A multi-site pilot study spanning 28 skilled nursing facilities across six states has demonstrated that AI-powered fall prevention systems can reduce fall incidents by 40% compared to traditional prevention protocols alone. The study, conducted over 12 months and published this week in the Journal of the American Medical Directors Association (JAMDA), represents the largest real-world evaluation of predictive analytics for fall prevention in long-term care settings.
How the Technology Works
The system, developed by health technology company SafeStride AI, combines three components: ambient sensor arrays installed in resident rooms and common areas that detect movement patterns and gait changes, a machine learning platform that integrates sensor data with electronic health record (EHR) information including medications, diagnoses, and recent fall history, and a real-time alert system that notifies nursing staff via mobile devices when a resident’s fall risk exceeds a configurable threshold.
Unlike camera-based monitoring systems that have faced pushback over privacy concerns, the SafeStride system uses radar and pressure sensors that detect motion patterns without capturing images. The AI component learns each resident’s baseline movement patterns and flags deviations—such as increased nighttime restlessness, changes in walking speed, or new patterns of bed exit behavior—that research has shown to be precursors to falls.
Study Results
The 28 pilot facilities were matched with 28 control facilities of similar size, acuity mix, and baseline fall rates. Over the 12-month study period, intervention facilities experienced a 40.3% reduction in total falls and a 52% reduction in falls resulting in injury requiring medical attention. The system generated an average of 4.2 alerts per resident per month, with a clinically validated true positive rate of 73%.
Nursing staff satisfaction surveys showed that 82% of direct care workers found the alert system helpful and not overly intrusive. The most commonly cited benefit was the ability to intervene proactively—for example, assisting a resident to the bathroom when the system detected increased restlessness rather than responding after a fall had already occurred.
Cost-Benefit Analysis
Fall-related injuries are among the most costly adverse events in skilled nursing, with the average fall with injury costing approximately $14,000 in direct medical expenses and liability exposure. The study estimates that participating facilities saved an average of $182,000 per year in avoided fall-related costs, against an annual technology investment of approximately $85,000 per facility. The resulting net savings of $97,000 per facility represents a return on investment that few quality improvement interventions can match.
Adoption Barriers
Despite the promising results, widespread adoption faces several hurdles. The upfront installation cost of $45,000-$65,000 per facility is prohibitive for smaller, independent operators. Integration with existing EHR systems remains technically challenging, particularly for facilities running older software platforms. Some staff and family members have expressed concerns about sensor-based monitoring, even without cameras, underscoring the need for clear communication about data use and privacy protections.
Looking Ahead
CMS has taken notice of the study and is reportedly exploring whether fall prevention technology investment could be incorporated into the SNF Value-Based Purchasing program or qualify for bonus payments under future payment models. Several state Medicaid programs are also piloting technology add-on payments that could offset adoption costs for Medicaid-heavy facilities.