Basing medical practice on science helps us avoid the pitfalls of relying on our own reasoning and experience. If I want to start a patient on a new medicine, the individual characteristics of the patient are important (Is the drug meant for their condition? Will it interact with other drugs they are on? Are they allergic to it? Can they tolerate it?) but at least as important is how the drug performs when used on large numbers of people. This attenuates the large differences that can be seen among individuals, and allows us to predict how in general the drug will act.
One of the metrics we can use is something called the “number needed to treat/harm” (NNT). This is a nice measure which is a bit more intuitive than other statistics. For example, in the recently released USPSTF recommendations on mammography, it was reported that nearly 2000 women in their 40s needed to be screened to avert one breast cancer death (the number in the 60′s age group was in the 300s). Deciding what that means is a value judgement. Is it worth it to screen so many for each life saved? Is it worth the pain of chasing down abnormal finding that are ultimately found to be benign?
These are questions that many of us will be asking over the next few months; every primary care doc in the states is already fielding calls. Whatever we decide to do with this data, it is the statistical analysis combined with societal and individual values that will help us decide what to do about breast cancer screening. We tend to think that our individual experience gives us more data than it really does. A woman may decide to have a screening mammogram based on the mistaken idea that she is more likely to benefit from it than most women. A similar cognitive fallacy affects us when contemplating flu shots.
One comment I hear daily is, “I’m not getting a flu shot; I never get sick.” This is what is sometimes called the “hot hand” or “reverse gambler’s” fallacy. Setting aside the problem of confounding “flu” with other illnesses, this kind of thinking is a problem. Whether one gets the flu in a given year is independent of whether one has gotten the flu in other years (mostly). In years such as this one, where the current strain has found a particularly susceptible population, the fallacy is even more likely to get you in trouble. Our own experiences are powerful, but often deceptive. With medical decisions, it’s often useful to write out the question, examine the data, and ask for an outside opinion, because our own judgments often lead us astray.