Timeslot allocation for waiting list control

Abstract

As pressure on the healthcare system increases, patients who require surgery experience longer access times to pre- and post-operative appointments and surgery. Hospitals can control their waiting lists by allocating timeslots to the different types of appointments they discern. To inform patients about their appointments in a timely manner, they need to make this decision several weeks in advance. However, the precise consequences of the timeslot allocation on the future waiting list are uncertain, as not all patients follow the same treatment pathway. Furthermore, as these planning decisions are made weeks in advance, they are based on an uncertain prediction of future waiting lists. In this paper, methods are developed with the aim to support hospitals in optimizing their timeslot allocation to reduce patient access times and utilize all available capacity in the outpatient department and operating room. The problem is modelled as a Markov decision process (MDP). However, as the state, decision and outcome spaces grow exponentially in size, even for a single surgeon, an exact solution cannot be determined. We thus compare four alternative solution methods to the static allocation method that is common in hospitals. Least-squares policy iteration is used to find an approximate solution, an (integer) linear program is formulated to solve a deterministic variant of the MDP, several heuristic decision rules are investigated, and a hybrid method is proposed that statically allocates a percentage of timeslots and then dynamically allocates the remaining timeslots with the linear program when sufficient information is available to effectively deal with variability. The solution methods are tested on a case study at the orthopedic care chain of the Sint Maartenskliniek hospital in the Netherlands. Simulation results indicate that in balanced capacity scenarios, static allocation achieves the highest performance when planning more than 4 weeks ahead. In contrast, in unbalanced systems, steering capacity toward patient groups incurring the highest costs yields better outcomes. The hybrid method offers flexibility as it can be adapted to both balanced and unbalanced situations. For the case study, we find that statically allocating 60% of timeslots and dynamically allocating the remainder 4 weeks in advance provides the best results in terms of meeting access time targets and efficient resource utilization.