Design of Teleconsultation System with Artificial Intelligence Based Chatbot Using Docker Platform

Abstract

The implementation of the National Health Insurance Program (JKN) caused an increase in the number of visits to the first-level health facilities, while there are still clinics that lack medical personnel. Teleconsultation is defined as the possibility of contact between patients and doctors through a telemonitoring platform. Chatbot is a program that automatically provides services by talking to users through many types of communication media. A study shows a positive result in the use of support-chatbot for breast cancer patients, and that doctor agrees that chatbot can help in carrying out some simple tasks in healthcare scenario. Microservice is a system architecture that allows improvement of stability and security. We proposed a chatbot with three main functions: as a teleconsultation platform where doctors can arrange consultation questions that can either be reused for some patients with similar conditions or a specific set of questions for a particular patient; an artificial intelligence based disease diagnosis function made with the decision tree model using data that was collected from the results of literature studies; and disease information search function where the data was obtained from the results of web scraping. Each function was made in their respective services, allowing the usage of different technology according to their specific needs while providing easier maintenance, testing, and deployment. The chatbot's ability to lead the conversation flow was tested using predetermined test cases and direct user testing, the predetermined test cases shows positive results with the chatbot being able to correctly answer all the input as expected, while the user testing shows some areas of the system that could be improved further, including the addition of more commonly used disease term in the disease information database and more training for the chatbot.