Outcome reporting in clinical trials and systematic reviews on medical interventions for primary open-angle glaucoma

Tags: Poster
Yu T1, Li T1, Dickersin K1
1Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

Background: Health decision-makers want to know ‘what works best?’ not just how well a specific intervention works. With this in mind, the Comparing Multiple Interventions Methods Group defines an overview as a review that synthesizes evidence from existing systematic reviews (SRs) to address a well-defined clinical question. Meta-analysis of multiple interventions using network meta-analysis (NMA) can be used to facilitate simultaneous inference on ‘what works best’ as long as the conditions treated and outcomes examined are similar. Suboptimal reporting on patient-centered outcomes in SRs and the underlying randomized clinical trials (RCTs) may hinder the translation of research evidence. Our objective was to examine the outcomes reported in SRs and RCTs of medical interventions for primary open-angle glaucoma (POAG), as a preliminary step for NMA.

Methods: As part of a larger project, we searched de novo for RCTs on any medical intervention for POAG. We also searched for SRs on the same topic to examine the usefulness and validity of relying on existing SRs for overviews and NMA. We examined reporting of 6 outcomes in the 497 RCTs and 23 SRs identified: intraocular pressure (IOP), optic nerve progression, visual field, visual acuity, development of glaucoma, and safety. We considered the last four outcomes to be patient-centered.

Results: IOP was a reported outcome in 95% of trials (see Table 1), while a smaller proportion of trials reported patient-centered outcomes: 16% reported visual field, 8% reported visual acuity, 2% reported development of glaucoma and 71% reported safety. Most SRs (83%) reported IOP, but only 9% reported visual field, 4% reported development of glaucoma, none reported visual acuity, and 70% reported safety.

Conclusion: Our data show that reporting on patient-centered outcomes is incomplete for those planning an overview or NMA, and may not provide enough evidence to inform patient-centered care.