Abstract: TH-PO112
An Artificial Intelligence-Based Program Significantly Reduces ESA (Erythropoiesis-Stimulating Agent) Use and Maintains Hemoglobin
Session Information
- Pharmacology
October 24, 2024 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Pharmacology (PharmacoKinetics, -Dynamics, -Genomics)
- 2000 Pharmacology (PharmacoKinetics, -Dynamics, -Genomics)
Authors
- Brier, Michael E., University of Louisville, Louisville, Kentucky, United States
- Gaweda, Adam E., University of Louisville, Louisville, Kentucky, United States
Background
ESAs have experienced dramatic dose fluctuations prompted by reimbursement challenges, shifting targets, and poorly designed anemia management protocols (AMP). The package insert recommends using the lowest dose to avoid transfusion. Our hypothesis is that SAA is superior to the commonly used expert system approach (AMP).
Methods
SAA is an AI-powered planning and decision support system based on two components: 1) pattern matching to identify individual dose-response profile, and 2) Model Predictive Control to perform look-ahead treatment planning. PEG epoetin beta data were obtained from the USRDS 2023 annual report and from the Dosis database and the University of Louisville dialysis facilities. Collection was approved by the University of Louisville institutional review board. PEG dose in the USRDS was reported as the mean monthly dose for the year. Individual monthly doses were obtained for the treatment groups.
Results
Table shows yearly means for all PEG treatments. 95% CI show that SAA resulted in a dose reduction of between 20 and 41%. A separate analysis of PEG use in the University of Louisville dialysis facilities shows that use of SAA results in immediate changes in dose decreasing when SAA is started and increasing when withdrawn Fac3 never used SAA and had poorest performance. Fac1 and Fac2 used SAA continuously and had best performance Fac4 and Fac5 stopped use and immediately increased consumption by 50 and 43%. Simultaneously Hb fell 6 to 11% following -SAA.
Conclusion
SAA is superior to physician-based AMP. The results of the RCT are confirmed in this analysis. There is a definitive cost benefit without sacrificing quality.
| 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | |
| USRDS PEG | 149 | 151 | 151 | 149 | 143 | 139 | 137 |
| SAA PEG | 82-88(4953) | 106-110 (n=12081) | 86-96(n=1911) | 95-111(n=847) |