AI and SDoH Drive 7% ED Reduction at Saint Peter’s

Saint Peter’s University Hospital, a 478-bed safety-net provider in New Jersey, faced a familiar challenge: social determinants of health (SDoH) were contributing to avoidable emergency department (ED) utilization, yet actionable insights were rarely available at the point of care. To address this gap, the organization partnered with Lightbeam Health Solutions to deploy an AI model that predicts 90-day ED or inpatient risk using social vulnerability data alone and generates patient-specific intervention recommendations.

These insights were embedded directly into weekly scheduling workflows, enabling care teams to proactively identify risks such as food insecurity and transportation barriers and connect patients to existing community resources. Supported by multidisciplinary collaboration, the approach shifted SDoH from retrospective reporting to operational decision-making.

As a result, Saint Peter’s achieved a 7.1% absolute reduction in 90-day ED visit rates compared to a matched control group. This session illustrates how SDoH-driven AI, when integrated into clinical operations, can translate health equity strategy into measurable clinical and financial impact that offers a practical blueprint for health systems seeking to reduce disparities and unnecessary utilization.