Dominic Thomas
Every year, a huge number of 911 calls that are fundamentally mental health crises get routed through police dispatch, jail, or the emergency room — the three most expensive, and often least appropriate, places to handle a behavioral health crisis. My 911 Mental Health Crisis Response project, run out of the Modular Agile Deployment (MAD) Lab I co-direct at Kennesaw State, is building the technology to catch those crises earlier, so co-responder teams can intervene with a better option.
The technical core of the project is a set of novel NLP and AI models, developed with KSU graduate students and faculty from Computer Science and Social Work, that can classify behavioral health content in messy, unstructured first-responder narratives while meeting real ethical constraints around transparency and explainability — this is design science in the fullest sense: building software that has to work inside actual first-responder processes and hold up under scrutiny, not just perform well on a benchmark. That work has been published in Smart Health (a Tier 1 journal in our field’s ranking system) and presented at venues including IEEE BigData and the IEEE/ACM Conference on Connected Health.
Getting there required building research infrastructure as much as models: working with KSU’s IT organization to develop and test secure enclave computing for highly sensitive identified criminal and health data, which in turn led to new policy and data-sharing templates supporting the university’s R2 research designation — infrastructure that now benefits other KSU faculty working with similarly sensitive data. One of my Ph.D. students, Martin Brown, successfully defended his Data Science dissertation on this project in April 2024, developing novel ensemble, active-learning, and fusion techniques that pushed classification accuracy on very messy, unstructured data to an extraordinary .99. Another student on the project, Adnan Azmee, won the CCSE Outstanding Researcher Award for 2024.
The project has been funded by a Partnership for Inclusive Innovation (PIN) grant that, in 2024, made Kennesaw State the first university included on the Georgia Department of Economic Development’s list of PIN-supported projects, plus a subsequent Fitzgerald Foundation grant. The work has been presented to the U.S. Department of Homeland Security and, most recently, to the National Academies’ U.S. Committee on the Maritime Transportation System, and is the basis of a current National Science Foundation Smart Connected Communities proposal, developed with colleagues in Computer Science and Social Work, to build data analytics and management software that helps divert mental health crisis cases away from emergency rooms and jails.