Fourth International Conference on Health Care Systems Engineering

Montreal, 30 May — 1 June 2019

Schedule Authors My Schedule

Home health care

May 30, 2019 03:00 PM – 05:00 PM

Location: Room: Marcel Lacoste

4 Presentations

  • 03:00 PM - 03:30 PM

    Adverse Event Prediction by Telemonitoring and Deep Learning

    • Antoine Prouvost, presenter, Polytechnique Montréal
    • Andrea Lodi, Polytechnique Montreal
    • Louis-Martin Rousseau, Polytechnique Montréal
    • Jonathan Vallée, AlayaCare

    Home health care comes as a potential solution to increasing stress on health-care systems, as well as concerns for medical patients comfort. However, additional distance from the care workers to the patients lead to more challenges, some of which can be addressed with machine learning (ML) and operations research (OR) algorithms. In this paper, we focus on automating a risk assessment of remote patients. Namely, we describe a risk prediction framework for home telemonitoring patients and show that learning a risk from weak signals in the patient's data outperforms simple risk threshold proposed by care workers to automate the task We combine recurrent neural networks with a ranking objective from survival analysis to evaluate the risk of patient's adverse events. Training and testing of our methodology is achieved on a retrospective dataset gathered by an Ontario home health care agency during the course of a multi-year pilot home telemonitoring program. Results are benchmarked against alerts that were manually engineered by registered nurses, and against a simple linear baseline.

  • 03:30 PM - 04:00 PM

    Multicriteria Scheduling Optimization in Home Health Care

    • Laura Musaraganyi, presenter,
    • Simon Germain, AlayaCare
    • Nadia Lahrichi, Polytechnique Montréal
    • Louis-Martin Rousseau, Polytechnique Montréal

    Home Health Care Services provide medical and paramedical services at the patients' home rather than in a facility (such as a hospital). This means these HHCS need to optimize the schedule of the caregivers to visit all their assigned patient in the most efficient way possible. Each agencies employ a number of schedulers whose task is to manually schedule the visits, taking into account multiple criteria and objectives.
    In this paper, we present an automatic scheduling assistant that fulfills this role to near optimality in a reasonable timeframe (seconds to minutes). The algorithm is based on a heuristic with a hierarchical approach, and submit an array of various near-equivalent solutions based on the user preferences

  • 04:00 PM - 04:30 PM

    A Two-Phase Method for Robust Home Healthcare Problem: A Case Study

    • Mahdyeh Shiri, presenter, Kurdistan university
    • Fardin Ahmadizar,
    • Houra Mahmoudzadeh, University of Waterloo
    • Mahdi Bashiri, Shahed University

    Home healthcare is a concept that is gently growing over time, defined to decrease pressure on in-patient beds of the hospital by preparing care to patients at home. This paper proposes a novel hybrid solution procedure for this problem, called a two-phase multi-attribute robust home healthcare problem. In the first phase, the best-qualified candidate locations for health facilities are evaluated by using a fuzzy analytic hierarchy process and grey rational analysis. In the second phase, a robust model considering several aspects such as overqualification cost, waiting time and overtime is proposed. The objective function minimizes the total cost. Finally, a case study from the city of Sanandaj, Iran, is utilized to validate the solution in real-world situations.

  • 04:30 PM - 05:00 PM

    Mass casualty events: a decision making tool for home health care to discharge conventional hospitals

    • Alain Guinet, presenter, INSA de Lyon
    • Eric Dubost, Centre Hospitalier Soins et Santé

    We need to admit in a Home Health Care structure a massive influx of patients requiring an early discharge from conventional hospitals, due to a terrorist attack, which requires freeing hospitalization beds at the earliest for the victims. The early discharge patients are transferable from a given release date and can be managed by the Home Health Care (HHC) structure until a due date in order to reach the patient's home and find the caregiver. The home Health care structure must plan the patient admissions with the objective to admit as soon as possible the most victims in conventional hospitals using the least amount of HHC human resources during the discharges of hospitalized patients. An admission-planning model is proposed, the bi-objective problem modelled is solved with CPLEX.