Some basic facts about medical treatments are first explained. New treatments are not necessarily better than existing ones, more of a treatment is not necessarily better, and early treatment is not always good (screening can cause harm by raising false alarms). And there are general problems involved with uncertainty about effects.
Three chapters explain how research can help with decisions about alternative treatment options, introducing the concept of a fair test (a randomized trial), explaining how fair tests can address the play of chance, and looking at combining multiple tests to assess all the evidence (the idea of a meta-analysis).
Testing Treatments proceeds to look at problems with regulation, where the same treatment may be effectively unregulated when used outside a research context, at some examples of good, bad, and unnecessary research, and at the importance of involving patients and lay people in research (but also the problems that can bring). It concludes with some general principles for using research to improve healthcare.
The approach in Testing Treatments is a little dry, but not academic (any formal statistics is avoided) and never turgid: without ever resorting to dramatisation of anecdotes, it makes good use of case studies and quotations in separate boxes. It doesn't try to offer advice on any specific health problems and it avoids entering into debates about the merits of alternative medicine, but is unlikely to appeal to those who insist that gut feelings override any kind of objective testing.
Testing Treatments restricts itself to medical research, with nothing on its applications to the economics of healthcare or broader policy-making. What happens, for example, when we have a number of treatments, of varying effectiveness, but can only afford to fund some of them, for some people? Or we have to choose which research programmes to fund?
This is a clear introduction to an important subject, which one hopes will reach a wide audience.
- Related reviews:
- - books about health + medicine