Published On Jun 05, 2015
Does This Cause That?
Whether it’s smoking and cancer or vaccines and autism, the Bradford Hill guidelines celebrate 50 years of tracing diseases to their proper roots.
Did a piece of shoddy medical research cause the measles outbreaks across the United States earlier this year? It’s the kind of question Sir Austin “Tony” Bradford Hill, an English epidemiologist and statistician, would have loved to answer.
“The Environment and Disease: Association or Causation,” which Hill published 50 years ago in The Proceedings of the Royal Society of Medicine—offered what came to be known as the “Bradford Hill criteria” as a way to determine whether a “causal relationship” exists between a specific factor and a disease.
Hill’s criteria cover nine points including such considerations as the strength of a possible association, the consistency of findings (did different researchers come to more or less the same conclusion?), whether things happen in the correct order (the supposed cause obviously has to precede the effect) and whether there’s laboratory evidence to back up what epidemiological observations suggest.
Hill’s path to what would become epidemiology was circuitous. Having contracted tuberculosis during World War I, he wasn’t able to follow his father into a life in medicine (because he might infect his patients). So he took a correspondence degree in economics, and then found his way to the statistics of human health. By the mid-1940s, as one of Britain’s foremost medical statisticians, he had begun planning what would become the world’s first randomized controlled trials—a study to gauge the benefits of a vaccine against whooping cough, and another to test the effects of the antibiotic streptomycin on tuberculosis.
In 1950, Hill and a colleague, Richard Doll, launched modern epidemiology with their efforts to gauge the dangers of tobacco smoke. Using “case control” studies, which compared people who had a particular condition to those who didn’t, their work ultimately demonstrated that smoking could lead to lung cancer.
Their approach was entirely new, says David Morens, an epidemiologist at the National Institutes of Health and past-president of the American Epidemiological Society. “The brain power that went into what these two did was remarkable,” Morens says. “Hill’s work has done more to save people’s lives than 100 Nobel laureates put together.”
Hill officially retired in 1961 (the year he was knighted) but spent the next several years honing his ideas about causality in medicine that he would express in his 1965 treatise. Although some critics contend that Hill’s criteria are excessively rigid, Morens believes the general principles of the framework have great value. “It’s less about the individual criteria than about how to wrap your brain around a huge problem—causality—so that you can say, ‘I think you shouldn’t smoke and here’s why.’”
The recent measles outbreaks provide only the latest example of why that’s so important. They came about as a result of parents refusing to vaccinate their children—a decision based on an anti-vaccination bias that leans heavily on a 1998 study claiming to find a causal link between the vaccine for measles, mumps and rubella (MMR) and autism. There’s no question now that, judged by the Bradford Hill criteria or any other, the infamous 1998 study was wrong. The journal that published the work retracted it, the British doctor who led the study lost his license and countless scientists have dismissed the findings while emphasizing the value of MMR and other childhood vaccines.
Yet champions of the anti-vaccination movement continue to hold to their ideas—which are most elegantly refuted by the 50-year old Bradford Hill report. In trying to establish a causal link that isn’t there, the antivaccine movement violates several Bradford Hill maxims, by giving more weight to anecdotes than to evidence, ignoring when supposed associations are in the wrong order and dismissing the lack of any reasonable biologic link between a vaccine and a harm. “I think the antivaccine people live in an alternate universe where the logic built into the Bradford Hill criteria does not apply,” Morens says.
But did the study that launched the antivaccine movement ultimately cause the measles outbreak? If Hill were able to examine the evidence, chances are he would say that it did.