Randomized Controlled Trials are not Always Necessary, Observation May Suffice

I just listen to an interview with Professor Sir Michael Rawlins, Chair of UK’s National Institute for Health and Clinical Excellence. You can find it on itconversations. He goes over very interesting topics. Two of which are deciding on what services should a public health service provide and evaluating evidence.

In evaluating evidence, he raise the point to the needless divide between using observation and using randomized controlled trials (RCTs). Both are needed methods for evaluating evidence. However, RCTs can be unnecessary and even unethical to do. He mentions the example of ganciclovir used to treat cytomegalovirus retinitis in AIDS patients. Where although observational study showed efficacy, RCT was carried out. after a year the benefit was shown, but the non-treatment arm got blind! (I need more time to dig up this study) This is an article commenting on Sir Rawlins views.