By Dorothy Brooks
The problem with vital sign measurements is those only give a snapshot of how a patient’s condition is at one moment. In a typical hospital, such measurements are taken only every four to six hours. Such measurements also can be inaccurate because of human error.
To circumvent these limitations, investigators at the University of Michigan Weil Institute for Critical Care Research and Innovation developed the Analytic for Hemodynamic Instability (AHI), an artificial intelligence-driven device that can provide continuous monitoring. This way, clinicians can pick up signs of patient deterioration promptly and deploy appropriate resources early.
“[The AHI] uses a single lead from an ECG monitor. Then, using a machine learning algorithm, [the device] analyzes the ECG waveforms,” explains Benjamin Bassin, MD, director of the Michigan Medicine Emergency Critical Care Center (EC3) at University Hospital.
Every two minutes, the AHI produces a report. A green bar indicates the patient is hemodynamically stable. A white bar indicates the patient is unstable. “The value of AHI is to provide continuous monitoring on a patient who would traditionally not have it. The goal is ... to detect hemodynamic instability earlier than you would in a traditional environment,” Bassin says.
Research suggests the AHI may be up to the task. Data captured by the AHI from more than 5,000 patients were compared with continuous heart rate and blood pressure readings measured by invasive arterial monitoring in several ICU units. Investigators found the AHI can detect standard indications of hemodynamic instability, a combination of elevated heart rate and low blood pressure, with nearly 97% sensitivity and 79% specificity.1
Essentially, the AHI is analyzing the autonomic nervous system. Bassin explains the device detects subtle changes that likely will be reflected in vital sign readings. However, the AHI detects these changes sooner, giving clinicians more time to act. “There is also a predictive piece to [the AHI] where it not only tells you what is happening in real time, but also gives you a predictive indicator of what is likely to happen over the next couple of hours,” Bassin adds.
Although there is some benefit to using the AHI in the ICU, these patients already are undergoing plenty of monitoring. “The real value is in places where there are fewer resources, less monitoring, and fewer eyes to identify that patient who is going to be your next rapid response call or your next patient to crash,” Bassin observes.
For instance, clinicians might consider using the AHI on patients who are not in the ICU or the resuscitation bays, but still might complain of chest pain, shortness of breath, or another systemic problem that could lead to significant deterioration. Investigators hope to show the AHI can positively affect resource allocation.
The AHI can be connected not only to a bedside monitor for patients who are on a hospital floor, but also with wireless ECG patches. This enables clinicians to keep tabs on patients who may be sitting in the back of a full waiting room. “The patches could also be used on patients in a less traditional medical environment and still be delivering the same level of monitoring,” adds Bassin, suggesting patches could be worn at home or even on a battlefield.
The AHI is web-based, so clinicians can monitor the reports anywhere there is internet access. Reports also could be viewed on a monitor at the nurse station or on the computer that houses the EMR. Further, clinicians can choose to monitor only those patients for whom they are responsible, from five patients to 20 patients to a whole floor of patients, to multiple floors of a facility, depending on assigned duties.
The AHI does not produce an alarm, but its readouts could be programmed to appear on a tracking board so clinicians would quickly become aware of any patient who is unstable. At press time, Bassin and colleagues started to study the effects of displaying AHI information on a large screen in the EC3, the ED’s nine-bed ICU, where all patients will be displayed at all times so staff can easily monitor their conditions.
The FDA approved the AHI as a software medical device, but it remains in relatively early clinical deployment.2 Still, Bassin notes the tool is used in many different ways by dozens of hospitals across the country. It will take some time to gather outcomes data, but there are some specific cases in which the AHI detected problems that otherwise would have been missed at an early stage.
For instance, Bassin recalls one patient who was about to be sent home, but the AHI signaled the patient was unstable. The physician sent the patient for testing, which revealed a massive pulmonary embolism. In another case, a patient was in recovery following an operation. Everything looked good, including the patient’s vital signs, but the AHI signaled the patient was unstable. It turned out the patient had a large, postoperative hemorrhage. The patient was taken back to the operating room immediately.
Bassin notes there are some limitations to the AHI device. It is only indicated for patients age 18 years and older. Also, the device has not been tested on patients with ventricular assistive devices or those with predominant or sustained arrhythmias.
REFERENCES
1. Schmitzberger FF, Hall AE, Hughes ME, et al. Detection of hemodynamic status using an analytic based on an electrocardiogram lead waveform. Crit Care Explor 2022;4:e0693.
2. Weil Institute. FDA provides clearance to market predictive analytic technology developed at U-M Weil Institute. Feb. 2, 2022.
Researchers developed the Analytic for Hemodynamic Instability, an artificial intelligence-driven device that can provide continuous monitoring. This way, clinicians can pick up signs of patient deterioration promptly and deploy appropriate resources early.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content