Use of network meta-analysis in guidelines: a retrospective analysis

Presentation video:




Oral session: Network meta-analysis


Wednesday 23 October 2019 - 11:00 to 12:30


All authors in correct order:

Yang N1, Kang S2, Zhou Q2, Zhang J3, Estill J4, Chen Y1, Yang K1
1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Chinese GRADE Centre, WHO Collaborating Centre for Guideline Implementation and Knowledge Translation, China
2 First Clinical Medical College, Lanzhou University, China
3 School of Public Health, Lanzhou University, China
4 Institute of Global Health, University of Geneva, Institute of Mathematical Statistics and Actuarial Science, University of Bern, Switzerland
Presenting author and contact person

Presenting author:

Xuan Yu

Contact person:

Abstract text
Background: the use of network meta-analyses (NMA) by guidelines has several advantages: NMA can analyze both direct and indirect evidence collectively, comparing all kinds of interventions simultaneously, and provide the ranking of all evaluated interventions for a specific outcome.

Objectives: to investigate how NMA have been cited in clinical practice guidelines.

Methods: we conducted a cited reference search with the cited titles including 'Network Meta-analysis' in the Web of Science database to identify articles using NMA. We screened all records identified in the search. We excluded guidelines published before 2018 or in languages other than English. Two researchers independently screened the records and extracted data; disagreements were resolved through discussion.

Results: we identified a total of 97 eligible guidelines. Most guidelines were developed in the USA (n = 22, 22.7%), UK (n = 11, 11.3%), and Canada (n = 8, 8.2%). Forty-seven (46.4%) guidelines were original versions; 66 (46.4%) were updates, and one (1.0%) guideline was an adapted version. Fifty-one (52.6%) guidelines had multiple topics, 33 guidelines (34.0%) were for treatment, and seven (7.2%) for prevention. The three most common classifications of diseases were circulatory system (n = 19, 19.6%), digestive system (n = 17, 17.5%), and musculoskeletal system and connective tissue (n = 7, 7.2%). One (1.8%) guideline cited nine NMA; one (1.8%) cited four NMA; six (6.2%) cited three NMA; 19 (19.6%) cited two NMA; and the remaining 70 (72.2%) cited one NMA. The first cited NMA was published in 2001, and the highest number were published in 2015 (n = 31, 32.0%). The duration of the reference period of the NMA ranged between 0 and 17 years and the median reference period was three years.

Conclusions: taking into account that other guidelines might not cite NMA, the guidelines published in 2018 cited very few NMA on average. Guideline developers should use more evidence from NMA to support recommendations.

Patient or healthcare consumer involvement: as target population of guidelines, patients could be involved in evidence-based decision making by reviewing evidence from NMA.