Can Patient-Reported Outcomes Serve as Endpoints in Trials?
By Lynda Seminara
Selected By: Stephen D. McLeod, MD
Journal Highlights
Ophthalmology, May 2019
Download PDF
On behalf of the United Kingdom Glaucoma Treatment Study (UKGTS) investigators, Jones et al. gathered and compared self-reported outcomes for UKGTS participants. In the flagship trial, patients with open-angle glaucoma (OAG) had been assigned to receive latanoprost or placebo drops, and visual field progression was the outcome of interest. Eligible for the subsequent study were patients from UKGTS with newly diagnosed OAG and self-reported outcome measures at both baseline and study completion. Because the average changes in patient-reported outcome measures (PROMs) for health- and vision-related quality of life were found to be similar for the placebo and active-treatment groups, the researchers surmised that PROMs may not be sensitive enough to function as primary endpoints in clinical trials of early-stage glaucoma.
The PROM study included 182 patients who received latanoprost and 168 placebo recipients. At baseline and trial exit, participants completed general health PROMs (European Quality of Life in 5 Dimensions [EQ-5D] and 36-item Short Form [SF-36]) as well as glaucoma-specific PROMs (15-item Glaucoma Quality of Life [GQL-15] and 9-item Glaucoma Activity Limitation [GAL-9]). The percentage change between PROM values was calculated for each patient and compared between treatment arms. Also compared were differences between patients whose glaucoma remained stable and those who experienced progression (determined by visual field changes).
The average percentage change in PROMs was similar for the placebo and latanoprost groups, with no significant between-group difference for any measure (EQ-5D overall, p = .98; EQ-5D visual analog scale, p = .88; SF-36, p = .94; GQL-15, p = .66; GAL-9, p = .87). As expected, there were significant differences in glaucoma-related PROMs between patients with and without progressing glaucoma (GQL-15, p = .02; GAL-9, p = .02), and the differences in general health PROMs were similar for these subgroups (EQ-5D, p = .62; EQ-5D visual analog scale, p = 0.23; SF-36, p = .65).
The low sensitivity of PROMs may render these tools inadequate as primary endpoints in trials of early-stage glaucoma. Although sensitivity to clinical meaningfulness is a common criterion for outcomes selection, quality of life is important as well. Even if PROMs cannot capture the disease-modifying effects of treatment, these tools may help to assess other consequences of therapy, such as side effects and dosing convenience. (Also see related commentary by Scott Wallace, MD, and Jane Edmond, MD, in the same issue.)
The original article can be found here.