Every few days, we will be publishing layman’s abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
The article featured today is from Statistics in Medicine and the full article, published in issue 38.22, is available to read online here.
Schuemie, MJ, Ryan, PB, Man, KKC, Wong, ICK, Suchard, MA, Hripcsak, G. A plea to stop using the case‐control design in retrospective database studies. Statistics in Medicine. 2019; 38: 4199– 4208. doi: 10.1002/sim.8215
Spurious results from observational studies often produce false headlines and cause patients to make uninformed and potentially dangerous decisions regarding healthcare. Case-control studies, while helpful in data-starved environments, are shown to provide irreproducible and sometimes inaccurate results.
“We have reached a time when there are better alternatives that should be used to ensure the reliability of findings when enough healthcare data, ranging from both electronic patient records and insurance claims, is available,” says Martijn Schuemie, Research Fellow at Janssen Research and Development.
Schuemie led a study that tested two recent case-control studies, one centered on the risk of colitis with an acne drug and another that focused on the risk of pancreatitis from diabetes drugs. Negative controls, exposures known not to cause the outcomes (colitis and pancreatitis), were used and showed incorrect strong effects between those controls and the outcomes. The same data and controls used in an appropriate study (a Self-Controlled Case Series design in this case) produced the expected effect of showing no association between the negative controls and the outcomes.
“The impact of these case-control studies, when brought to the public, can cause people to stop taking crucial drugs based on erroneous findings,” Schuemie says. “This is not only harmful for those patients, but it also hurts public trust. More reliable designs should be the scientific standard when dealing with existing healthcare data.”