On its own, a single piece of research can be misleading. Separate but similar small studies can produce apparently conflicting results, often due to chance.
Collecting information from different trials in an organised way is a good method of showing what the overall evidence is. By using a careful "study of studies", called a systematic review, it is possible to distinguish the effects of treatment from the effects of chance.
For example, the first trial to show the benefits of a short course of steroid injections given to pregnant women at risk of having a premature birth was carried out in 1972. A number of later trials had results that were not quite so positive.
Only in 1989 was a "study of studies" of all the trials carried out. It found that babies born to mothers who took steroids were much less likely to die than those whose mothers did not take steroids.
It was not new research that made the difference in this case, but collecting together results from all the existing research.
What are systematic reviews?
Systematic reviews are now an established part of research. They look at all the studies on a topic, not just selected studies that may have a particular point of view and could result in biased conclusions.
For example, contributors to the Cochrane Collaboration, the Centre for Reviews and Dissemination and the Health Technology Assessment programme review all the reliable research about preventing and treating specific health problems and publish it to help patients and clinicians make choices in healthcare.
Systematic reviews often combine findings from separate but similar studies and calculate an overall result. This process is called meta-analysis.
Systematic reviews and meta-analyses are now an important part of health research, because they can identify findings that might otherwise be missed in individual studies.
Research should take place only when a systematic review of previous research has been carried out, and the need for new research has been established.
Read more about systematic reviews on the Testing Treatments interactive and James Lind Library websites.
