Thanks to scientists working under the auspices of the World Health Organization, you can be fairly sure your toothbrush won’t give you cancer. Over four decades, a WHO research agency has assessed 989 substances and activities, ranging from arsenic to hairdressing, and found only one was “probably not” likely to cause cancer in humans. It was an ingredient in nylon used in stretchy yoga pants and toothbrush bristles.

All the other 988 substances, however, pose some level of risk or need further research, according to the International Agency for Research on Cancer (IARC), which is an arm of the WHO. Some things in IARC’s top category of carcinogens are pretty obvious nasties, such as plutonium, mustard gas and smoking tobacco. Others are more surprising: Also ranked as “Group 1 Carcinogens” are wood dust and Chinese salted fish.

IARC has said that working as a painter causes cancer, using a mobile phone possibly does, and working shifts as a pilot or a nurse, for example – is “probably carcinogenic.” Last October, it ranked processed meats in its top category of known carcinogens, alongside plutonium.

The findings have caused consternation, not least for non-scientists puzzled by what IARC’s rankings mean.

As a global authority on cancer – a disease that kills more than 8 million people a year worldwide, with more than 14 million new cases appearing annually – IARC has enormous influence and commands much respect, even among its critics. Yet experts from academia, industry and public health say IARC confuses the public and policymakers. Some critics say the way IARC considers and communicates whether substances are carcinogenic is flawed and needs reform.

Even the WHO, which oversees IARC, was caught off guard by the agency’s announcement that red and processed meat should be classified respectively as probable and known carcinogens. The WHO’s official spokesman, Gregory Hartl, issued a statement saying WHO’s Geneva headquarters had been flooded with queries and requests for clarification. IARC’s ruling did not mean people should stop eating meat, he said.

Asked about the relationship between IARC and the WHO, Hartl told Reuters: “WHO works closely and continually with IARC to improve the way the two bodies collaborate and communicate on the knowledge of potential and real hazards and risks to the public.”

At stake are judgments that can affect the lives of millions of people and the economic activities of states and multinational companies. IARC’s rulings influence many things, from whether chemicals are licensed for use in industry to whether consumers choose or spurn certain products or lifestyles.

But its methods are poorly understood and do not serve the public well, according to Bob Tarone, a statistician formerly at America’s National Cancer Institute and now Biostatistics Director at the International Epidemiology Institute. He said of the way IARC works: “It’s not good for science, it’s not good for regulatory agencies. And for people? Well, they are just being confused.”

Paolo Boffetta worked at IARC for 19 years, rising to become head of the genetics and epidemiology team, and describes himself as “still a strong supporter” of the agency. Nevertheless, Boffetta, now at the Mount Sinai School of Medicine in the United States, said IARC’s approach sometimes lacks “scientific rigour” because its judgments can involve experts reviewing their own research or that of colleagues.

Some institutions have also clashed with IARC. The agency is currently embroiled in an acrimonious dispute with the European Food Safety Authority (EFSA) over glyphosate, an ingredient of widely-used pesticides. IARC says glyphosate is “probably carcinogenic.” EFSA says it isn’t. The glyphosate row has thrown up concerns about potential conflicts of interest at IARC: It involves an adviser to the agency who is closely linked to the Environmental Defense Fund, a U.S. campaign group opposed to pesticides.

In the face of its critics, IARC steadfastly defends its methods and aims. “This is really the strongest possible process,” Kurt Straif, the head of IARC’s classification programme, told Reuters when asked about the way his agency evaluates possible causes of cancer.

IARC’s director, Chris Wild, has also defended the agency against criticism in scientific journals. In a letter to one of the journals, he said the scientists involved in its classification decisions “are motivated by a desire to improve public health by identifying the causes of human cancer and thereby contributing to disease prevention.”

Richard Sullivan, a professor of cancer policy and global health at King’s College London, says any confusion is due to a widespread misunderstanding of IARC’s role.

“IARC is purely there to do the science. And the science is absolutely fine,” he told Reuters. “But there is a disjunction between the pure science and the policy and public health messaging. That’s where problems arise.”

SEMI-DETACHED AGENCY

From the beginning, IARC has been a compromise. Born out of a French initiative, it was originally envisaged as an independent agency with a huge budget. It ended up as a semi-autonomous part of the WHO with modest funds. IARC, based in Lyon, had revenue of about 30 million euros ($34 million) in 2014, whereas the British charity Cancer Research UK had income of about $875 million and the U.S. government’s National Cancer Institute had a budget of $4.9 billion in 2014.

Despite its limited financial heft, IARC was a pioneer and established itself as a world-leading authority. Its assessments of whether something is a cause of cancer catch the eyes and ears of policymakers and the public.

To produce its assessments, IARC assembles groups of experts who review the existing scientific evidence and then place a substance or activity in one of five categories: carcinogenic to humans; probably carcinogenic; possibly carcinogenic; not classifiable as carcinogenic; and probably not carcinogenic. These reports are known as “monographs.”

The public sometimes misunderstands what IARC means by its classifications. The agency says it assesses “hazard” – the strength of evidence about whether a substance or activity can cause cancer in any way. It takes no account of typical levels of human exposure or consumption. So it is not measuring “risk” or the likelihood of a person getting cancer from something.

IARC gives no view on the relative levels of risk of getting cancer from, say, plutonium or alcohol; what it does say is that there is clear evidence that both are capable of causing cancer. Therefore it ranks both substances in its top category of being carcinogenic.

Geoffrey Kabat, a cancer epidemiologist at the Albert Einstein College of Medicine in the United States who has publicly criticised IARC, says the classifications do the public “a disservice.”

“What the public wants to know is: What are the agents in our surroundings that are likely to have palpable effects on our health? Not theoretical exposures which might, under some far-fetched conditions, possibly have an effect,” said Kabat, who is also author of the book “Hyping Health Risks.”

The risks of public misapprehension were evident in some of the media reaction to IARC’s announcement on red and processed meats. The Huffington Post declared: “Meat is the new tobacco.” Britain’s Daily Mail said “health chiefs” had “put processed meat on same level as cigarettes.”

Such interpretations are misleading, in IARC’s view. Straif told Reuters that the blame for any confusion lies with industry, activist groups and the media.

“There are stakeholders on various sides that want to make it look ridiculous,” he said. “There are activist groups who want to say, ‘This is now an IARC carcinogen and we need to take all actions against it’. And then there is a third dimension – the media, who have their own interests in being sensational.”

Straif defended the decision to place processed meat in the same hazard category as plutonium, saying that “for both of these things there is clear evidence that these are human carcinogens.”

“NAIVE, IF NOT ANTI-SCIENTIFIC”

Some critics say the problems with IARC’s monographs begin well before they become headlines. Their concerns focus on the composition of the “expert working groups” that decide which of the five IARC categories a substance or activity should go in. These experts sometimes include people who have spent years publishing research on whether the substance or activity under scrutiny can cause cancer. They may be part of IARC working groups that review their own research or that of close colleagues.

Between 2012 and 2015, for example, IARC published or started 18 monographs involving 314 scientists. A Reuters analysis found that at least 61 of those scientists served on monograph working groups that considered their own scientific research. The analysis did not include the number of scientists on working groups that reviewed the research of close colleagues.

In letters, commentaries and articles in scientific journals, Tarone of the International Epidemiology Institute and other scientists have questioned whether such people “are the best judges of the validity and methodological soundness of their own and allied work.”

IARC’s Straif said the agency’s working groups consist of “the world’s best experts” who critically review the scientific evidence and are not swayed by previous findings in their own work, or that of close colleagues. “IARC has a strong belief, for good reasons, that those who know the most about certain exposures are those who have worked on such exposures,” he said.

Straif said IARC’s rules ensure no “author or associated colleague” can directly evaluate a study they have published. And neutrality is assured, he said, because the discussions involve 20 to 30 people in an environment where “advocacy of any kind … is not tolerated.”

Tarone regards IARC’s assumption that all experts will be detached and independent as “naive, if not anti-scientific.” He told Reuters: “It’s absurd to assert there are no issues of bias related to self-interest, reputation or careerism. It has nothing to do with bad motives, it’s just human nature.”

Tarone and other critics say IARC is inconsistent in its treatment of potential conflicts of interest and cite as an example a study of radiation of the sort emitted by mobile phones. In June 2011, IARC concluded such radiation is “possibly carcinogenic.” That ranking put mobile phone use in the same category as lead and chloroform.

Anders Ahlbom, a senior professor at the Karolinska Institute in Sweden, was originally invited to chair the working group on electromagnetic radio frequencies in May 2011. However, Ahlbom, who thinks there is little evidence to suggest mobile phones cause cancer, was asked to step down about a week before the meeting was due to begin after he told IARC he had been contacted by a journalist. The journalist had questioned him about being on the board of his brother’s consulting firm, which helps clients lobby on telecoms issues.

IARC decided that Ahlbom had a perceived conflict of interest. Ahlbom accepted this decision, although he said there was no conflict since he had no financial interest in his brother’s company. Ahlbom said his departure upset the balance of the IARC working group, which, he and five other scientists said, included researchers who already viewed mobile phones as likely to raise the risk of brain tumours.

“It appears that IARC handles conflicts of interest differently depending on who the person is and which ‘side’ he is assumed to represent,” Ahlbom told Reuters.

Straif said IARC “takes all conflicts of interest seriously, regardless of the individuals or organisations involved.” He said that Ahlbom’s exit from the working group did not leave it with an imbalance.

“It is difficult to perceive how a strong working group of 32 internationally renowned experts would suddenly not have a balance anymore because of one single expert who had a conflict,” he said.

SEEING RED

At the meetings of IARC working groups, invited observers who have “relevant scientific credentials” are allowed to attend; but they have to sign confidentiality agreements and cannot discuss the proceedings. Straif says this is to ensure the scientists can speak candidly, without fear of having their disagreements or discussions reported externally without their consent.

One observer, a specialist in food and animal science who attended the working group on red and processed meats in 2015, spoke to Reuters on condition of anonymity. The person alleged that the expert panel reviewing the scientific evidence appeared to aim for a specific result.

In its meat assessment, IARC went beyond its normal remit of assessing hazard, not risk. It gave specific warnings about the risk of eating red and processed meat products.

IARC said, for example, that for each 50 gram piece of processed meat eaten daily, the risk of a person developing colon cancer increases by 18 percent. The observer who spoke to Reuters said these data appeared “to come from nowhere, overnight.”

The observer said: “I expected that the science would be reviewed with a high level of rigour. But quite frankly, at the end of the 10 days, from a scientific standpoint I was really quite shocked.”

Straif said the numbers came from “a combined analysis” of the scientific papers under review, and were issued by IARC because there was sufficient evidence in human epidemiological studies for the working group experts to feel confident in them.

Straif told Reuters parts of the working group discussions may have been missed by some observers: “We really worked around the clock, up late into the night and all weekend, so I’m not sure if the observers were there at all times.”

In a subsequent email to Reuters, he said the risk estimates and corresponding scientific papers were part of the monograph discussions from “the very first working drafts and through all revisions.” He added: “It is very difficult to understand how any participant could have missed this discussion.”

While not disagreeing with IARC’s assessment of meat as a carcinogen, the WHO headquarters issued a series of tweets giving context. The WHO stressed that “the health risks of processed meat are vastly different of those cigarettes and asbestos” and that “meat provides a number of essential nutrients and, when consumed in moderation, has a place in a healthy diet.”

The controversy has raised questions at WHO headquarters about the organisation’s control over IARC. “There is talk here now of needing to rein IARC in,” said a Geneva-based WHO insider.

Charles Clift, a global public health specialist at the Centre on Global Health Security at Britain’s Chatham House, said the WHO should have taken more of a role in IARC’s presentation of its conclusions on red and processed meat.

“The WHO should be there to give authoritative guidance,” Clift said, “not just endorse things that can be misinterpreted – either from IARC or anybody else.”

The WHO’s spokesman, Hartl, told Reuters that IARC was a “functionally independent” agency, and that when IARC flags up cancer hazards, the “WHO assesses or re-assesses the levels of risk associated with those hazards. Based on the risk assessment, WHO reaffirms existing or issues new guidance aimed at safeguarding public health.”

Straif of IARC said: “I’m very happy with the way we do things at the moment. We are really at the head of the scientific community.”