CaltechAUTHORS
  A Caltech Library Service

Enabling Personalized Decision Support with Patient-Generated Data and Attributable Components

Mitchell, Elliot G. and Tabak, Esteban G. and Levine, Matthew E. and Mamykina, Lena and Albers, David J. (2021) Enabling Personalized Decision Support with Patient-Generated Data and Attributable Components. Journal of Biomedical Informatics, 113 . Art. No. 103639. ISSN 1532-0464. https://resolver.caltech.edu/CaltechAUTHORS:20201224-085808256

[img] PDF - Accepted Version
See Usage Policy.

1462Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20201224-085808256

Abstract

Decision-making related to health is complex. Machine learning (ML) and patient generated data can identify patterns and insights at the individual level, where human cognition falls short, but not all ML-generated information is of equal utility for making health-related decisions. We develop and apply attributable components analysis (ACA), a method inspired by optimal transport theory, to type 2 diabetes self-monitoring data to identify patterns of association between nutrition and blood glucose control. In comparison with linear regression, we found that ACA offers a number of characteristics that make it promising for use in decision support applications. For example, ACA was able to identify non-linear relationships, was more robust to outliers, and offered broader and more expressive uncertainty estimates. In addition, our results highlight a tradeoff between model accuracy and interpretability, and we discuss implications for ML-driven decision support systems.


Item Type:Article
Related URLs:
URLURL TypeDescription
https://doi.org/10.1016/j.jbi.2020.103639DOIArticle
ORCID:
AuthorORCID
Levine, Matthew E.0000-0002-5627-3169
Mamykina, Lena0000-0001-5203-274X
Albers, David J.0000-0002-5369-526X
Additional Information:© 2020 Elsevier. Received 22 November 2019, Revised 3 August 2020, Accepted 26 November 2020, Available online 13 December 2020. Data Sharing Statement: The data used in this project is considered protected health information (PHI) and therefore cannot be made openly available for general use. The authors are open to collaborating with researchers, to build on or reproduce the methods and analyses described in this work. Please email the corresponding author if you are interested in collaborating or reproducing this work (Elliot G Mitchell: egm2143@columbia.edu). CRediT authorship contribution statement: Elliot G Mitchell: Conceptualization, Formal analysis, Investigation, Data curation, Writing - original draft. Esteban G Tabak: Methodology, Writing - original draft, Writing - review & editing. Matthew E Levine: Data curation, Writing - review & editing. Lena Mamykina: Writing - review & editing. David J Albers: Supervision, Methodology, Writing - review & editing. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Subject Keywords:Type 2 diabetes; Patient-generated health data; Machine learning; Patient decision support
Record Number:CaltechAUTHORS:20201224-085808256
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20201224-085808256
Official Citation:Elliot G. Mitchell, Esteban G. Tabak, Matthew E Levine, Lena Mamykina, David J. Albers, Enabling personalized decision support with patient-generated data and attributable components, Journal of Biomedical Informatics, Volume 113, 2021, 103639, ISSN 1532-0464, https://doi.org/10.1016/j.jbi.2020.103639. (http://www.sciencedirect.com/science/article/pii/S1532046420302665)
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:107282
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:04 Jan 2021 15:41
Last Modified:05 Jan 2021 00:15

Repository Staff Only: item control page