A study published recently (Tatonetti et. al. 2012) has looked at the huge problem of detecting adverse side effects of medications and in detecting the potential problems caused when one drug interacts with another if the prescribed courses run at the same time. The paper is summarised by Heidi Ledford in a recent copy of Nature magazine.
The US drug control body Food and Drug Administration (FDA) are the recipients of hundreds of thousands of reports of adverse side effects every year and the new research provides a way to sift those reports and arrive at a listing of new adverse events.
This is not an easy task. There are many variables to take in for account whenever an adverse event is reported; age, gender, health status all vary tremendously and profoundly affect conclusions.
The example Heidi uses is one where there are high rates of heart attack reported in people taking a particular drug (compared with the general population). That sounds like an adverse effect of that drug! However if we delve a bit deeper we find that only older people take the drug, and of course amongst older people there are far higher rates of heart attack - the impression that this drug is causing heart attack as a side effect is merely a coincidence.
How do we tell if an adverse effect is the real thing? In this case we would compare older people who don't take the drug with older people who do take the drug - if there are still more cases of heart attack amongst the people taking the drug then we have found a genuine adverse effect.
The computer software developed by Tatonetti et. al. looks at all of the reported adverse effects and carries out comparisons using corrections for the type of bias I just described. This approach has successfully identified hundreds of previously unknown side effects and drug interactions in only 1332 of the drugs currently on the market. Application of this software should reveal many more when run against all of the drugs currently on the market, making us all more aware of the hazards of taking all drugs, separately OR in combination.