Using the results of systematic reviews, overviews of systematic reviews, and network meta-analyses to inform clinical practice guideline recommendations

Presentation video:




Oral session: Knowledge translation and communicating evidence (1)


Tuesday 22 October 2019 - 11:00 to 12:30


All authors in correct order:

Lunny C1, Ramasubbu C2, Liu T2, Salzwedel DM3, Gerrish S3, Puil L3, Mintzes B4, Wright JM3
1 Cochrane Hyptertension Review Group, Therapeutics Initiative, University of British Columbia, Canada
2 Faculty of Pharmaceutical Sciences, University of British Columbia, Canada
3 Cochrane Hypertension Review Group, University of British Columbia, Canada
4 the University of Sydney, Charles Perkins Centre, and School of Pharmacy, Australia
Presenting author and contact person

Presenting author:

Carole Lunny

Contact person:

Abstract text
Background: high quality and comprehensive systematic reviews (SRs), overviews of systematic reviews (OSRs), and network meta-analyses (NMAs) reduce research waste. By using the results of research that has already been published, SRs, OSRs, and NMAs can represent the best available evidence. Many clinical practice guideline developers, such as the World Health Organization (WHO) and the Australian National Health and Medical Research Council (NHMRC), recommend the use of these research syntheses to underpin guideline recommendations.

Objectives: our objectives were initially to determine if SRs, OSRs, or NMAs are being cited in clinical practice guidelines (CPGs). When this was the case, we determined whether SRs, OSRs, or NMAs (Cochrane and non-Cochrane) were used and cited to inform treatment recommendations within the guideline. We also evaluated if and how:
1) these review types were assessed for risk of bias, and
2) the recommendations were assessed to determine the certainty (strength) of the body of evidence.

Methods: we searched for CPGs using the TRIP and Epistemonikos databases. The retrieved citations were sorted randomly and then screened sequentially by two independent reviewers until 50 CPGs were identified that meet eligibility criteria. We included CPGs that provided at least two explicit recommendations for the management of any clinical condition. We first evaluated whether SRs, OSRs, or NMAs were used in CPGs, and second whether they were cited to inform recommendations. Full data extraction was done independently by two authors and then compared. We conducted an assessment of risk of bias of the guideline. We calculated the number and frequency of citations of SRs, OSRs, or NMAs in the CPGs, and then their recommendations. We described, tabulated, and categorised the results based on review type (SRs, OSRs, or NMAs).

Results: the 691 records retrieved were randomly sorted and screened sequentially. We screened a total of 478 records in full text to identify 50 CPGs. We excluded a total of 428 CPGs, where 47 CPGs did not have a methods section, and 16 did not include a reference section. The majority of CPGs were open access, published in a medical journal, and had guideline authors affiliated with/honoraria from pharmaceutical companies. Less than one-quarter of CPGs used a systematic review process to inform the recommendations, and less than one per cent used OSRs or NMAs. One-third of CPGs reported objectives and eligibility criteria based on PICOS elements (Population, Intervention, Comparison, Outputs, Study), a full search strategy, and a process for screening studies. Two-thirds reported that two or more databases were searched. One-third used or adapted the GRADE approach.

Conclusions: high quality research syntheses have the potential to make use of published primary and secondary research to aid in practice and policy decision making, and reduce waste. Our study found that less than one-quarter of Clinical Practice Guidelines made use of the different review types.