BACKGROUND
A growing telehealth industry, the online health community (OHC) bridges the gap between physicians and patients by offering an e-service of a new sort through telemedicine - online medical consultation (OMC), a solution that alleviates problems associated with unbalanced distribution and inadequate high-quality medical resources.
OBJECTIVE
A personalized recommendation for OMC services is imperative to reduce patient information overload and optimize physician resource utilization. This paper aims at providing an overview of current trends and outlining e-Service-oriented paradigms and approaches to recommending OMCs.
METHODS
Our systematic literature search was limited to quantitative, qualitative, and mixed-method papers published from 2001 to 2020 through ACM Digital Library, EBSCO, Springer, PubMed, and Google Scholar. Using cross-disciplinary literature on medicine, information systems, and artificial intelligence, this paper summarizes personalized recommendations for online services, which involve many aspects, such as two-sided matching for patients and physicians, interpretable recommendations for patients, and workload balancing for physicians.
RESULTS
The paper finds that e-service-oriented recommendations are an emerging concept that has not yet been clearly defined and fully investigated. We try to break the inertia associated with tinkering with the traditional personalized recommendation models, establish an innovative theoretical framework for e-service-oriented recommendations, and propose some critical technical issues of two-sided personalized recommendations.
CONCLUSIONS
OMC is a knowledge-intensive and labor-intensive service, where patients lack expert knowledge and demand interpretable recommendations; physicians have varied energy levels and cannot afford overwork. E-service recommendations need to face two-sided users with different cognitive abilities, expectation levels, decision-making perspectives, and preferences, so they require an entirely different paradigm needs to develop distinct attributes and study unique contents.