The Effect of Online Reviews on Physician Demand: A Structural Model of Patient Choice

Introduction. Consumers in the healthcare market typically face uncertainty regarding the doctor’s quality, and may rely on imperfect signals to infer quality measures such as accuracy of diagnosis, bedside manner, and operational factors such as service time, waiting time, and appointment availability. By enabling large-scale distribution of information from countless other consumers, those consumer review websites can help resolve information asymmetries among a much broader peer group than has been traditionally possible. This paper provides empirical evidence on the impact of patient-generated reviews on demand in the outpatient healthcare market. We integrate and extend the advances in online marketing and healthcare operations literature, taking a step toward a more detailed understanding of patients’ choice behavior in the healthcare online markets and how it can help doctors improve the demand of outpatient care through operational strategies. The key challenges of this paper are to integrate the various sources of social media (both textual and numeric information), and leverage the aggregated level demand data into a consumer choice model. With the rapid growth and popularity of user-generated content (UGC) on the web, a new area of research has emerged that applies text mining techniques to product reviews. We have seen more and more papers in the information systems and marketing literature incorporate text information into the consumer choice models; however, limited research has looked at mining professional service reviews with operational factors and choice models in mind, and limited research in empirical operations management has utilized the information from text mining. One potential difficulty in mining doctor service reviews is deriving the operations based service features, namely, waiting time and service time. Compared to product reviews, the features for doctor service reviews have a more diverse format, due to service intangibility. Our goal here is to fill this gap by proposing a patient choice model that allows for doctors to be differentiated along multiple dimensions, contains richer distributions of taste parameters, and incorporates operational factors derived from online user-generated information.