Surgical scheduling via optimization and machine learning with long-tailed data
Using data from cardiovascular surgery patients with long and highly variable post-surgical lengths of stay (LOS), we develop a modeling framework to reduce recovery unit congestion. We estimate the LOS and its probability distribution using machine learning models, schedule procedures on a rolling...
| Main Authors: | , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Springer US
2023
|
| Online Access: | https://hdl.handle.net/1721.1/153163 |
| _version_ | 1811079426031484928 |
|---|---|
| author | Shi, Yuan Mahdian, Saied Blanchet, Jose Glynn, Peter Shin, Andrew Y. Scheinker, David |
| author2 | Massachusetts Institute of Technology. Operations Research Center |
| author_facet | Massachusetts Institute of Technology. Operations Research Center Shi, Yuan Mahdian, Saied Blanchet, Jose Glynn, Peter Shin, Andrew Y. Scheinker, David |
| author_sort | Shi, Yuan |
| collection | MIT |
| description | Using data from cardiovascular surgery patients with long and highly variable post-surgical lengths of stay (LOS), we develop a modeling framework to reduce recovery unit congestion. We estimate the LOS and its probability distribution using machine learning models, schedule procedures on a rolling basis using a variety of optimization models, and estimate performance with simulation. The machine learning models achieved only modest LOS prediction accuracy, despite access to a very rich set of patient characteristics. Compared to the current paper-based system used in the hospital, most optimization models failed to reduce congestion without increasing wait times for surgery. A conservative stochastic optimization with sufficient sampling to capture the long tail of the LOS distribution outperformed the current manual process and other stochastic and robust optimization approaches. These results highlight the perils of using oversimplified distributional models of LOS for scheduling procedures and the importance of using optimization methods well-suited to dealing with long-tailed behavior. |
| first_indexed | 2024-09-23T11:14:49Z |
| format | Article |
| id | mit-1721.1/153163 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T11:14:49Z |
| publishDate | 2023 |
| publisher | Springer US |
| record_format | dspace |
| spelling | mit-1721.1/1531632024-09-11T05:14:47Z Surgical scheduling via optimization and machine learning with long-tailed data Shi, Yuan Mahdian, Saied Blanchet, Jose Glynn, Peter Shin, Andrew Y. Scheinker, David Massachusetts Institute of Technology. Operations Research Center Using data from cardiovascular surgery patients with long and highly variable post-surgical lengths of stay (LOS), we develop a modeling framework to reduce recovery unit congestion. We estimate the LOS and its probability distribution using machine learning models, schedule procedures on a rolling basis using a variety of optimization models, and estimate performance with simulation. The machine learning models achieved only modest LOS prediction accuracy, despite access to a very rich set of patient characteristics. Compared to the current paper-based system used in the hospital, most optimization models failed to reduce congestion without increasing wait times for surgery. A conservative stochastic optimization with sufficient sampling to capture the long tail of the LOS distribution outperformed the current manual process and other stochastic and robust optimization approaches. These results highlight the perils of using oversimplified distributional models of LOS for scheduling procedures and the importance of using optimization methods well-suited to dealing with long-tailed behavior. 2023-12-14T16:32:09Z 2023-12-14T16:32:09Z 2023-09-04 2023-12-09T04:17:46Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/153163 Shi, Yuan, Mahdian, Saied, Blanchet, Jose, Glynn, Peter, Shin, Andrew Y. et al. 2023. "Surgical scheduling via optimization and machine learning with long-tailed data." en https://doi.org/10.1007/s10729-023-09649-0 Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature application/pdf Springer US Springer US |
| spellingShingle | Shi, Yuan Mahdian, Saied Blanchet, Jose Glynn, Peter Shin, Andrew Y. Scheinker, David Surgical scheduling via optimization and machine learning with long-tailed data |
| title | Surgical scheduling via optimization and machine learning with long-tailed data |
| title_full | Surgical scheduling via optimization and machine learning with long-tailed data |
| title_fullStr | Surgical scheduling via optimization and machine learning with long-tailed data |
| title_full_unstemmed | Surgical scheduling via optimization and machine learning with long-tailed data |
| title_short | Surgical scheduling via optimization and machine learning with long-tailed data |
| title_sort | surgical scheduling via optimization and machine learning with long tailed data |
| url | https://hdl.handle.net/1721.1/153163 |
| work_keys_str_mv | AT shiyuan surgicalschedulingviaoptimizationandmachinelearningwithlongtaileddata AT mahdiansaied surgicalschedulingviaoptimizationandmachinelearningwithlongtaileddata AT blanchetjose surgicalschedulingviaoptimizationandmachinelearningwithlongtaileddata AT glynnpeter surgicalschedulingviaoptimizationandmachinelearningwithlongtaileddata AT shinandrewy surgicalschedulingviaoptimizationandmachinelearningwithlongtaileddata AT scheinkerdavid surgicalschedulingviaoptimizationandmachinelearningwithlongtaileddata |