Guidelines for scheduling in primary care under different patient types and stochastic nurse and provider service times

Abstract

Scheduling in primary care is challenging because of the diversity of patient cases (acute versus chronic), mix of appointments (pre-scheduled versus same-day), and uncertain time spent with providers and non-provider staff (nurses/medical assistants). In this paper, we present an empirically driven stochastic integer programming model that schedules and sequences patient appointments during a work day session. The objective is to minimize a weighted measure of provider idle time and patient wait time. Key model features include: an empirically based classification scheme to accommodate different chronic and acute conditions seen in a primary care practice; adequate coordination of patient time with a nurse and a provider; and strategies for introducing slack in the schedule to counter the effects of variability in service time with providers and nurses. In our computational experiments we characterize, for each patient type in our classification, where empty slots should be positioned in the schedule to reduce waiting time. Our results also demonstrate that the optimal start times for a variety of patient-centered heuristic sequences consistently follow a pattern that results in easy to implement guidelines. Moreover, these heuristic sequences and appointment times perform significantly better than the practice's schedule. Finally, we also compare schedules suggested by our two-service-stage model (nurse and provider) with those that only consider the provider stage and find that the performance of the provider-only model is 21% worse than that of the two-service-stage model.