Patient-Specific Effects of Medication Using Latent Force Models with Gaussian Processes

Li-Fang Cheng, Bianca Dumitrascu, Michael Zhang, Corey Chivers, Michael Draugelis, Kai Li, Barbara Engelhardt
Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, PMLR 108:4045-4055, 2020.

Abstract

A multi-output Gaussian process (GP) is a flexible Bayesian nonparametric framework that has proven useful in jointly modeling the physiological states of patients in medical time series data. However, capturing the short-term effects of drugs and therapeutic interventions on patient physiological state remains challenging. We propose a novel approach that models the effect of interventions as a hybrid Gaussian process composed of a GP capturing patient baseline physiology convolved with a latent force model capturing effects of treatments on specific physiological features. The combination of a multi-output GP with a time-marked kernel GP leads to a well-characterized model of patients’ physiological state across a hospital stay, including response to interventions. Our model leads to analytically tractable cross-covariance functions that allow for scalable inference. Our hierarchical model includes estimates of patient-specific effects but allows sharing of support across patients. Our approach achieves competitive predictive performance on challenging hospital data, where we recover patient-specific response to the administration of three common drugs: one antihypertensive drug and two anticoagulants.

Cite this Paper


BibTeX
@InProceedings{pmlr-v108-cheng20c, title = {Patient-Specific Effects of Medication Using Latent Force Models with Gaussian Processes}, author = {Cheng, Li-Fang and Dumitrascu, Bianca and Zhang, Michael and Chivers, Corey and Draugelis, Michael and Li, Kai and Engelhardt, Barbara}, booktitle = {Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics}, pages = {4045--4055}, year = {2020}, editor = {Silvia Chiappa and Roberto Calandra}, volume = {108}, series = {Proceedings of Machine Learning Research}, month = {26--28 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v108/cheng20c/cheng20c.pdf}, url = { http://proceedings.mlr.press/v108/cheng20c.html }, abstract = {A multi-output Gaussian process (GP) is a flexible Bayesian nonparametric framework that has proven useful in jointly modeling the physiological states of patients in medical time series data. However, capturing the short-term effects of drugs and therapeutic interventions on patient physiological state remains challenging. We propose a novel approach that models the effect of interventions as a hybrid Gaussian process composed of a GP capturing patient baseline physiology convolved with a latent force model capturing effects of treatments on specific physiological features. The combination of a multi-output GP with a time-marked kernel GP leads to a well-characterized model of patients’ physiological state across a hospital stay, including response to interventions. Our model leads to analytically tractable cross-covariance functions that allow for scalable inference. Our hierarchical model includes estimates of patient-specific effects but allows sharing of support across patients. Our approach achieves competitive predictive performance on challenging hospital data, where we recover patient-specific response to the administration of three common drugs: one antihypertensive drug and two anticoagulants.} }
Endnote
%0 Conference Paper %T Patient-Specific Effects of Medication Using Latent Force Models with Gaussian Processes %A Li-Fang Cheng %A Bianca Dumitrascu %A Michael Zhang %A Corey Chivers %A Michael Draugelis %A Kai Li %A Barbara Engelhardt %B Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2020 %E Silvia Chiappa %E Roberto Calandra %F pmlr-v108-cheng20c %I PMLR %P 4045--4055 %U http://proceedings.mlr.press/v108/cheng20c.html %V 108 %X A multi-output Gaussian process (GP) is a flexible Bayesian nonparametric framework that has proven useful in jointly modeling the physiological states of patients in medical time series data. However, capturing the short-term effects of drugs and therapeutic interventions on patient physiological state remains challenging. We propose a novel approach that models the effect of interventions as a hybrid Gaussian process composed of a GP capturing patient baseline physiology convolved with a latent force model capturing effects of treatments on specific physiological features. The combination of a multi-output GP with a time-marked kernel GP leads to a well-characterized model of patients’ physiological state across a hospital stay, including response to interventions. Our model leads to analytically tractable cross-covariance functions that allow for scalable inference. Our hierarchical model includes estimates of patient-specific effects but allows sharing of support across patients. Our approach achieves competitive predictive performance on challenging hospital data, where we recover patient-specific response to the administration of three common drugs: one antihypertensive drug and two anticoagulants.
APA
Cheng, L., Dumitrascu, B., Zhang, M., Chivers, C., Draugelis, M., Li, K. & Engelhardt, B.. (2020). Patient-Specific Effects of Medication Using Latent Force Models with Gaussian Processes. Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 108:4045-4055 Available from http://proceedings.mlr.press/v108/cheng20c.html .

Related Material