Abstract
Purpose :
Clustering of visual outcomes for neovascular age-related macular degeneration (nAMD) within practitioners is likely to occur due to a combination of shared patient demographics and adherence, practitioner treatment strategy including choice of treatment and re-treatment intervals, and access to resources. Such clustering would subsequently result in variation in outcomes between practitioners. We report both the intraclass correlation and variation in 12-month outcomes within and between practitioners in a large cohort of patients from routine clinical practice.
Methods :
Treatment-naïve patients in the Fight Retinal Blindness! registry initiating treatment with anti-vascular endothelial growth factor (anti-VEGF) injections from 2012 onwards with at least 12 months of follow-up were identified. Only practitioners with at least 10 eligible eyes were included. Data from Australia, France, Italy, Netherlands, New Zealand, and Spain were included.
The primary outcome was the practitioner intraclass correlation (ICC) for the change in visual acuity (VA) at 12 months. Other outcomes included the standard deviation of the 12-month VA change within and between practitioners. Mixed-effects modelling with the practitioner specified as a nested random-effect and baseline age and vision as fixed effects was used to estimate these outcomes.
Results :
There were 7191 eyes from 6023 patients treated by 94 practitioners that met the inclusion criteria. The mean 12-month VA change and 95% confidence interval for each practitioner is shown in Figure 1. The ICC for the 12-month VA change was 0.016, which is negligible, and the between practitioner standard deviation was 2.25. In contrast, the within practitioner standard deviation was much higher at 14.8. This result was consistent across countries. The between practitioner standard deviation ranged from 0.73 in France to 4.14 in New Zealand.
Conclusions :
We found a negligible practitioner ICC in VA change due to the substantial variation in outcomes within practitioner relative to the variation between practitioners. This indicates that the patient-level response is contributing far more to the variation in outcomes than the practitioner. It is possible that most practitioners these days are treating nAMD patients fairly similarly, at least in the short-term.
This is a 2021 ARVO Annual Meeting abstract.