The study determined that 58% of NeuroFlow users whose suicidal ideation was detected by NLP may not have been identified otherwise

PHILADELPHIA, Feb. 29, 2024 /PRNewswire-PRWeb/ — NeuroFlow, a behavioral health technology and analytics company helping to manage and measure population risk, has published a new paper titled, “Using Natural Language Processing to Detect Suicidal Ideation and Prompt Urgent Interventions” in the journal of Innovations in Digital Health, Diagnostics, and Biomarkers. The findings shed light on the new and emerging capabilities of artificial intelligence (AI) for improving the detection of suicide risk among populations.

Previous studies have shown that 45% of people who completed suicide visited their primary care doctor a month prior to their death – a devastating example of suicide risk hiding in plain sight in the current healthcare ecosystem. The challenges of timely identification of suicidal ideation are substantial; many individuals aren’t forthcoming when asked about suicidal ideation directly, and many healthcare providers hesitate to pursue conversation about suicide with patients, fearing liability.

With suicide claiming roughly 48,000 lives in the United States each year and costing the country over a trillion dollars annually, NeuroFlow’s research showcases the value of AI as an additional safety net beyond conventional screening for flagging high-risk individuals and immediately connecting them to appropriate resources. By analyzing free-form text entries submitted by users of NeuroFlow’s engagement platform, the company’s proprietary natural language processing (NLP) algorithm successfully identified indicators of suicidal ideation in 425 users’ self-directed written reflections, leading to timely interventions and potentially life-saving support.

When evaluating activity of these users in the 30 days prior to their NLP-detected suicidal ideation, NeuroFlow found that:

  • 81% of users had completed a PHQ-9 assessment (the most commonly used clinical tool for surfacing suicidal ideation), almost half of whom indicated not having suicidal ideation.
  • The remaining 19% did not complete a PHQ-9 at all – missing an opportunity for detection.
  • In conclusion, more than half of these users (58%) may not have been identified as having suicidal ideation without NLP.

These results highlight the pivotal importance of a multi-faceted approach to identifying and supporting at-risk populations. NeuroFlow’s technology can understand the context of language, making it a critical tool for identifying high-risk words or phrases. When NLP identifies this language, an alert is generated, triggering immediate delivery of localized crisis resources and timely outreach from Response Services, a team of coordinators who are trained in crisis response. Response Coordinators screen for suicide risk and navigate patients to the appropriate level of care as needed, helping organizations optimize use of their limited care resources and diverting patients from emergency settings like the ED when lower levels of care are more appropriate.

NeuroFlow Chief Medical Officer Tom Zaubler, MD, a co-author of the study, expressed enthusiasm about how these findings translate to impactful suicide prevention at scale: “This research is a significant step forward in demonstrating the meaningful application of AI and digital health technologies in mental health care. Effective suicide prevention requires intervening upstream, engaging people in settings where they are comfortable being forthcoming about how they are feeling. By taking an innovative approach to identification that expands beyond traditional screening alone, NeuroFlow is resetting the standard of care for suicide prevention.”

The technology’s scalability and location-agnostic nature make it particularly valuable in supporting marginalized communities with limited access to psychiatric resources. Moreover, the research shows how the thoughtful application of AI combined with compassionate human intervention provides life-saving value in healthcare. “Our commitment to leveraging technology for mental health extends far beyond this study,” NeuroFlow Chief Product Officer Julia Kastnersaid. “We believe in the power of AI and digital health to transform the way we approach mental health care, and this research is a testament to NeuroFlow’s dedication to advancing the field and improving outcomes for individuals facing mental health challenges.”

NeuroFlow has worked with leading researchers across the field of healthcare to validate its approach to supporting mental health care. The team’s data-driven analysis has led to a series of reports published with peer review. For the complete study, and access to other published work from NeuroFlow, please visit here.

About NeuroFlow

NeuroFlow helps risk-bearing healthcare organizations improve outcomes and cost of care in medically complex populations by surfacing and supporting behavioral health needs that typically go undetected and under-addressed. Across payors, providers, and the federal government, NeuroFlow’s scalable technology and analytics capabilities empower organizations with the behavioral health insights they’re missing to manage these populations in a financially sustainable way. Powered by deep expertise in whole-person care, NeuroFlow offers a path to risk predictability and proactive care that helps overcome the systemic challenges in today’s healthcare ecosystem.

Enter your keyword