DoseMate: A Real-world Evaluation of Machine Learning Classification of Pill Taking Using Wrist-worn Motion Sensors

Antoine Nzeyimana, Anthony Campbell, James M Scanlan, Joanne D Stekler, Jenna Marquard, Barry G Saver, Jeremy Gummeson
Proceedings of the fifth Conference on Health, Inference, and Learning, PMLR 248:566-581, 2024.

Abstract

Non-adherence to medication is a complex behavioral issue that costs hundreds of billions of dollars annually in the United States alone. Existing solutions to improve medication adherence are limited in their effectiveness and require significant user involvement. To address this, a minimally invasive mobile health system called \myname{} is proposed, which can provide quantifiable adherence data and imposes minimal user burden. To classify a motion time-series that defines pill-taking, we adopt transfer-learning and data augmentation based techniques that uses captured pill-taking gestures along with other open datasets that represent negative labels of other wrist motions. The paper also provides a design methodology \updated{that generalizes to other} systems and describes a first-of-its-kind, in-the-wild, unobtrusively obtained dataset that contains unrestricted pill-related motion data from a diverse set of users.

Cite this Paper

BibTeX
@InProceedings{pmlr-v248-nzeyimana24a, title = {DoseMate: A Real-world Evaluation of Machine Learning Classification of Pill Taking Using Wrist-worn Motion Sensors}, author = {Nzeyimana, Antoine and Campbell, Anthony and Scanlan, James M and Stekler, Joanne D and Marquard, Jenna and Saver, Barry G and Gummeson, Jeremy}, booktitle = {Proceedings of the fifth Conference on Health, Inference, and Learning}, pages = {566--581}, year = {2024}, editor = {Pollard, Tom and Choi, Edward and Singhal, Pankhuri and Hughes, Michael and Sizikova, Elena and Mortazavi, Bobak and Chen, Irene and Wang, Fei and Sarker, Tasmie and McDermott, Matthew and Ghassemi, Marzyeh}, volume = {248}, series = {Proceedings of Machine Learning Research}, month = {27--28 Jun}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v248/main/assets/nzeyimana24a/nzeyimana24a.pdf}, url = {https://proceedings.mlr.press/v248/nzeyimana24a.html}, abstract = {Non-adherence to medication is a complex behavioral issue that costs hundreds of billions of dollars annually in the United States alone. Existing solutions to improve medication adherence are limited in their effectiveness and require significant user involvement. To address this, a minimally invasive mobile health system called \myname{} is proposed, which can provide quantifiable adherence data and imposes minimal user burden. To classify a motion time-series that defines pill-taking, we adopt transfer-learning and data augmentation based techniques that uses captured pill-taking gestures along with other open datasets that represent negative labels of other wrist motions. The paper also provides a design methodology \updated{that generalizes to other} systems and describes a first-of-its-kind, in-the-wild, unobtrusively obtained dataset that contains unrestricted pill-related motion data from a diverse set of users.} }
Endnote
%0 Conference Paper %T DoseMate: A Real-world Evaluation of Machine Learning Classification of Pill Taking Using Wrist-worn Motion Sensors %A Antoine Nzeyimana %A Anthony Campbell %A James M Scanlan %A Joanne D Stekler %A Jenna Marquard %A Barry G Saver %A Jeremy Gummeson %B Proceedings of the fifth Conference on Health, Inference, and Learning %C Proceedings of Machine Learning Research %D 2024 %E Tom Pollard %E Edward Choi %E Pankhuri Singhal %E Michael Hughes %E Elena Sizikova %E Bobak Mortazavi %E Irene Chen %E Fei Wang %E Tasmie Sarker %E Matthew McDermott %E Marzyeh Ghassemi %F pmlr-v248-nzeyimana24a %I PMLR %P 566--581 %U https://proceedings.mlr.press/v248/nzeyimana24a.html %V 248 %X Non-adherence to medication is a complex behavioral issue that costs hundreds of billions of dollars annually in the United States alone. Existing solutions to improve medication adherence are limited in their effectiveness and require significant user involvement. To address this, a minimally invasive mobile health system called \myname{} is proposed, which can provide quantifiable adherence data and imposes minimal user burden. To classify a motion time-series that defines pill-taking, we adopt transfer-learning and data augmentation based techniques that uses captured pill-taking gestures along with other open datasets that represent negative labels of other wrist motions. The paper also provides a design methodology \updated{that generalizes to other} systems and describes a first-of-its-kind, in-the-wild, unobtrusively obtained dataset that contains unrestricted pill-related motion data from a diverse set of users.
APA
Nzeyimana, A., Campbell, A., Scanlan, J.M., Stekler, J.D., Marquard, J., Saver, B.G. & Gummeson, J.. (2024). DoseMate: A Real-world Evaluation of Machine Learning Classification of Pill Taking Using Wrist-worn Motion Sensors. Proceedings of the fifth Conference on Health, Inference, and Learning, in Proceedings of Machine Learning Research 248:566-581 Available from https://proceedings.mlr.press/v248/nzeyimana24a.html.

Related Material