Patient safety is a top priority for healthcare systems, but traditional video monitoring solutions have fallen short of customer expectations.
The good news is that artificial intelligence and machine learning are revolutionizing the way that organizations monitor and manage patient risk.
During a June webinar hosted by Becker's Hospital Review and sponsored by Caregility, two Caregility experts discussed how advanced augmented video analysis is transforming care for patients and providers:
- Mike Brandofino, president and chief operating officer
- Sushant Mongia, video communications software architect and developer
Five key takeaways:
1. Augmented information technology can help health systems manage patient risk and more. Clinicians are seeking more ways to leverage the eyes and ears in patient rooms. The most urgent need is to identify patients who may be at risk from falls or from visitor actions. Healthcare organizations also want to track the equipment in rooms, monitor activity around patient beds and verify whether rooms are available. Caregility's iObserver solution offers augmented information based on AI and machine learning technologies. "Augmented information captures information automatically and assesses data points to trigger warnings and alerts," Mr. Brandofino said. "Our goal is to support clinical decision-making, not to fully automate it."
2. Advanced video analysis is superior to traditional video monitoring solutions. With traditional patient monitoring systems, virtual sitters monitor a static box that focuses only on the patient bed. Many healthcare systems abandoned these solutions entirely because they generate many false alerts and sitter fatigue. To address these challenges, Caregility developed a machine learning-based application called advanced video analysis (AVA). "AVA leverages real-time video and historical knowledge to identify potential activities that could be a precursor to increased patient risk," Mr. Brandofino said. AVA ingests video feeds from existing wall units, carts and security cameras. Unique algorithms determine the potential severity and scope of actions and then launch the appropriate alerts as needed. The sensitivity of the analysis can be adjusted for each patient.
3. iObserver is designed with security and network performance in mind. Caregility's iObserver solution has three core components: a distributed architecture, security and customer-owned hardware. The distributed architecture includes a bedside unit, on-premise media server and UHE cloud. "We have fully penetration-tested the security system for iObserver's media server, bedside unit and UHE cloud," Mr. Mongia said. Video from the bedside unit can be sent to both the iObserver virtual sitter solution and the media server where the AVA engine resides. Since the media server is deployed on-premise, this minimizes the performance impact to the network.
4. Patient privacy is a top priority. All video content captured by iObserver is de-identified. The AVA engine focuses on patient limbs and all other information is pixelated once it reaches the media server. All faces are blurred and no patient information or audio is captured for analysis.
5. The possibilities for advanced video analysis extend beyond patient safety. Most health systems adopt iObserver to have eyes and ears in every room. Once iObserver has been deployed, however, the potential applications are endless. "We can identify when a patient's posture in bed is appropriate and we can analyze a patient's gait as they ambulate around the room," Mr. Brandofino said. "It's all about the art of the possible."