Store cards and ovarian cancer story not quite what it seems…
Most professional footballers have fancy cars so most people with fancy cars must be profesional footballers, say researchers…
A wow-science story got blanket coverage in the UK last week. The stories ran that researchers at London’s elite Imperial College had shown that women with hard-to-spot-early ovarian cancer had bought various pharmacy products in the previous few months. The BBC’s headline was typical; “Shop loyalty data card may help spot ovarian cancer”.
Here’s the actual paper’s methodology section. Researchers looked at the store card purchases made by women in the months before they were diagnosed with ovarian cancer. The results give us the maybe that not surprising news that women who turned out to have ovarian cancer bought more pain medication before they were formally diagnosed.
This of course tells us nothing at all about whether there is a useful correlation between women who increase their purchase of pain medication and those who turn out to have ovarian cancer. There are obviously a million reasons a person might be suffering more pain than usual and so might have an uptick in their painkilling purchases.
My guess is that the college media folk (certainly not the research team itself) worded a press release so that media organisations would leap at a massively newsworthy, but entirely false, implication that researchers have somehow shown that storecard purchases can have predictive validity in respect of medical conditions. But who knows? You can ask them yourself.
