Cautions in Interpreting Study Results

Animal or laboratory study: Because the study was based on animals, researchers cannot be certain how well the findings will apply to people.

Small study: These findings are based on a small study population; larger studies are needed to understand how well the intervention really works.

Uncontrolled study: Everyone in this study took the drug X. By not including patients who did not take the drug, it is impossible to be sure how much (or even if) drug X accounted for the findings.

Controlled but not randomized study: Because the study was not a true experiment, researchers do not know whether it was drug X or something else about the people who happened to take drug X (for example, were they younger? less likely to smoke?) that accounted for the differences observed.

Any intervention study: The benefit observed should be weighed against the adverse effects (or other downsides such as inconvenience or cost). Important adverse effects should be specified and quantified; if there are none, this should be reported.

Unpublished scientific meeting presentations: The findings presented are from work in progress. Because a full study report has not undergone peer review, the results have yet to be independently verified and may change.

Source: Steve Woloshin, M.D., M.S., and Lisa Schwartz, M.D., M.S., Medical Journal of Australia, June 2006