Predictors of problems reported on the EQ-5D-3L dimensions among people with impaired vision in northern Portugal.

Background: The EQ-5D index often fails to detect the effect of ophthalmic diseases and sight loss. Investigating predictors of individual EQ-5D health dimensions might reveal the underlying reasons. The aim of this study was to investigate predictors of health dimension ratings obtained with the EQ-5D-3L from participants with impaired vision representing a spectrum of eye diseases.

Methods: Observational cross-sectional study with participants recruited at four public hospitals in Portugal. Outpatients with visual acuity of 0.30 logMAR(6/12) or worse in the better-seeing eye were invited to participate. Participants completed two instruments: the EQ-5D-3L (measures participants' perceived health-related quality-of-life) and the Massof Activity Inventory (measures visual ability-ability to perform vision-related activities). This study used logistic regression models to identify factors associated with responses to the EQ-5D-3L.

Results: The study included 492 participants, mean age 63.4 years (range = 18-93), 50% females. The most common diagnosis was diabetic retinopathy (37%). The mean visual acuity in the better seeing eye was 0.65 logMAR (SD = 0.48) and the mean visual ability was 0.62 logits (SD = 2.04), the correlation between the two was r = - 0.511 (p < 0.001). Mobility and self-care were the health dimensions with the fewest problems (1% reported extreme problems), anxiety and depression the dimension with the most problems (24% reported extreme problems). ROC curve analysis showed that the EQ-5D index was a poor predictor of cases of vision impairment whilst visual ability given was a good predictor of cases of vision impairment. Visual ability was an independent predictor of the response for all dimensions, higher ability was always associated with a reduced odds of reporting problems. The odds of reporting problems were increased for females in 3 out of 5 dimensions. Comorbidities, visual acuity and age-category were predictors of the odds of reporting problems for one dimension each.

Conclusions: The odds of reporting problems for the five health dimensions of the EQ-5D-3L were strongly influenced by the ability to perform vision-related activities (visual ability). The EQ-5D index showed poor performance at detecting vision impairment. These findings are informative and relevant for the clinic and for research evaluating the impact of eye diseases and disease treatments in ophthalmology.