- 1.Randomized trials should be considered when:
There is uncertainty regarding the effect of an exposure or treatment
The exposure can be modified in a trial setting
Phase I and phase II studies evaluate the tolerability and biological activity of a drug; phase III and phase IV studies are randomized trials that evaluate clinical endpoints.
- 3.Potential limitations of randomized trials include:
Limited generalizability of the study population
Limited generalizability of the study environment
Randomized trials address a narrow study question
Randomized design accounts only for confounding
Common measures of effect in randomized trials are relative risk and risk difference.
The number needed to treat or harm = 1/risk difference.
The intention-to-treat analysis predictably leads to bias toward the null.
- 7.Criteria used to judge whether results of a subgroup analysis are valid include:
Biological plausibility for a particularly strong effect in the subgroup
The subgroup analysis was pre-specified
Reasonably large number of outcomes in the subgroup
KeywordsBone Mineral Density Risk Difference Impaired Kidney Function Subgroup Finding Pain Reliever