Global Deep Forecasting with Patient-Specific Pharmacokinetics

Willa Potosnak, Cristian Ignacio Challu, Kin G. Olivares, Keith A Dufendach, Artur Dubrawski
Proceedings of the sixth Conference on Health, Inference, and Learning, PMLR 287:306-336, 2025.

Abstract

Forecasting healthcare time series data is vital for early detection of adverse outcomes and patient monitoring. However, it can be challenging in practice due to variable medication administration and unique pharmacokinetic (PK) properties of each patient. To address these challenges, we propose a novel hybrid global-local architecture and a PK encoder that informs deep learning models of patient-specific treatment effects. We showcase the efficacy of our approach in achieving significant accuracy gains in a blood glucose forecasting task using both realistically simulated and real-world data. Our PK encoder surpasses baselines by up to 16.4% on simulated data and 5.3% on real-world data for individual patients during critical events of severely high and low glucose levels. Furthermore, our proposed hybrid global-local architecture outperforms patient-specific PK models by 15.8%, on average.

Cite this Paper

BibTeX
@InProceedings{pmlr-v287-potosnak25a, title = {Global Deep Forecasting with Patient-Specific Pharmacokinetics}, author = {Potosnak, Willa and Challu, Cristian Ignacio and Olivares, Kin G. and Dufendach, Keith A and Dubrawski, Artur}, booktitle = {Proceedings of the sixth Conference on Health, Inference, and Learning}, pages = {306--336}, year = {2025}, editor = {Xu, Xuhai Orson and Choi, Edward and Singhal, Pankhuri and Gerych, Walter and Tang, Shengpu and Agrawal, Monica and Subbaswamy, Adarsh and Sizikova, Elena and Dunn, Jessilyn and Daneshjou, Roxana and Sarker, Tasmie and McDermott, Matthew and Chen, Irene}, volume = {287}, series = {Proceedings of Machine Learning Research}, month = {25--27 Jun}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v287/main/assets/potosnak25a/potosnak25a.pdf}, url = {https://proceedings.mlr.press/v287/potosnak25a.html}, abstract = {Forecasting healthcare time series data is vital for early detection of adverse outcomes and patient monitoring. However, it can be challenging in practice due to variable medication administration and unique pharmacokinetic (PK) properties of each patient. To address these challenges, we propose a novel hybrid global-local architecture and a PK encoder that informs deep learning models of patient-specific treatment effects. We showcase the efficacy of our approach in achieving significant accuracy gains in a blood glucose forecasting task using both realistically simulated and real-world data. Our PK encoder surpasses baselines by up to 16.4% on simulated data and 5.3% on real-world data for individual patients during critical events of severely high and low glucose levels. Furthermore, our proposed hybrid global-local architecture outperforms patient-specific PK models by 15.8%, on average.} }
Endnote
%0 Conference Paper %T Global Deep Forecasting with Patient-Specific Pharmacokinetics %A Willa Potosnak %A Cristian Ignacio Challu %A Kin G. Olivares %A Keith A Dufendach %A Artur Dubrawski %B Proceedings of the sixth Conference on Health, Inference, and Learning %C Proceedings of Machine Learning Research %D 2025 %E Xuhai Orson Xu %E Edward Choi %E Pankhuri Singhal %E Walter Gerych %E Shengpu Tang %E Monica Agrawal %E Adarsh Subbaswamy %E Elena Sizikova %E Jessilyn Dunn %E Roxana Daneshjou %E Tasmie Sarker %E Matthew McDermott %E Irene Chen %F pmlr-v287-potosnak25a %I PMLR %P 306--336 %U https://proceedings.mlr.press/v287/potosnak25a.html %V 287 %X Forecasting healthcare time series data is vital for early detection of adverse outcomes and patient monitoring. However, it can be challenging in practice due to variable medication administration and unique pharmacokinetic (PK) properties of each patient. To address these challenges, we propose a novel hybrid global-local architecture and a PK encoder that informs deep learning models of patient-specific treatment effects. We showcase the efficacy of our approach in achieving significant accuracy gains in a blood glucose forecasting task using both realistically simulated and real-world data. Our PK encoder surpasses baselines by up to 16.4% on simulated data and 5.3% on real-world data for individual patients during critical events of severely high and low glucose levels. Furthermore, our proposed hybrid global-local architecture outperforms patient-specific PK models by 15.8%, on average.
APA
Potosnak, W., Challu, C.I., Olivares, K.G., Dufendach, K.A. & Dubrawski, A.. (2025). Global Deep Forecasting with Patient-Specific Pharmacokinetics. Proceedings of the sixth Conference on Health, Inference, and Learning, in Proceedings of Machine Learning Research 287:306-336 Available from https://proceedings.mlr.press/v287/potosnak25a.html.

Related Material