Analytics-Driven Design of Multi-Phase Multi-Provider Appointment System for Patient Scheduling

Restricted (Penn State Only)
Author:
Srinivas, Sharan
Graduate Program:
Industrial Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
June 02, 2017
Committee Members:
  • Dr. A. Ravi Ravindran, Dissertation Advisor
  • Dr. A. Ravi Ravindran, Committee Chair
  • Dr. M. Jeya Chandra, Committee Member
  • Dr. Vittal Prabhu, Committee Member
  • Dr. Dennis K. J. Lin, Outside Member
Keywords:
  • patient scheduling
  • open access
  • multi-phase multi-provider
  • scheduling rules
  • appointment system design
  • hybrid appointment system
  • data analytics
  • machine learning
  • patient no-shows
Abstract:
Outpatient scheduling plays a key role in matching the healthcare provider capacity to patient demand and improving clinic performance measures, such as patient waiting time, patient satisfaction, and resource utilization. In addition to the traditional pre-booked appointments, outpatient hospitals and clinics are also experimenting with same day appointments. Designing a hybrid appointment system (combination of same-day and pre-booked) involves multiple decisions such as determining the appointment types, patient sequence, and appointment time. Further, various factors such as patient flow, demand uncertainty, and patient no-shows (patients who do not arrive for scheduled appointments) must be considered to develop an effective design. Inefficiencies in the appointment system design and patient no-shows cost the U.S. healthcare system more than $150 billion a year. In addition, they also reduce productivity and timely access to care. Most of the previous work on outpatient appointment systems consider a simplified clinic setting with single phase (one-stop service) and single provider. Further, they rarely consider patient’s provider preference, patient availability, patient specific no-show rate, uncertainty in patient demand and service times. However, in practice, most outpatient departments have multi-phase settings (e.g., pre-screening, visit nurse, visit doctor, checkout) with multiple providers. A detailed simulation analysis indicated that ignoring the multi-phase nature of patient flow, patient’s provider preference and patient’s availability lead to unmet demand, patient dissatisfaction and inefficient resource utilization. Further, the associated uncertainties complicate the task of designing the appointment system. This research focuses on designing a data-driven multi-phase multi-provider appointment system for outpatient clinics with the objective of improving resource utilization and patient satisfaction. First, a new approach to design a hybrid appointment system, a combination of pre-booking and open access (same day) appointment types, is proposed. The objective is to determine the schedule configuration of a hybrid appointment system under uncertainty for a multi-phase multi-provider clinic that incorporates patient’s provider preference and availability. A mathematical programming model is proposed to determine the optimal percentage of appointments reserved for pre-booking and open access, and a scenario-based Monte Carlo approach is used to account for uncertainty. Finally, heuristics are developed to determine the best configuration for the hybrid appointment system. Next, a new framework for sequentially scheduling patients is proposed by using a combination of data analytics and simulation. In the proposed framework, patient-related data from various sources are used to develop predictive models to identify the risk of patient no-show. Finally, different scheduling rules that leverage the patient specific no-show risk are proposed. Their effectiveness is evaluated with respect to current scheduling practices. The results indicate that the proposed rules consistently outperform the current practice for all the clinic settings tested. A case study with real data from a Family Medicine Clinic in Pennsylvania is used to show the feasibility and applicability of the proposed models. The analysis of the results provided several key insights in designing an appointment system, which are applicable to both researchers and practitioners. Further, the proposed approaches are generic and can be adopted by any outpatient clinic by incorporating their clinic parameters, such as operating hours, slot duration and others.