Drug companies should prove their products are completely safe before they sell them

I suspect that most people agree with this statement. I also suspect that journalists would be very critical of any politician who disagreed with it!

But how can we prove that a medicine will always be safe for everyone?

Let’s forget ethical considerations and pretend that it would be acceptable to test drugs on people. If we gave the drug to one person there may not be an adverse effect. But it could adversely affect some-one else because we are not all the same in the way in which we react to things. For example, some people react adversely to common foods like bread (because it contains wheat flour) and nuts. But this doesn’t mean that the rest of us shouldn’t eat them.

So, to overcome this problem, we could test the drug on many people. But how many? Adverse reactions could be very rare. Suppose the drug is tested on 1000 people and there are no problems; we can’t be sure, if the number had been increased to 1001, that the last person would not have suffered an adverse effect.

All that we can do is to test the drug a very large number of times and identify any adverse effects that may occur in some people. Most drugs will have some adverse effects because they are chemicals that don’t naturally belong in our bodies.

Whether it is safe for you to take the drug then becomes a question of probability. If only 5 people in a 1000 tested showed any adverse effects, the probability that it is safe for you is 5/1000 = 1/200, i.e. 0.5%. Of course tests could also identify groups of people who are especially at risk and so should avoid taking the drug. But they can’t prove that the drug will be safe for everyone else.

 

Follow-up posts

16.10 Will driving an expensive car make you healthier?
16.36 Good and bad
16.46 The placebo effect
16.47 Fats
19.1 Hard water
19.4 Dietary information