Outpatient Appointment Scheduling Study: Utilization Projection, No-Show Prediction, and Capacity Allocation
Abstract
Long waiting times could result in many negative effects, such as low capacity utilization, high patient no-show rates, and loss of social benefits, which will also lead to a waste of public resources. Therefore, to better utilize healthcare resources and serve the community, my dissertation will focus on three objectives: To study the relationship between appointment utilization and indirect waiting time (IWT); to predict the patient’s no-shows without profiling them; to develop an optimization model for appointment capacity allocation.
To achieve these objectives, multiple models and approaches have been developed in this dissertation. For the first model, two mixture distribution models, including a beta geometric (BG) and a discrete Weibull (BdW) model were carried out to project the appointment utilization over IWT. The results indicated that appointment utilization is positively related to the IWT but tends to fluctuate after the first couple of weeks. Two mixture distribution models were also proved to be more accurate for projecting the appointment utilization when compared with commonly used curve-fitting models.
For the second objective, a conditional inference tree model was applied to predict the patient’s no-show probability and classified the no-show probability without profiling patients. This model was also compared with the general linear model and typically used logistic model, the result showed that using the conditional inference tree model with classified data will lead to a more accurate prediction and higher R-squared value.
For the final objective, three optimization methods and two scheduling strategies were examined. The proposed solution of capacity allocation provided a more robust, flexible, and efficient allocation plan for outpatient appointments, which significantly improved the average daily profit and capacity utilization rate.
By completing those three objectives, this dissertation did not only provide a more accurate way to monitor and predict outpatient appointments but also proposed a more practical and efficient appointment capacity allocation strategy. This will help our society save healthcare resources, reduce unnecessary costs for the healthcare providers, and provide better healthcare services to the community.