Menu
PREPLEX
An algorithm to automate the balancing of supply-demand and optimize the management of resources in the outpatient department of a hospital
Hospital Universitario del Sureste
The ‘Hospital Universitario del Sureste’, is a public hospital within the Autonomous Community of Madrid in Spain that provides primary and secondary care for around 200.000 people. Its sphere of influence covers both big towns and small rural areas. The whole region has suffered a massive increase in population in recent years (the projected number was around 170.000 by 2025 when the hospital was built) so the resources are always strained.
The Hospital Information System (HIS) currently used is Selene, a solution developed by CompuGroup Medical. Everything related to management of supply and demand in the outpatient department is stored within this system, so only one data source is needed for the purpose of solving the challenge.
Our organization has a strong data engineering department with direct access to the system’s database and sound knowledge of the data models so access to the information should not be a problem.
PREPLEX
As in many other healthcare organizations around Europe, the demand in our outpatient department is structured around a system of waiting lists implemented using slots to help manage the available resources. This is how it works:
- The hospital has resources (like ultrasound scanners).
- Each resource has a schedule composed of slots.
- Every slot is predefined to accommodate only certain healthcare services (like abdominal echography) and priorities (urgent, preferential and normal).
- Physicians make service requests for an available slot against the resources.
- Each service request includes parameters like the requested healthcare service, priority and indication date.
- Once the service request has been processed, an appointment is created occupying a slot.
It is very important to stress that these schedules and slots are predefined for a certain period of time before any appointment is even admitted. The reason behind this is that we are not talking about a pure first-come, first-served basis. We want to segment patients into different waiting lists, each with different waiting times as not every healthcare service and priority require the same response times.
Applications closed on the 24th of January 2024 at 17:00 CET.
GRAPHENUS
Graphenus is a Big Data distribution, which manages and controls the end-to-end life cycle of data: from infrastructure management and monitoring to data storage, analysis and governance.
Graphenus distribution was born with a clear purpose: to provide maximum security and flexibility to respond to any data processing need. Graphenus allows, based on information collected from different sources and stored reliably, to perform complex analytics in batch or streaming, ensure the persistence of data models, or develop machine learning models. In addition, the platform offers companies and public administrations pre-built and secured solutions to respond to specific cases, being easily deployable, scalable and versionable.
The aim is to develop, train and validate an algorithm that support clinicians and administrative staff to schedule the different medical services, presenting them with the most efficient options. This will be achieved considering demand forecasting, the constraints of the service catalogue and the different types of medical appointments, each with different waiting times and priorities. Our algorithm will be able to solve the two main issues that are causing the bottleneck: The expected number of patients that will require medical services in the following months. Provide an optimal resource scheduling proposal taking into account the estimated demand and available resources. GRAPHENUS solution does not replace any existing system but will be integrated with the current Selene and enables low-cost and frequent simulations to be performed. The pilot will be carried out in 3 medical specialties: Dermatology, ENT and Radiology, as they will allow to add time constraints based on their surgical activity and to test the algorithm when two health services are linked together.
