A scoping review of prospective meta-analyses in health research


Oral session: Overviews, rapid reviews, and other types of evidence synthesis (2)


Wednesday 23 October 2019 - 14:00 to 15:30


All authors in correct order:

Cheyne S1, Seidler AL1, Hunter K1, Ghersi D2, Berlin J3, Askie L1
1 NHMRC Clinical Trials Centre, University of Sydney, Australia
2 National Health and Medical Research Council and the University of Sydney, Australia
3 Johnson & Johnson, USA
Presenting author and contact person

Presenting author:

Saskia Cheyne

Contact person:

Abstract text
Background: prospective meta-analyses (PMA) may reduce many of the issues that can occur in traditional (retrospective) meta-analyses. They can reduce biases in publication and selective reporting. Yet, to date there is no clear understanding of the definition of PMA or how to report a PMA. A summary of PMA literature is needed to gain greater clarity and help inform future guidance and reporting for those intending to conduct and publish PMA.

Objectives: to identify and describe the key features, methods and reporting characteristics of PMA in health research.

Methods: we searched for studies using search terms derived from previously identified PMA, and by consulting topic experts. We systematically searched PubMed, Embase, Cochrane Database of Systematic Reviews and PROSPERO, and performed grey literature searches. One review author screened the search results, and a second review author screened a sample, to identify any potential PMA or methods articles on PMA. One review author then extracted data, which a second review author checked. We then used these data to create a survey, which we sent to authors of potential PMA.

Results: title and abstract screening identified 1056 articles, of which we screened 274 at full-text. Of these, we identified 51 as potential PMA and contacted authors. Of the 23 who responded to the survey 58% were confirmed as PMA. Of these, some reported not having a published protocol and had varied reporting on details such as committees involved or methods used to identify studies for inclusion. The remaining 42% of respondents did not plan the meta-analysis before the results of any of the studies were known and this did not classify as PMA. Despite this, the articles used terms such as PMA, prospectively planned meta-analysis, pre-defined or pre-specified meta-analysis to describe their studies. The identified PMA methods papers were outdated and inconsistent.

Conclusions: identification of PMA was difficult. There is a lack of information reported and when details were reported, this was done with much variation. Yet, the number of PMA appears to be increasing and the methods advice available is lacking. Therefore, a standardised way of reporting PMA is greatly needed.

Patient or healthcare consumer involvement: PMA have the potential to decrease bias, reduce research waste and increase statistical power. This means that research can help patients in areas that may not have been possible before. For example, increased statistical power for rarer outcomes means that questions can be answered that may have never before been answered with any certainty.