DEPRA: An Early Depression Detection Analysis Chatbot

Abstract

The application of Artificial Intelligence (AI) in the assessment and treatment of mental health has gained momentum in recent years due to the evidential development of chatbots. There are promising outcomes from recent attempts such as facilitation to detect the depression level in patients’ profiles, which have improved the aspiration of finding a solution to assist medical professionals in detecting depression. However, experts believe the promise is still far from the expectations since most of the chatbots found in literature has conscious decision from selectable answer. In addition, the participants are required to have longer period of the interactions with the chatbot which suffer great losses of the participation. Furthermore, the user privacy and scientific evaluations of early depression detection are not guaranteed due to the customized chatbot platforms. Motivated by these, we proposed and developed DEPRA based on contemporary bot platforms with early depression detection to tackle the mental health symptoms. DEPRA is built on Dialogflow as a conversation interface and uses personalized utterances collected from a focused group to train it. Moreover, the interaction time was reduced remarkably by the setup of DEPRA. A structured early detection depression interview guide for the Hamilton Depression Scale (SIGH-D) and Inventory of Depressive Symptomatology (IDS-C) underpins the formation. DEPRA can act as the proxy between the health professional and the patient. Moreover, the DEPRA integrated with social network platform which provide convenience of the attractions for the participants. More than 40 participants interact with DEPRA and the analysis of their response establishes the promise of its use in mass screening for early detection of depression. User experience survey demonstrates that the overall user satisfaction level is approbatory.