
DME Cost Analysis: Strengths and Limits of Trial-Based Economic Modeling
A model built on trial data offers standardized, controlled inputs but reflects preset practice patterns. The panel weighs the strengths and limits of this approach and the value of counting indirect costs.
Episodes in this series

This episode, "DME Cost Analysis: Strengths and Limits of Trial-Based Economic Modeling," examines how a trial-based economic model was built and where its limits lie.
Dr. Yannuzzi notes that the analysis is a theoretical model layered on trial data rather than claims or registry data. He asks Dr. Patel about the strengths and limitations of that approach. Dr. Patel first congratulates the medical students and residents who built the analysis and credits their sound simulation methodology.
He argues that starting with trial data is important for cost analyses. The data are well controlled, loss to follow-up is lower, and clear data points exist at each visit. Visual acuity is measured in a standardized way, which allows quality-of-life differences to be assessed when acuity is converted to utility. Clinic patients, by contrast, are sometimes not even refracted.
On limitations, Dr. Patel says the model is preset, so real-world practice patterns may differ. A strength that cuts across both settings, he adds, is the inclusion of indirect costs. Many analyses look only at what insurers pay physicians and practices for drugs and imaging. Injections also involve caregivers. Many young, healthy patients cannot drive themselves home, so a companion loses time from work as well, and self-employed patients lose revenue when they take a day off.
These factors show that visits, not just drug prices, are a large cost driver. Even though the 8 mg dose is more expensive, cutting injections by about 25% eliminates those visit costs and lowers overall cost. He suggests registry studies as a next step but warns they blur efficacy, since patients and physicians may treat more than needed out of routine or incentives.
Dr. Yannuzzi agrees that trial data are the gold standard and worries about garbage in, garbage out with registries. He sees Monte Carlo as more robust than the decision-tree models he has used, while noting that protocol criteria differ from clinic practice.
In "DME Cost Analysis Results: Fewer Injections Drive Lower Total Costs," Dr. Yannuzzi presents the 96-week cost results, and the panel considers why injection counts matter.
Related to this article








