazole, a donation rate the two companies have promised to sustain until at least 2020. In addition, a range of charities and non-profit organisations, mainly in the US, contribute more than 100 million deworming pills a year, which they acquire either free or for a few cents a pill. Some of the charities enter the pills in their accounts as costing as much as $10 (£6; €8), making them look like hugely munificent philanthropies, Forbes magazine reported in December 2011.7 For example, Oklahoma City based Feed the Children overstated its gifts-in-kind contributions by hundreds of millions of dollars by pricing the deworming pills it distributed at $9.07 each. In 2011 it changed this policy, lowering the price to $0.35 a pill and reporting a fall of $668m in its gift-in-kind contributions. But even $0.35 is 1600% above the world market price, Forbes drily remarked.
While some US charities overstate their generosity, others overstate the benefits of deworming, according to the Cochrane review. A group of them—Deworm the World, Innovations for Poverty Action, the Center for Effective Global Action, and the Abdul Latif Jameel Poverty Action Lab—responded to the review with a critique headed: “Cochrane’s incomplete and misleading summary of the evidence on deworming.”8 It cites three studies that it claims the Cochrane team ignored, a 2004 study that drew on data collected early in the 20th century in the deep south of the US and two unpublished studies. But none of these was a randomised controlled trial comparing those given pills against those given no drug or a placebo, so they did not meet the Cochrane criteria and were appropriately excluded from the analysis, Garner and his coauthors responded.
What does the evidence show?
The Cochrane reviewers did identify 42 trials that met the criteria for inclusion. Excluding one for which the data are still awaited, the remaining 41 included 65 168 participants. The trials included three with 149 children who had been screened for helminth infection and treated with a single dose of deworming drugs; although the numbers were small and the quality of the evidence low or very low, there was some evidence of improvements in weight gain, haemoglobin levels, and formal tests of cognition.
The evidence was weaker in trials that mimic the mass medication without screening, with either a single or multiple doses. For single doses, a positive gain in weight was reported in two trials from a single location but not in seven others published since then. Two trials measured cognition: one reported no effect, the other that deworming made cognition scores worse. The two older studies that showed weight gain also found an improvement in physical wellbeing, not surprising since nobody has ever denied deworming has health benefits. Studies of multiple doses, which are recommended by deworming advocates, show no significant weight gain on average and no benefit to cognition; two studies, both in Kenya, showed school attendance 4% higher, a result that was not significant. Over longer follow-up periods the results were similar: no studies showed cognitive benefits, one showed weight gains, and one showed school attendance 5% higher, which again was not significant. In almost all cases the quality of the evidence was low or very low.
The authors concluded that screening and treating children “appears promising, but the evidence base is small.” Routine deworming had not shown benefits on weight in most cases, except for three studies conducted 15 years ago or more. For haemoglobin and cognition, community deworming “seems to have little or no effect,” and the evidence in relation to school attendance and school performance is “generally poor, with no obvious or consistent effect.”
The review concludes: “Our interpretation of this data is that it is probably misleading to justify contemporary deworming programmes based on evidence of consistent benefit on nutrition, haemoglobin, school attendance or school performance as there is simply insufficient reliable information to know whether this is so.”
These conclusions remain little changed from the 2000 review, which prompted a number of rapid responses when summarised in the BMJ.9 One response came from Dr Don Bundy of the World Bank, the person most strongly identified with pushing deworming programmes, and Professor Richard Peto of Oxford University. They said: “The remarkable cost effectiveness of deworming derives not from some easily measured and immediate clinical benefit of a single intervention but from the longer term preventive value of an annual investment of less than 7p.”
Long awaited data
At the time, the two were following up a trial carried out in the Indian city of Lucknow in the mid-1990s that had shown significant benefits in children living in slums who were given vitamin A every six months, half of whom were also given albendazole.10 After two years the albendazole group showed greater weight gains of a highly significant 1 kg. Peto explains that he and Professor Richard Doll had won a prize worth £500 000 and had decided to spend it on a much bigger trial in India involving a million children, divided into 72 areas and randomised to be given albendazole, vitamin A, both, or neither. The primary outcome was mortality, because it was easy to measure and it was reasonable to assume that any improvement in health would be reflected in a reduction in mortality.
This is by far the largest trial of deworming ever carried out, but despite having been completed in 2005, it is yet to be published. The failure to publish has naturally led others to draw their own conclusions. “This is a very important study,” says Garner. “If the results had shown an obvious positive finding, do you think we’d still be waiting?” That was more than 18 months ago. The study is now in press in the Lancet and is expected to appear in January.
The long delay, Peto says, arose because when the vitamin A part of the study was reported at a conference in 2007, it did not show positive results. “It was received with a deafening silence,” he says. “The WHO’s claim that vitamin A supplementation would knock 25% off mortality couldn’t be sustained. The subject is bitterly polarised in India, where some people have been exaggerating the benefits of vitamin A—not misrepresenting them, but exaggerating—while others have claimed that supplementation may be of zero value, or even toxic.
“So we organised a meeting in Oxford to discuss the vitamin A results, where feelings ran high. Some of them really tore into us, saying our mortality data couldn’t be trusted. We were afraid that if any trivial defects were found in the data, they would be misused to undermine the credibility of the study. So we did a lot more data checks to try to weed out any duplicated records. We didn’t want to be in a position where people could pick holes.”
All this took time, with more than a year spent checking the data. Without that, he said, “we would have been discredited by a malicious attack on details.” He blames himself, and a shortage of resources to complete the trial, for the delays. “I agree it isn’t right that things should be delayed like this, but we failed to foresee the vast amounts of work that would be involved.”
The albendazole findings, which were shared with the Cochrane team 18 months ago, show that the intervention was successfully delivered, with good compliance and a decline in the levels of worm infestation but no significant effect on either weight gain or mortality. “It’s a beautiful study,” says Peto sadly. “It’s a pity the results are what they are.”
Although unpublished studies risk distorting the literature, published studies also need careful reading, Garner says. In 2006 the BMJ published a study led by Harold Alderman, a World Bank economist, and coauthored by Bundy, which reported a significant weight gain in Ugandan children aged 1-7 years who had been given deworming pills.3 The paper reported that the children put on 10% more weight when treated twice a year, and 5% when treated once a year.
But the statistical analysis was faulty. The authors had analysed the data on extra weight gained as if the 27 955 children had been randomised individually rather than in clusters of 50 parishes. When this is corrected for, the confidence intervals are wider and the result no longer significant. In response to an email from Garner, Alderman conceded the mistake, and a correction has now been published on bmj.com. However, the authors maintain that the study’s main conclusion was based on the multivariable analyses, which were adjusted for the study design. The BMJ has asked them for further clarification.
Value for money
The charity GiveWell, which helps donors by rating other charities for value for money, also found an error in Disease Control Priorities in Developing Countries, a major report funded by the Gates Foundation and considered to be gold standard evidence. The report estimates that the cost effectiveness of deworming is $3.41 per disability adjusted life year, making it one of the most cost effective interventions for global health. But close examination of the figures by Alexander Berger of GiveWell has established five separate errors on the spreadsheet that was used to make the calculation.11 The true figure, based on the same data, is $326.43—nearly a hundred times as much. “We find it a source of major concern that these errors stood uncorrected in the several years between the publication of the report and our 2011 investigation,” GiveWell says.
Its revised figure puts deworming on the same cost effectiveness footing as treating drug resistant tuberculosis or providing family planning services and makes it worse value for money than providing vaccination or insecticide treated bed nets. GiveWell concludes, however, that that may be too harsh a judgment, and it still rates deworming as a worthwhile health intervention.12
Bundy has proved reluctant in the past to concede any ground to critics of deworming, despite coauthoring with Peto the Indian study that found no significant effects. In a blog on the World Bank website in July 2011, he claimed that some experts (unnamed) believed deworming to be the “closest we have come to finding a magic bullet” in education policy.13 “While deworming has proven to be one of the most cost-effective interventions to get children into school, promising new research suggests that deworming children can also result in many long-term benefits, including higher wages, healthier individuals and stronger communities” he added.
In an interview with the BMJ in which he made clear he was speaking in a personal capacity and not as a representative of the World Bank, he conceded some ground. “I think certainly my views have changed as the evidence hasn’t come in,” he said. “Well meaning people overstated the case. There is a good deal that has been written that isn’t substantiable, and that has led to a polarisation of views.”
Where people are screened and then treated, he said, they show considerable improvements. “What’s less apparent is if you take a whole population and treat them all, you’ll get results that are measurable. When averaged over the whole population, it doesn’t show effects.”
But some studies, he said, have produced compelling results, citing the work in Kenya of Edward Miguel of the University of California at Berkeley and Michael Kremer of Harvard, who found better school attendance among children treated for worms. The Cochrane review includes this study but points out that the differences are small. Its view of this frequently cited study, on which the whole enterprise has come increasingly to rest, is that because the researchers did not compile any baseline data about the schools randomised to get treatment or placebo, it is impossible to say if any subsequent differences are due to the treatment or to different baseline values. With others, Miguel and Kremer have followed up the children, reporting in an unpublished study that they subsequently had higher incomes than those left untreated.14
Bundy cites this study in a forthcoming review15 of the subject in Trends in Parasitology, together with another unpublished study carried out by Owen Ozier when at the University of California at Berkeley (he is now on the staff of the World Bank).16 Ozier found that the benefits were shared by children who were too young to be treated at the time but who lived in communities that were treated. Bundy and Ozier attribute this benefit to the treatment having reduced the spread of infection in these communities, but it could also be explained by the treated communities differing from the untreated ones at baseline.
Bundy’s conclusion in his new review is that “the paucity of randomised trial data suggesting benefit does not equate to data suggesting a lack of benefit.” Will deworming continue, the BMJ asked him? “Yes. Costs have substantially declined, with the availability of free treatments, and are now exceedingly low,” he replied. “Deworming should continue to be delivered alongside other drugs. Those people who have worms are extremely keen to get rid of them.”
Cite this as: BMJ 2012;345:e8558
Competing interests: The author has completed the ICMJE unified disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declares no support from any organisation for the submitted work; no financial relationships with any organisation that might have an interest in the submitted work in the previous three years; and no other relationships or activities that could appear to have influenced the submitted work.
Provenance and peer review: Commissioned; not externally peer reviewed.