Scheduling physiotherapy treatment in an inpatient setting

Whilst many areas of healthcare have benefitted from purpose-built automated scheduling programs (e.g. nursing rosters, doctor’s surgeries, and operating theatres), physiotherapy has been somewhat left behind. This is a matter of concern, since inpatient physiotherapy is an important part of many treatment programmes in general hospitals and specialist rehabilitation units. This study seeks to help address this deficiency by detailing a computer program that has been introduced to timetable physiotherapy treatment at a major neurological rehabilitation unit in Cardiff, UK. The aim is to show how automated scheduling can provide a far more effective and efficient alternative to manual scheduling by hand. Whilst the scheduling procedure will undoubtedly differ from unit to unit, this paper outlines what are thought to be some of the most important objectives and constraints in addition to an adaptable methodology that can be employed to arrive at a good-quality solution in a timely fashion. The scheduling problem considered here is formulated as a multi-objective combinatorial optimisation problem due to the many competing objectives and constraints and the significant size of the solution space. An approximate solution is sought by means of a three stage local search based approach. The first stage is to construct a valid and good quality initial solution. The second stage is to improve this with respect to some of the more important soft constraints by applying a series of steepest descent based algorithms. Like hard constraints, these ’moderate’ constraints represent a discrete number of violations rather than contributions to an objective function. This allows for a more targeted approach in their solution. The third stage is then the optimisation of the (remaining) soft constraints using simulated annealing and tabu search. Results are favourable with typical timetable quality (as measured by constraint violations), being considerably better than under the former ’by hand’ approach. Since its introduction, employee time expended on timetable creation has been cut from eight hours per week to fewer than two, freeing up therapist time for clinical work.