A recent publication reported that increasing numbers of systematic reviews are being published and, although standards have improved, many are still poorly conducted and reported, especially non-Cochrane systematic reviews (1). The aim of this study was to assess the quality of the conduct and reporting of systematic reviews undertaken for the United Kingdom (UK) Health Technology Assessment (HTA) programme and published in the International Journal of Technology Assessment in Health Care (2) and compare those undertaken in 2004 and 2014.

A comparative sample of all systematic reviews published in 2004 and 2014 in the UK HTA monograph series was identified by a structured search of MEDLINE in August 2016. After piloting of the form, two reviewers each extracted relevant data. These data were tabulated and summarized.

The search identified twenty-three systematic reviews from 2004 and thirty from 2014. By 2014, compared with 2004, a smaller proportion of treatment (53 percent versus 70 percent) and pharmaceutical (20 percent versus 57 percent) reviews were being published. In 2014, there were much higher percentages of review registrations (70 percent versus 0 percent) and available protocols (90 percent versus 17 percent); increased explicit inclusion of unpublished literature (65 percent versus 39 percent); less frequent use of local checklists (32 percent versus 61 percent) for critical appraisal; more complete reporting of study flow for inclusion (97 percent versus 57 percent) and exclusion (91 percent and 65 percent) of studies; and there were more reviews reporting limitations affecting the review itself (73 percent versus 49 percent). The process had clearly become more reflective and rigorous. However, some previous weaknesses persisted, including the general absence of any assessment of publication bias and the failure to report overall numbers of patients in the review.

Marked improvements can be seen in the conduct and reporting of systematic reviews published by the UK HTA programme as a result of the publication and general acceptance of the PRISMA statement (3) and the increased application of a smaller number of relevant standards.