Abstract
Problem definition Ensuring timely patient discharges is central to managing a hospital’s patient flow; however, discharges are dependent on the coordination of multiple care teams and thus are highly decentralized in nature. Therefore, central capacity teams often lack transparency into how to prioritize scarce resources (e.g., imaging, post-hospital facility placement) for patients who need them for timely discharge. Our goal is to identify a small subset of such barriers for hospitalized surgical patients by balancing two aims: a high likelihood that the patient will be discharged from the hospital in the next 24 hours if these barriers are resolved; and a high likelihood that these barriers will indeed be resolved.
Methodology/results We leverage a machine learning model that predicts the likelihood of a patient’s discharge in the next 24 hours and train another set of machine learning models that predict, for each open barrier of a patient, the likelihood that the barrier will be resolved prior to discharge. We then combine these predictive models with a mixed-integer prescriptive optimization model to identify a subset of issues called minimal barriers. We empirically demonstrate the efficacy of the proposed formulation and solution methodology in identifying a small number of minimal barriers using real data from a large academic medical center. We also study the influence of variability in discharge prediction estimates on the prescribed barriers; we introduce a robust optimization variant which incorporates this directly, and we demonstrate that this is capable of overcoming several shortcomings of the nominal approach.
Managerial implications The modeling framework we propose can significantly enhance the ability of capacity management teams to identify significant barriers to care progression and understand which barriers are most important to address (based on their relationship with patient discharge) while taking into account the heterogeneity of barriers and their resolution.
Competing Interest Statement
The authors have declared no competing interest.
Funding Statement
This study did not receive any funding.
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
Approved by Mass General Brigham Institutional Review Board, protocol 2011P001124
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Yes
Footnotes
Contact: taghi.khaniyev{at}bilkent.edu.tr
Addition of extensive sensitivity analyses and additional model formulations; model retraining with temporal data splits; and updated model exposition
Data Availability
No data is publicly available.
