Evaluating Conversational Artificial Intelligence for Depression Management

The conversational AI intervention utilizes interactive dialogue to collect patient medical histories. Powered by advanced Large Language Models (LLMs), the system engages participants in open-ended, natural language conversations. A Dialogue Manager coordinates the dialogue, ensuring conversations remain medically relevant, personalized, and safe. Clinicians monitor the interactions in real-time, validating and ensuring accuracy of antidepressant recommendations informed by an evidence-based Antidepressant Knowledgebase.

"The structured questionnaire intake utilizes a predefined, multiple-choice survey that systematically collects patient medical history without conversational interaction. Questions are algorithmically selected based on the patient's earlier responses to maximize information efficiency, typically requiring fewer than 13 items. Patients select from standardized responses (including options like I do not know or I do not want to say). Contextual help messages are provided as needed. Upon completion, the system generates antidepressant recommendations using the same Antidepressant Knowledgebase as the Conversational AI system, clearly explaining its rationale. A prototype version is available at http://MeAgainMeds.com, with ongoing development to enhance features and usability."