Stochastic Models for Capacity Planning in Healthcare Delivery: Case Studies in an Outpatient, Inpatient and Surgical Setting

U.S. healthcare system has become far too complex and costly to sustain and operations research has much to contribute in improving health systems by addressing a large spectrum of problems. We study capacity planning in healthcare while considering the case-mix of patients, using stochastic modeling in different application areas: primary care, inpatient bed allocation and (spine) surgery scheduling. This body of work was developed over four years of collaborative research with hospitals and healthcare providers. The main objective of our research in primary care is to optimize the patient mix of primary care physicians in a group practice to maximize patient-clinician continuity and access. To model case-mix, we use the number of simultaneous chronic conditions (comorbidities) a patient has as a predictor of the number of appointment requests. We later extend the optimization framework and use queuing theory to develop methodologies to quantify and evaluate access to care and continuity of care for patient visits with different urgencies. From an inpatient care perspective, we develop an empirically calibrated simulation model to represent a time-varying multi-server queuing network model with multiple patient classes. Our main focus has been on quantifying the impact of discharge profiles to alleviate inpatient bed congestions. The main objective of our research in surgical care is to create better patient access and improve revenue as a result of increased surgical capacity with more efficient schedules and an improved patient mix, using a multi-stage mixed integer optimization