The review was based on broadly defined inclusion criteria. Searching encompassed a range of resources but was limited to publications in English, which opened up the possibility of language and publication biases. Two reviewers selected studies, extracted data and assessed study quality (which helps to minimise bias and error). A narrative synthesis was appropriate given the differences between studies in participants, interventions, outcomes and theories. Study quality was assessed and used to inform the inclusion of studies in the review.
The role of a range of intervention variables (such as use of a theoretical framework or use of tailored feedback) was examined in the synthesis. Conclusions on effectiveness were based on counting the numbers of studies with statistically significant results.
The authors' preliminary conclusions on the potential of computer- and web-based interventions are fair but effects tended not to be maintained after the intervention and that most of the outcomes in the review were based on self-report from an intervention that was not blinded.