Assessing the Credibility of Findings From Nonrandomized Studies of Interventions

Randomized clinical trials (RCTs) are widely accepted as the gold standard in the assessment of health interventions. However, evidence from RCTs is often unavailable. Randomized clinical trials may also be impractical, for example when the outcome of interest is so rare or so long term that appropriately sized studies would be too expensive. In addition, RCTs require clinical equipoise, and withholding established treatments may be considered unethical even if evidence for the effects of these treatments is limited. Finally, RCTs may be underway but not ready for reporting for some years.