When should research in a given area end?

Steve Simon

2010-07-26

Categories: Blog post Tags: Systematic overviews

Someone asked a rather philosophical question, is there ever an end to research in a given area? Will there ever be a “last word” on a research topic. Here’s what I wrote in response.

In theory, once the confidence interval for a meta-analysis lies entirely inside the range of clinical indifference, all research should stop for futility. Also in theory, once the confidence interval for a meta-analysis lies entirely above the minimum clinically important difference, then any more research is unnecessary because the case is proven.

This is not done in practice, and leads to a lot of redundancy in research. There are suggestions that Institutional Review Boards should demand a systematic overview prior to approval of research so that they can determine if the proposed research will have potential value or if the area is pretty much settled.

Now the fly in the ointment is that in many areas, there are too many complicating factors to rely just on a simple confidence interval. This is especially true for observational research, where many of the assumptions that you need to make to assure a valid confidence interval are untestable. As we learn more about an area, some of the untestable assumptions may seem more reasonable or less reasonable and may be cause for re-opening the investigation.

You can find an earlier version of this page on my original website.