College of Public Health receives NIH grant to pilot AI chatbot for African Americans with depression

(Press-News.org) As a leader in innovative health solutions, George Mason University’s College of Public Health received a National Institutes of Health (NIH) AIM-AHEAD program grant to pilot an artificial intelligence (AI) chatbot for Black and African Americans with depression. Health Informatics Professor Farrokh Alemi will enhance his first-of-its-kind, evidence-based artificial intelligence tool to address the medication needs of African Americans with depression.  

The existing AI tool recommends antidepressants for 16,775 general-population patient subgroups, each representing a unique combination of medical history. For each of these subgroups, the current project will analyze the effectiveness and appropriateness of the recommendations for African Americans, using the NIH All of Us database and existing published literature. 

To the researchers' knowledge, this is the first research focused on developing and evaluating an antidepressant recommendation system for Black and African American people.   

“Antidepressant medications are a first-line treatment for depression; however, a majority of depressed patients do not experience improvement with their first antidepressant. Additionally, minority populations, including Black and African Americans, are not well represented in antidepressant studies, contributing to reduced antidepressant effectiveness in these populations,” said Alemi. “There is a significant need to synthesize available evidence regarding antidepressant effectiveness and provide personalized treatment recommendations, and this project addresses a major gap in the management of Black and African Americans with depression.” 

Researchers will develop a Knowledge-enhanced Antidepressant Recommendation Dialogue System (KARDS), which will engage users in a back-and-forth conversation to acquire the patient information needed to identify the appropriate antidepressant medication. The AI will provide the patient with a list of recommended medications, list of the relevant studies, and an explanation for the medication decisions. The system will automatically send the patient’s clinician a brief point-of-care recommendation and explanation, with an option to examine a complete record of the conversation and the supporting evidence. 

“Chatbots—or patient-facing dialogue systems like the one we will create—hold transformative potential for the health care sector and are increasingly prominent in psychiatric applications, predominantly through therapy-bot implementations,” said Alemi. “Our chatbot will help improve the detailed, time-consuming, medical history intake process, and provide point-of-care summary and prescription recommendations to the patients’ clinicians. The chatbot will make patients more comfortable because the natural language modality provides an intuitive, empathetic, stigma-free interface.” 

Once the AI chatbot is developed, the team will test the dialogue system with Black and African American patients to evaluate system functionality and user preferences. Additionally, the project will train a Black or African American doctoral or master’s student in AI, expanding the available workforce and building the community’s capacity to address AI. 

Alemi will lead the research team, which includes Janusz Wojtusiak, a George Mason professor of Health Informatics and the director of the Machine Learning and Inference Laboratory, and Kevin Lybarger, a George Mason assistant professor in the Department of Information Sciences and Technology in the College of Engineering and Computing. All three members have collaborated previously to diagnose COVID at home from presenting symptoms. 

The $70,906 grant is part of the NIH’s AIM-AHEAD (Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity) program, which aims “to establish mutually beneficial and coordinated partnerships to increase the participation and representation of researchers and communities currently underrepresented in the development of AI/machine learning models and enhance the capabilities of this emerging technology, beginning with electronic health record data.”

END