Machine learning based system for optimizing the allocation and availability of beds to deal with uncertainties and enable semi-automated bed-management.
This project would broadly fall under the category of hospital management. Managing the allocation and availability of beds in a hospital is an ongoing problem in hospital managenent, and something of this would be a necessity in almost any hospital that sees large volumes of patients and has an infrastructure consisting of a large but fixed number of beds. The problem is essentially that of managing the inherent uncertainty in the 'occupancy' of a bed by a patient at any given time --- discharge dates, different categories of occupancy (ICU, recovery/observation, general ward for waiting), patient admission types (private, special, shared wards), uncertainty in the arrivals of new patients who may need to be accommodated (emergency/trauma patients, discretionary admissions,...), inherent uncertainty in patient condition, their individual responses to treatment, doctor's diagnosis/assessment and consequently the amount of time different patients spend in the hospital. Irrespective of any of these factors it is necessary that every patient who is admitted needs a bed allotted to the patient. In view of all these uncertainties, planning the bed infrastructure, allotting future dates for discretionary admissions to patients, dynamic bed allocation depending on the current level of care and/or procedure different patients need to undergo becomes a huge challenge. Though this is a crucial problem that all hospitals need to deal with for their effective functioning, it is unfortunate that medical interns/resident doctors spend enormous amount of time dealing with this logistics. Clearly this has no direct relevance to their professional learning as a medical practitioner. The attempt in this project is to build a machine learning based system that can deal with this uncertainty and enable semi-automated bed-management. This in turn would allow the medical professionals to spend most of their time with what they need to be doing ideally --- treating patients.