Comparison of Non-AI and AI-Enabled M-Health Platforms for COVID-19 Self Screening in Indonesia

Abstract

This study aimed to analyze and differentiate the role of AI and no AI-supported m-health platforms for COVID-19 self-screening in Indonesia. We utilized a mysterious shopping method to develop four standardized cases with various severity levels of COVID-19 tested in Indonesia's most popular mHealth platforms. We selected seven apps from the top 200 free mHealth apps in the 'Medical' category in the Google Play Store equipped with COVID-19 symptom checkers. A total of 36 teleconsultations were performed in four chatbot-based, two apps supported with AI combined with a human-based approach, and three apps with the human-based process. Teleconsultations were recorded, classified, and analyzed compared with the COVID-19 guideline by the MoH of Indonesia. The study indicated that most of the self-screening provided questions that had consistently led to the COVID-19 condition such as cough, fever, and shortness of breath and followed the guideline from the national health authority.