Use of network meta-analysis in clinical guidelines.

Abstract

The World Health Organization (WHO) has conducted two major updates to global guidance for the management of human immunodeficiency virus (HIV) and hepatitis C virus (HCV) infections in 2016. (1,2) Recommendations made by WHO can have an important effect on policy and practice, particularly for low- and middle-income countries. For instance, a 2014 survey found that over three-quarters of all 158 recommendations for HIV and tuberculosis were incorporated into national guidelines. (3) Clinical guidelines are developed through multi-step processes that ensure that guidelines are feasible within the current clinical environment and that they are based on the best available evidence. For the latest HIV and HCV guidelines, WHO used network meta-analysis to inform treatment recommendations. An expansion of conventional pairwise meta-analysis, network meta-analysis includes multiple interventions within a single analysis and estimates the relative treatment effect between each two treatments compared using direct or indirect evidence. Although it is often acknowledged that having the most up-to-date evidence is critical to the development of clinical guidelines, it is equally important that the optimal analytical methods are used to appraise the evidence. We explain here why network meta-analysis lends itself to the development of clinical guidelines and why it may be used more often in this context. WHO has a history of improving the methods it uses for developing its guidelines. Following calls for greater transparency in the use of evidence for decision-making, (4) WHO adopted a process for guideline development in 2007 that included the use of the GRADE approach (grading of recommendations, assessment, development and evaluation) to support decision-making. In addition to clinical evidence, WHO guideline groups take account of factors such as patient preferences, feasibility, costs and human rights. (5) To support decision-making beyond clinical effectiveness, mathematical models have been used to appraise the cost-effectiveness and epidemiological impact of HIV testing and treatment strategies in 2013. (6) WHO guidelines are regularly revised, generally every 3-5 years, to provide up-to-date guidance on interventions to contain or reduce major public health threats. With an ever-growing number of treatment options for HIV and HCV, methods used to synthesize the evidence within the guideline development procedures should maximize the use of all evidence and generate results that can be translated into recommendations. Although the use of network meta-analysis for WHO guideline development is recent, the method is well established within national health technology assessment agencies. It has become essential to formulating recommendations on reimbursements by health care agencies around the world. The method has also recently been adopted by Cochrane, and we found that since 2015, 10% (23/230) of systematic reviews published by this organization used network meta-analysis. The method has numerous qualities that lend themselves to decision-making processes such as clinical guidelines development. In 2015, GRADE working groups published guidance on how to use GRADE in conjunction with network meta-analysis. (7) Furthermore, the National Institute for Health and Clinical Excellence (NICE) in the United Kingdom of Great Britain and Northern Ireland has included recommendations on network meta-analysis within its clinical guidelines manual. (8) However, in 2012 only 8% of 138 NICE clinical guidelines had used network meta-analysis, (9) but the percentage is likely to be higher today. Systematic evidence appraisal, but not the network meta-analysis approach, was used to formulate recommendations in other recent, major guidelines for HIV and HCV disease. (10-12) Network meta-analysis offers several advantages to the process of developing guidelines. One advantage is the ability to make a quantitative comparison of interventions that have not been directly compared in studies. …