Doctors don’t go to medical school to become expert at interpreting statistics. Yet most of the articles that tell us how safe drugs might be rely on complex statistical analyses that go far beyond what I learned in my one credit course on Statistics for Medical Practitioners. For the most part, doctors have to take the word of the statisticians on whether a drug has a given benefit or harm.
An example is a recent Lancet review article has been hailed as the “definitive” summary of the potential risks versus potential harms of statin drugs. The authors conclude that most of the side effects we’ve thought statins had are actually nonexistent. They predict just 2 adverse events per 15 cases of heart attack or stroke prevented, and make rather strong statements about doctors not prescribing statins often enough.
A couple of folks I respect have pointed out that the authors have long served as apologists for the statin industry. In my video critique of the Lancet article I discuss Malcolm Kendrick, and HealthInsightUK.
My concern about this particular article is that, as far as being the last word on risk versus harm, it’s seriously flawed by failure to include ALL the common side effects in the analysis. Some side effects simply weren’t discussed, like heart failure. But most side effects were dismissed based on the results of randomized controlled trials. However, a little-known stage of most such trials called the “run-in” or “wash out” period, can be used to screen out massive numbers of study subjects. I believe that this little trick is often used to sway the results in favor of a given drug.
View my video critique of the Lancet article here, and let me know what you think. Do you believe doctors should also consider the potential harms from statin side effects that I review in the video, do you think they are irrelevant like the Lancet article claims?