The challenge of external validity in policy-relevant systematic reviews: a case study from the field of substance misuse

Abstract

Aim: To critically evaluate the methods utilized in the conduct of a systematic review in the field of substance misuse.

Design: Participant-observation in the review process, semi-structured interviews with review team members and management and structured observation of the process of guidance development.

Setting: An 'arm's-length' government body.

Participants: Review team members, management and the committee responsible for producing evidence-based guidance for policy and practice.

Measurements: Data from interviews and (participant-)observation were reflected upon critically in order to increase understanding of the systematic review process.

Findings: The application of systematic review methods produced an evidence base that did not inform the development of guidance to the extent that it could have done: (i) an emphasis upon internal research validity produced an evidence base with an emphasis on short-term interventions at the level of the individual; (ii) criteria for appraising the external validity of studies were not developed sufficiently; and (iii) the systematic review of evidence and development of guidance are strongly reliant upon the judgement of reviewers and committee members.

Conclusions: Prioritizing internal validity in a systematic review risks producing an evidence base that is not informed adequately by the wider determinants of health and which does not give sufficient consideration to external validity. The use of appropriate methods requires that commissioners of systematic reviews are clear at the outset how the review is proposed to be utilized. Review methods such as meta-ethnography and realist synthesis could contribute to making the frameworks within which judgements are made more explicit.