Systematic reviews as a tool for planning and\ud interpreting trials
Bath, Philip M.W.
Gray, Laura J.
Background Systematic reviews followed by ameta-analysis\ud are carried out in medical research to combine the results of two or more related studies. Stroke trials have struggled to show beneficial effects and meta-analysis should be used more widely throughout the research process to either speed up the development of useful interventions, or halt more quickly research with hazardous or ineffective interventions.\ud \ud Summary of review. This review summarises the clinical research process and illustrates how and when systematic\ud reviews may be used throughout the development programme.\ud Meta-analyses should be performed after observational\ud studies, preclinical studies in experimental stroke, and\ud after phase I, II, and III clinical trials and phase IV clinical surveillance studies. Although meta-analyses most commonly work with summary data, they may be performed to assess relationships between variables (meta-regression) and, ideally, should utilise individual patient data. Meta-analysis techniques may alsoworkwith ordered categorical outcome data (ordinal meta-analysis) and be used to perform indirect comparisons where original trial data do not exist.\ud \ud Conclusion Systematic review/meta-analyses are powerful\ud tools in medical research and should be used throughout\ud the development of all stroke and other interventions