The methodological quality of the primary studies included in a systematic review may influence its results and final conclusions. Methodological quality may be defined in various ways. Partially because of this there are many different assessment lists. The most important dimension of quality is internal validity, defined as the confidence that the design, performance and report of a trial prevent or reduce systematic errors (bias) in the outcomes. For only a limited number of internal validity items a relationship with bias has been proven in empirical studies: concealment of randomisation and blinding of patients and outcome assessors. Preferably, quality should be assessed by at least 2 assessors independently. There is no consensus whether assessment should be done blinded for authors, journal, results and conclusions. Internal validity can be incorporated into statistical pooling in various ways: as a selection criterion, to be used as weight or to hierarchically order studies in a presentation. Well-designed comparative studies are needed to provide clearer guidelines for methodological assessment in the future.