A Bayesian Nonparametric Approach for Estimating Individualized Treatment-Response Curves

Yanbo Xu, Yanxun Xu, Suchi Saria
Proceedings of the 1st Machine Learning for Healthcare Conference, PMLR 56:282-300, 2016.

Abstract

We study the problem of estimating the continuous response over time of actions from observational time series—a retrospective dataset where the policy by which the data are generated are unknown to the learner. We develop a novel method based on Bayesian nonparametrics (BNP) that can flexibly model functional data and provide posterior inference over the treatment response curves both at the individual and population level. On a challenging dataset containing time series from patients admitted to a hospital, we estimate treatment responses for 8 different treatments used in managing blood pressure and kidney function and show that the resulting fits are more accurate than alternative approaches. Accurate methods for obtaining ITRs from observational data can dramatically accelerate the pace at which personalized treatment plans become possible.

Cite this Paper

BibTeX
@InProceedings{pmlr-v56-Xu16, title = {A Non-parametric Bayesian Approach for Estimating Treatment-Response Curves from Sparse Time Series}, author = {Xu, Yanbo and Xu, Yanxun and Saria, Suchi}, booktitle = {Proceedings of the 1st Machine Learning for Healthcare Conference}, pages = {282--300}, year = {2016}, editor = {Doshi-Velez, Finale and Fackler, Jim and Kale, David and Wallace, Byron and Wiens, Jenna}, volume = {56}, series = {Proceedings of Machine Learning Research}, address = {Northeastern University, Boston, MA, USA}, month = {18--19 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v56/Xu16.pdf}, url = {https://proceedings.mlr.press/v56/Xu16.html}, abstract = {We study the problem of estimating the continuous response over time of actions from observational time series—a retrospective dataset where the policy by which the data are generated are unknown to the learner. We develop a novel method based on Bayesian nonparametrics (BNP) that can flexibly model functional data and provide posterior inference over the treatment response curves both at the individual and population level. On a challenging dataset containing time series from patients admitted to a hospital, we estimate treatment responses for 8 different treatments used in managing blood pressure and kidney function and show that the resulting fits are more accurate than alternative approaches. Accurate methods for obtaining ITRs from observational data can dramatically accelerate the pace at which personalized treatment plans become possible.} }
Endnote
%0 Conference Paper %T A Bayesian Nonparametric Approach for Estimating Individualized Treatment-Response Curves %A Yanbo Xu %A Yanxun Xu %A Suchi Saria %B Proceedings of the 1st Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2016 %E Finale Doshi-Velez %E Jim Fackler %E David Kale %E Byron Wallace %E Jenna Wiens %F pmlr-v56-Xu16 %I PMLR %P 282--300 %U https://proceedings.mlr.press/v56/Xu16.html %V 56 %X We study the problem of estimating the continuous response over time of actions from observational time series—a retrospective dataset where the policy by which the data are generated are unknown to the learner. We develop a novel method based on Bayesian nonparametrics (BNP) that can flexibly model functional data and provide posterior inference over the treatment response curves both at the individual and population level. On a challenging dataset containing time series from patients admitted to a hospital, we estimate treatment responses for 8 different treatments used in managing blood pressure and kidney function and show that the resulting fits are more accurate than alternative approaches. Accurate methods for obtaining ITRs from observational data can dramatically accelerate the pace at which personalized treatment plans become possible.
RIS
TY - CPAPER TI - A Bayesian Nonparametric Approach for Estimating Individualized Treatment-Response Curves AU - Yanbo Xu AU - Yanxun Xu AU - Suchi Saria BT - Proceedings of the 1st Machine Learning for Healthcare Conference DA - 2016/12/10 ED - Finale Doshi-Velez ED - Jim Fackler ED - David Kale ED - Byron Wallace ED - Jenna Wiens ID - pmlr-v56-Xu16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 56 SP - 282 EP - 300 L1 - http://proceedings.mlr.press/v56/Xu16.pdf UR - https://proceedings.mlr.press/v56/Xu16.html AB - We study the problem of estimating the continuous response over time of actions from observational time series—a retrospective dataset where the policy by which the data are generated are unknown to the learner. We develop a novel method based on Bayesian nonparametrics (BNP) that can flexibly model functional data and provide posterior inference over the treatment response curves both at the individual and population level. On a challenging dataset containing time series from patients admitted to a hospital, we estimate treatment responses for 8 different treatments used in managing blood pressure and kidney function and show that the resulting fits are more accurate than alternative approaches. Accurate methods for obtaining ITRs from observational data can dramatically accelerate the pace at which personalized treatment plans become possible. ER -
APA
Xu, Y., Xu, Y. & Saria, S.. (2016). A Bayesian Nonparametric Approach for Estimating Individualized Treatment-Response Curves. Proceedings of the 1st Machine Learning for Healthcare Conference, in Proceedings of Machine Learning Research 56:282-300 Available from https://proceedings.mlr.press/v56/Xu16.html.

Related Material