Topic Modeling for Online Health Consultation on Low-Risk Diseases

Online Health Consultation (OHC) can improve healthcare quality, efficiency, and equity by offering medical services online and overcoming geographical barriers for those far from health institutions. However, this huge dataset of Q&As from OHCs has yet to reveal Indonesians' online health information needs for low-risk diseases. Therefore, the distribution of topics should be analyzed and identified to find trends in low-risk diseases and group the topics into the most important clusters. The proposed approach for this research is topic modeling using LDA. This study aims to identify topics derived from textual analysis of physicians' answers on low-risk diseases in OHC and investigate the trends of low-risk disease topics based on topic modeling results in order to provide personalized information service recommendations and improve OHC service quality, especially for information seekers or users of OHC services on low-risk diseases. This research contributes to identifying trends in low-risk diseases based on OHC history from Indonesia and describing topic clusters from topic modeling based on physician responses. The results show that based on topic modeling with LDA, physicians identified 33 topics and two trends in online health consultations for low-risk diseases (12 diseases) and non-diseases, such as healthy lifestyle, pregnancy, and medical procedures.