Abstract
Purpose :
While ophthalmic screening is effective at mitigating diabetic retinopathy (DR)-related vision loss, current guidelines do not consider newer screening technologies like teleretinal imaging (TRI) and how they should be best applied to populations in a large urban setting which tend to be at greatest risk for DR. In order to benefit as many patients as possible, optimal screening type and frequency as well as cost-effectiveness must be assessed. Cost-effectiveness analysis was conducted on 4 hypothetical screening policies to determine the most cost-effective DR screening policy for an urban safety-net hospital system.
Methods :
A Monte Carlo discrete event simulation model was utilized to evaluate the cost-effectiveness of 4 screening policies. These policies were defined as annual and biennial TRI (ATRI and BTRI, respectively) where a follow-up clinical screening (CS) is done if TRI indicates sight-threatening disease, and conventional annual and biennial CS (ACS and BCS, respectively). For each policy, a hypothetical cohort of 500,000 patients were simulated where costs ($) and QALYs were collected over each patient’s lifetime. Accuracies of TRI and cost estimates for both CS and TRI were based on previously published data (Walton OB IV et al. JAMA Ophthalmol. 2016, Garoon RB et al. Opthalmology Retina 2018). Compliance with CS exams was fixed to 50% as is observed in similar urban populations. Sensitivity analysis was performed to evaluate the effects of higher TRI compliance rates.
Results :
The average costs per QALY (assuming 50% compliance rate) for ATRI, BTRI, ACS, and BCS were $375.63, $424.02, $374.91, and $405.86, respectively. Once compliance with TRI reached 55%, ATRI was most cost-effective with an average cost/QALY of $368.92 (Figs 1 and 2). Once compliance with TRI exceeded 68%, ATRI generated the most accumulated QALYs for the least cost compared to other policies.
Conclusions :
For a safety-net healthcare system in Texas, ATRI was shown to be the best practice once the TRI compliance rate reached 68%. This provides valuable insights into the implementation of a TRI-based DR screening program and optimization of resource allocation in a large, heterogeneous patient population.
This is a 2020 ARVO Annual Meeting abstract.