Skip navigation

Best available truth (Editorial)

Whenever "evidence" is discussed, arguments often break out which seem to Bandolier to be much like baying for the moon. If a review chooses only randomised trials, then what about other types of evidence? If a randomised trial is chosen, then is the trial big enough to draw implementable conclusions? Are case-control or cohort studies good enough, other than from which to develop hypotheses for future testing? What about ethics? What should we do next?

It all seems impossible if every sentence ends with a question mark and every conclusion is "more research". We live in a real world and have to make the best decisions we can. We were helped by a thoughtful Irish biochemist who said that what we are seeking is the "best available truth".

That may be from a great systematic review - as in this issue on analgesics in dysmenorhoea . It may be from a single great randomised trial with useful outcomes - as in this issue on second cataract surgery . It may be from epidemiological work - as in this issue looking at the links between homocysteine and heart disease . But if we know our source is the best available truth, and we update that truth regularly, then we can be confident that we're likely to be doing OK.

That is why electronic libraries are so important, because they can be relatively a painless way of updating ourselves. The Cochrane Library is one, but there are other ways, and electronic library/information resources are delivering the goods, and it is difficult to remember how we lived without them.

next story in this issue