Overviews of randomized clinical trials.

Therapeutic decisions should obviously be based on all available relevant evidence. Research reviews are important for bringing together and critically assessing such evidence. The conclusions of individuals who perform research reviews are subjective, however, because the determination of what evidence is relevant and how the pieces of evidence should be fitted together is generally not straightforward. Although the findings of studies often do not fit together neatly, discrepancies sometimes provide an opportunity to better understand the factors that influence treatment effectiveness. Consequently, although reviews are not substitutes for well-done clinical trials, individual clinical trials are also not substitutes for well-done reviews.