Dynamic Recommendation of Physician Assortment With Patient Preference Learning

Abstract

Web-based appointment systems are emerging in healthcare industry providing patients with convenient and diversiform services, among which physician recommendation is becoming more and more popular tool to make assignments of physicians to patients. Motivated by a popular physician recommendation application on a web-based appointment system in China, this paper gives a pioneer work in modeling and solving the physician recommendation problem. The application delivers personalized recommendations of physician assortments to patients with heterogeneous illness conditions, and then, patients would select one physician for appointment according to their preferences. Capturing patient preferences is essential for physician recommendation delivery; however, it is also challenging due to the lack of data on patient preferences. In this paper, we formulate the physician recommendation problem based on which the preference learning algorithm is proposed that optimizes the recommendations and learns patient preferences at the same time. Since the illness conditions of patients are heterogeneous, the algorithm aims to make personalized recommendation for each patient. Besides demonstrating the effectiveness of algorithm performance in terms of regret bound, we also provide extensive numerical experiments to show the expected algorithm performance under heterogeneous reward scenarios and performance comparison with algorithms in the literature under fixed reward scenarios. We introduce the flexibility of adjusting preference estimate update interval into our algorithm and conclude that short update interval contributes to short-term performance while long update interval leads to good results in the long run. Furthermore, we analyze how preference bound helps the algorithm to make explorations, which constitute two major contributions of our algorithm. Finally, we discuss the relevance between patient preferences and physician utilization and present a utilization-balancing approach that is effective in numerical experiments. Note to Practitioners —This paper aims at addressing the challenge of recommending physicians to patients on web-based appointment systems, so as to achieve efficient and effective utilization of physician resources. Currently, the recommendations are delivered manually by a huge number of general practitioners hired by the platforms based on their expertise and experiences. Upon each arrival of patient, they upload symptom statements of their own illness, and then, general practitioners will give a score or a rating on how well these patients are matched to physicians on the basis of symptom descriptions. With the development of artificial intelligence and information technology, these works might be done efficiently and effectively by algorithms automatically. To accomplish that, challenges will be presented: 1) physicians with adequate expertise might conflict with patient preferences, i.e., patients have strong preference toward senior specialists but their illness should be taken care of by a junior physician. This conflict indicates the importance of capturing patient preferences and choices in web-based appointment system operations and 2) the lack of patient preferences information makes the problem even harder. For newly developed web-based system, there are no available historical data to infer the preference parameters. On the other hand, the historical data in the web-based appointment system might not be able to reveal full information on patient preferences, because the system did not conduct effective and comprehensive explorations over all possible operations in the past. We capture the heterogeneous illness conditions in terms of rewards to imply the potential health improvement of patients. This paper assumes complete information of patient illness descriptions, and therefore, we focus on the operations management of appointment system, i.e., how to deliver the optimized physician assortment.