Planning and Optimisation

Planning and Optimisation

Reducing waiting lists and times for people requiring elective surgery.

 The focus of much recent Australian government policy has been on reducing waiting lists and times for people requiring elective surgery.
Optimal and timely resource management is key to the successful operational management of modern hospital systems. Predicting and scheduling the use of hospital beds, theatres, staff and equipment presents a challenging problem where the complexity and changeability of interacting factors demands a flexible and dynamic solution in order to achieve a high level of utilization.

This work also encompasses public health research such as disease surveillance and outbreak/anomaly detection via spatio-temporal modelling and epidemiology. For example, a new strain of flu might hit the state but affect a particular age group more dramatically, say pre-schoolers; early identification of this group would allow for a shift in paediatric resources as well as the possibility of targeted public interventions/awareness campaigns. Similarly a biohazard event might cause an increase in gastrointestinal illness in a particular area, early detection of which areas would enable public health officials to begin determining the source as quickly as possible.

These measures can inform and shape appropriate response policy for state health departments.

It is proposed to incorporate the important aspect of resource scheduling to work beside the prediction models. The health system has a number of high cost health service assets – from theatres to imaging machines – dependent on many resources being available – such as doctors, technicians, nurses, patients. Many of these assets and resources are left idle due to inefficient planning and late changes to schedules. Initial work by an AEHRC PhD student has devised a new algorithm that builds and maintains a single schedule for hospital theatres based on multiple departmental schedules. The schedule then uses agents to negotiate between the departmental schedules and a distributed constraint model to find the best solutions to conflicts. The challenge is to show that such an approach can lead to better utilisation.

Key Research Questions

  1. What are current practices regarding patient flow decision making?
  2. How can prediction tools be used daily and weekly to assist in bed management decision making (e.g. trigger points to escalate concerns and responses to those escalations)?
  3. Can we ultimately model the health system based on the number of patients discharged, the time of day, the number of elective surgeries performed, and derive the affect this has on occupancy and access block?
  4. What choice of syndrome groupings should be defined (eg. by ICD-10 codes)? Modelling of the counts of presentations for these syndromes affect the nature of the presentations and the ability to forecast ahead.
  5. What methods (eg. Surveillance Trees) should be implemented for monitoring counts on a daily basis?
  6. Do optimised schedules (eg. for operating theatres) lead to better resource utilisation?