Clinical trials have a number of characteristics that make them the most reliable way of assessing the efficacy and safety of medical interventions.
Clinical trials use randomisation to produce groups which are comparable for both known and unknown factors influencing outcome.
Studies are blinded so that biases of participants or investigators cannot influence assessments of efficacy or safety.
Endpoints are pre-specified and consistently defined.
Analyses are planned a priori to prevent data dredging with a view to generating specific outcomes.
Studies are powered to be able to detect a clinically meaningful difference in endpoints.
These safeguards allow a high level of confidence in the reliability of any statistically significant difference found between the treatment groups. No other form of clinical investigation has comparable rigour. Observational studies have a role to play in hypothesis generation and in monitoring events too rare to be assessed in trials. The repeated instances of major discrepancies between trial outcomes and analyses of observational data reinforce the belief that all major treatment policies in medicine should be based on clinical trial results.