A hybrid IT framework for identifying high-quality physicians using big data analytics. (August 2019)

• Record Type: Journal Article
• Authors: Ye, Yan; Zhao, Yang; Shang, Jennifer; Zhang, Liyi
• Is Part Of: International journal of information management. Volume 47(2019)
• Page Start: 65
• Page End: 75
• Publication Date: 2019-08
• Subjects: Online healthcare communities -- Physician identifying -- Signaling theory -- Machine learning -- Topic modeling -- Multi-criterion analysis; Social sciences -- Information services -- Periodicals; Social sciences -- Research -- Periodicals; Information science -- Periodicals; Management information systems -- Periodicals; Knowledge management -- Periodicals; Sciences sociales -- Documentation, Services de -- Périodiques; Sciences sociales -- Recherche -- Périodiques; Sciences de l'information -- Périodiques; Systèmes d'information de gestion -- Périodiques; Information science; Management information systems; Social sciences -- Information services; Social sciences -- Research; Periodicals; Electronic journals; 025.52068
• DOI: 10.1016/j.ijinfomgt.2019.01.005
• Languages: English
• ISSNs: 0268-4012

Abstract

Highlights: Expertise similarity match alone is not enough for recommending physicians. Feedback, basic profiles and service quality are also valuable for doctor-finding. Propose a four-level model to identify high-quality doctors using signaling theory. Use Binary Long Short-Term Memory (Bi-LSTM) method to mining feedbacks. Elite factors that influence recommendation intention using the regression model. Abstract: Patients face difficulties identifying appropriate doctors owing to the sizeable quantity and uneven quality of information in online healthcare communities. In studying physician searches, researchers often focus on expertise similarity matches and sentiment analyses of reviews. However, the quality is often ignored. To address patients' information needs holistically, we propose a four-dimensional IT framework based on signaling theory. The model takes expertise knowledge, online reviews, profile descriptions (e.g., hospital reputation, number of patients, city) and service quality (e.g., response speed, interaction frequency, cost) as signals that distinguish high-quality physicians. It uses machine learning approaches to derive similarity matches and sentiment analysis. It also measures the relative importance of the signals by multi-criterion analysis and derives the physician rankings through the aggregated scores. Our study revealed that the proposed approach performs better compared with the other two recommend techniques. This research expands the boundary of signaling theory to healthcare management and enriches the literature on IT use and inter-organizational systems. The proposed IT model may improve patient care, alleviate the physician-patient relationship and reduce lawsuits against hospitals; it also has practical implications for healthcare management.