Healthcare providers are swamped with false alarms from remote patient monitoring systems. Here’s how AI can help solve the problem.
Remote patient monitoring systems have made enormous inroads in healthcare in recent years, especially for chronically ill patient populations. The ability to passively monitor patients for adverse events and warning signs can save lives, but it also creates substantial logistical hurdles for healthcare providers– especially cardiologists, electrophysiologists, and their support staff.
In particular, we’re seeing increased adoption of implantable loop recorders (ILRs) which are implanted underneath the skin of a patient’s chest and then used to detect abnormal rhythms that can be warning signs for stroke. Such devices are now connected and able to transmit data to the ‘cloud’ in order to be reviewed by healthcare professionals. Obviously, the data from these devices must be remotely monitored for alerts. ILRs are designed to be extremely sensitive so that they don’t miss any critical events, but this sensitivity often leads to an unwieldy number of false positives – which can quickly overwhelm healthcare teams.
How We Got Here
Historically, most healthcare data collected was for a specific diagnostic reason or during routine checkups. Clinicians were able to process this volume of data manually considering most patients only produced a few dozen data points per year.
However, over the past few years, there has been a dramatic rise in the amount of passive healthcare data being produced and collected. Today, we’re often generating health-related data not for a diagnostic purpose but in hopes the information might lead to healthier individual choices. If you’re wearing a smartwatch right now, it might be collecting data on your heart rate, steps, sleep, and other metrics. Most of this information doesn’t need to be processed by a healthcare provider because it isn’t meant to raise alarms about potentially disastrous medical events.
Data from cardiac implantable electronic devices (CIEDs) like ILRs are completely different. Like most health-related data collected in the past, it is meant to be actionable. But as with the more mundane health-related data collected by consumer wearable devices, there’s simply too much of it for healthcare providers to handle on their own. And with remote monitoring systems producing such a large number of false alarms, clinicians struggle to keep up with the deluge of data and provide just-in-time interventions to their patients.
Many people don’t realize that remote monitoring of implantable medical devices started out as a way to detect equipment failures – not as a method for diagnosing medical problems. Makers of devices like pacemakers and implantable cardioverter-defibrillators (ICDs) faced huge lawsuits due to devices that unfortunately sometimes failed before they were supposed to. These companies responded by promoting remote monitoring features that could help predict early signs of failure. Connectivity was developed over time for diagnosis devices such as ILR in order to get faster access to relevant events (often asymptomatic), while avoiding unnecessary monthly patient visits.
Prevalence of False Positives
According to industry estimates, false positives make up between 50 and 80 percent of alerts from ILR remote monitoring systems. Implantable devices – as well as most remote monitoring systems – only look at patient data points without applying additional context. A doctor or nurse might see an alert and instantly notice that the reading is within a normal range for a given patient, but most healthcare providers struggle to monitor the influx of alerts that come across for all their patients. As a result, providers need to lean on outside solutions to manage the sheer volume of alerts.
The Promise of AI
The adoption of AI in healthcare is no longer just a promise of the future. A recent study published in the European Heart Journal demonstrated that Implicity’s newly FDA-cleared, novel AI algorithm reduced the number of false-positive episodes by 79% when analyzing ECG recordings from patients implanted with Medtronic ILRs while maintaining 99% sensitivity.
Medical device maker Medtronic also recently announced an AI algorithm to address the false positive issue, but it is only for use with their LINQ II device. Implicity’s solution is compatible with all previous Medtronic models (i.e., Reveal LINQ, Reveal XT, and Reveal DX) that are implanted in the majority of patients.
Interest in leveraging intelligent solutions like those above as instruments for improving efficiency is at an all-time high. The use of ILR devices is increasing and is not expected to slow down. By adopting AI solutions for remote monitoring, clinicians can better serve their patients – and prevent themselves from floundering in a sea of false alarms.