Systematic reviews using individual patient data: a map for the minefields?

An earlier paper in this series showed why systematic reviews are more reliable than other forms of review, such as the narrative commentary [1]. This paper takes this further by examining why systematic reviews which use individual patient data (IPD) are regarded as more reliable than other forms of systematic review [2]. It illustrates how the IPD approach can help achieve the ultimate aim of any systematic review of randomised controlled trials: the inclusion of all randomised patients in all relevant trials, analysed in accordance with the treatment allocated at randomisation. This paper highlights why IPD reviews might be particularly useful in cancer. It does not argue that all systematic reviews need to be done in this way. The most appropriate way in which to conduct any systematic review depends very much on the health care question being addressed by that review.