I consulted my husband Bryan, who has a degree in philosophy. He explained some elementary points about false induction in logic. For example, pink flamingos have backbones. Most pets have backbones. Ergo, most pets are pink flamingos. You start with the particular (flamingos and backbones) and falsely induce the general.
Although it's been used for some 100 years to fight pain, fever and inflammation, in the last 25 years, medicine discovered that aspirin could stop blood from clotting. Although studies into this capability had been carried out, they were small and not particularly conclusive. So, in the late 1980s, the Antiplatelet Trialists' Collaboration (ATC), headed by a prestigious group at Oxford, combined these disparate little studies into what is called a 'systematic review'.
In such a review, the reviewers use rigorous methods to synthesise and appraise all relevant studies from the body of published medical research. It stems from a noble impulse to construct a beach from so many individual grains of sand. In this case, the land to be staked out was nothing less than the prevention of many of the West's biggest killers.
In 1988, the ATC published their first review, which concluded that low-dose antiplatelet treatment could prevent stroke or heart attack by 25 per cent in people who'd had either before (BMJ, 1988; 296: 320-31). A few months later, the ATC delivered its knock-out punch. After combining 145 randomised studies, they concluded that low-dose aspirin was effective for preventing heart attacks and stroke in healthy people as well (BMJ, 1994; 308: 81-106). Hard on its heels was another ATC study (BMJ, 1994; 308: 235-46), showing that antiplatelet therapy was also effective as a just-in-case measure against deep vein thrombosis or pulmonary embolism in patients at risk.
Aspirin blocks platelets. All cardiovascular events proceed from platelet-clotting. Therefore, aspirin prevents heart disease and stroke.
One or two dissenters pointed out the logic problems inherent in the systematic review. A group from the Thrombosis Research Institute in London pointed out that one ATC review included a load of poorly designed clinical trials, and had tried to compare apples with oranges. They'd even mucked up their arithmetic percentages (BMJ, 1994; 309: 1213-5). In fact, a year after the study, The Cochrane Group set up a Cochrane Collaboration Stroke Review group precisely because current reviews were felt to be small, unsystematic and either incomplete or biased. On examining a number of individual trials, no benefit is actually found (Lancet, 1992; 340: 1421-5; Circulation, 1980; 62: 449-61; JAMA, 1980; 243: 661-9).
When the large-scale Thrombosis Prevention Trial was finally carried out, which included more than 5000 men at high risk of heart disease, although aspirin or the equivalent did significantly reduce heart attacks, this was balanced out by a significant increase in stroke. The net result was nil. Overall mortality did not change (Lancet, 1998; 351: 231-41).
Doctors are particularly fond of euphemisms. When aspirin does not work, they speak of aspirin 'resistance', as though you are being stubbornly intractable about taking your medicine and foolishly 'resisting' it - as though choice plays a part here.
But it turns out that aspirin, like so much of medicine - as Leonard Cohen puts it - is just one more Joseph looking for a manger.