Harnessing Artificial Intelligence to Improve the Quality of Answers in Online Question-answering Health Forums

Effects of emotional expression and platform prompts on patient selection: A multi-method study on online question and answer platforms

Patients lacking medical expertise in online health Q&A platforms find it difficult to assess the quality of physician responses. Emotional support provided by doctors and prompts from the platform may influence patient selection. We developed a conceptual model based on emotions as social information theory and the Stimulus-Organism-Response framework. The model was validated using a multi-method and multi-study approach. In Study 1, we analyzed platform data using natural language processing and machine learning techniques to explore the impacts of physician expression characteristics on patient selection. In Study 2, we conducted an experiment to explore patient perceptions and influence mechanisms in the process. Study 1 showed the effects of empathy, encouragement, and using default responses on patient selection. Study 2 revealed the importance of psychological empowerment and uniqueness neglect on patient decisions. The findings have significant implications for physicians and platforms in designing improved health services and optimizing systems.

Knowledge mapping of online healthcare: An interdisciplinary visual analysis using VOSviewer and CiteSpace

Background

Online healthcare has been regarded as a permanent component and complementation in routine worldwide healthcare. Although there have been large-scale related studies in this field, studies are scattered across disciplines. Numerous publications are needed to systematically and comprehensively identify the status quo, development, and future hotspots in this field.

Methods

Publications on online healthcare were screened from the WoS database. By using VOSviewer and CiteSpace, this study analyzed 4636 articles in this field with 60,306 associated references. First, countries/regions distributions, institutions distributions, influential journals, and productive authors were obtained. Then, co-citation analysis, co-occurrence analysis, timeline analysis, and burst detection were further conducted to sketch the panorama of online healthcare.

Results

There were 147 countries/regions participated in and contributed to this field in total. Accounting for over half of the total number of publications, the USA, England, Australia, China, and Canada played significant roles in this area. Among the 24,362 authors, Guo XT was the most influential author. The International Journal of Environmental Research and Public Health was the journal with the most publications and citations. Studies in this field have shifted from basic research to applied practice research. COVID-19, mental health, healthcare, and healthcare workers were the most common keywords, indicating that studies on the impact of online healthcare on healthcare workers, online healthcare service for COVID-19, and mental health will be promising areas in the future.

Conclusions

Research on online healthcare is booming, while worldwide cooperation is still regionalized. Cross-regional cooperation among institutions and scholars is needed to enhance in the future. Online healthcare services for specific health fields and specific groups are the current and developing topics in this field.

The Effects of Online Text Comments on Patients' Choices: The Mediating Roles of Comment Sentiment and Comment Content

As an increasingly important application of mobile social media usage, online healthcare platforms provide a new avenue for patients to obtain and exchange information, referring not only to online doctor's advice but also to the patients' comments on a doctor. Extant literature has studied the patients' comments facilitated with the direct numeric information gathered in the web pages including the frequencies of "thanks letter," "flowers," and "recommendation scores." Adopting the text analysis method, we analyzed patients' comments on the healthcare platform, focusing on the comments from two aspects, namely, comment contents and content sentiment. Based on the analysis of the data collected from one of the most popular healthcare apps named "Haodaifu" in China, the results show that the vast majority of the comments are positive, which basically follows the L-shaped distribution. Meanwhile, comment sentiment covering sentiment tendency and proportion of positive comments demonstrates significant effects on recent 2-week consultation by a doctor. One of the comment contents "patience explanation" has significant effects both on the total consultation and recent 2-week consultation by a doctor. The research findings indicate that the online preferences for and evaluations on doctors provide strong support and guidance for improving doctor-patient relationships and offer implications for medical practices and healthcare platforms improvement.

Q&A participation and behavioral patterns on academic social networking sites: a comparative study of academic, corporate and government institutions on ResearchGate

Purpose

This paper aims to explore question and answer (Q&A) participation and behavioral patterns on academic social networking sites (ASNSs) from the perspective of multiple subjects such as academic, corporate and government institutions.

Design/methodology/approach

Focused on the Q&A service of ASNSs, this study chooses ResearchGate (RG) as the target ASNS and collects a large-scale data set from it, involving a sample of users and a Q&A sample about academic, corporate and government institutions. First, it studies the law of Q&A participation and the distribution of the type of user according to the sample of users. Second, it compares question-asking behavior and question-answering behavior stimulated by questions among the three types of institutions based on the Q&A sample. Finally, it discusses the Q&A participation and behavioral patterns of the three types of institutions in academic Q&A exchanges with full consideration of institutional attributes, and provides some suggestions for institutions and ASNSs.

Findings

The results show that these three types of institutions generally have a low level of participation in the Q&A service of RG, and the numbers of questions and answers proposed by institutional users conform to the power-law distribution. There are differences in Q&A participation and Q&A behavioral patterns among academic, corporate and government institutions. Government and academic institutions have more users participating in the Q&A service and their users are more willing to ask questions, while corporate institutions have fewer users who participate in the Q&A service and their users are inclined to provide answers. Questions from corporate institutions attract much more attention than those from the other two types of institutions.

Originality/value

This study reveals and compares the Q&A participation and the behavioral patterns of the three types of institutions in academic Q&A, thus deepening the understanding of the attributes of institutions in the academic information exchange context. In practice, the results can help guide different institutions to use the Q&A service of ASNSs more effectively and help ASNSs to better optimize their Q&A service.