Cumulative Network Meta-Analysis and Clinical Practice Guidelines: A Case Study on First-Line Medical Therapies for Primary Open-Angle Glaucoma
Embargo until
Date
2015-05-19
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Background: Clinical practice guidelines are statements of recommendations for patient care. Studies have shown that guideline recommendations do not always depend on evidence from clinical trials or systematic reviews. It is unknown whether no high quality evidence exists, evidence exists but authors were unaware of it, or advanced statistical methods were not available to them to address their questions. Our objective was to compare the guideline recommendations for first-line medical therapy for primary open-angle glaucoma (POAG) from each major update of the American Academy of Ophthalmology’s (AAO) Preferred Practice Patterns (PPPs) with the actual evidence base available at the time.
Methods: We identified and extracted recommendations relevant to first-line medical therapy for POAG from each version of the AAO PPP. We searched MEDLINE, EMBASE, and CENTRAL for randomized controlled trials published up to March 2014. We analyzed intraocular pressure (IOP) outcome data as available at the time of each major guideline update. We used network meta-analysis to determine which of all drugs “works best.”
Results: We identified 9 versions of AAO’s guideline for POAG published between 1989 and 2010. Based on similarity in treatment recommendations or discussion, we grouped these guidelines into 5 sets: 1989-1992, 1996, 2000-2003; 2005-2006, and 2010. The 2010 guideline recommended prostaglandins as initial treatment, but previous sets presented treatment options without recommending one drug (or class) over another. Based on a series of network meta-analyses of trials published up to around the time of the latest guideline in each set, all drugs are more effective than placebo or no treatment at each time point, but effect size appears to decrease over time. Network meta-analysis indicated that the most effective drug and class (at time point analyzed) were: levobunolol and beta blockers (1991), levobunolol and alpha agonists (1995), travoprost and prostaglandins (2002), bimatoprost and prostaglandins (2004 and 2009).
Conclusions: Network meta-analysis improves our understanding of the comparative effectiveness of multiple interventions. Had network meta-analysis been available, the AAO POAG PPP could have recommended prostaglandins (current first-line treatment) seven years before it actually did. Guideline developers should consider using results from network meta-analyses in forming future recommendations.
Description
Keywords
Network meta-analysis, Systematic review, Evidence-based medicine, Clinical practice guidelines, Primary open-angle glaucoma