Nurses clad in scrubs and running shoes move through hospital corridors like commuters at Grand Central during rush hour.
But they can’t be everywhere at once.
Add to that staffing shortages — and in the worst-case scenario, another pandemic — and the already daunting task of monitoring patients on the floor becomes impossible for nurses deep in the trenches.
This is a problem Eske Aasvang, a professor and anesthesiologist at Rigshospitalet, a teaching hospital in Copenhagen, Denmark, set out to solve.
Aasvang and his colleague Christian Meyhoff, a fellow professor and anesthesiologist at Copenhagen’s Bispebjerg and Frederiksberg Hospital, collaborated with the Technological University of Denmark to develop the wireless assessment of circulatory and respiratory distress-clinical support system, or WARD-CSS, an around-the clock patient monitoring system that alerts front-line personnel to dangerous deviations in vital signs.
Together, they also founded the company WARD 24/7.
As Aasvang explained to Medtech Insight, complications that arise after surgery often happen outside of ICUs when patients are returned to the general ward and no longer have eyeballs on them 24/7.
In fact, there’s about a 30% chance of adverse events during the first month after surgery, according to a study in The Lancet, with 4.2 million deaths within 30 days of post-op recovery. This accounts for nearly 8% of all deaths worldwide.
Reducing Burnout
As Aasvang noted, nurses in general wards are already responsible for high volumes of patients — “in some cases, one nurse for every 20 patients” — which makes responding to every call impossible, especially when most are false alarms.
This problem also leads to “alarm fatigue,” according to Ashish Khanna, a critical care anesthesiologist at Wake Forest University School of Medicine in North Carolina, who spoke to Medtech Insight about the potential for WARD-CSS to free up nurses by allowing them to focus on genuine emergencies and not chase down every beep and buzz echoing down the halls.
The ability to separate “noise” from genuine alerts, Khanna said, not only enhances patient care, but can cut down on burnout.
“We need systems that will also free up mental space so nurses can focus on the patient and on real problems.”
Ashish Khanna
Khanna also pointed out that there are other duties outside of clinical care, such as clerical tasks, which add to the burden of already taxed hospital staff.
“We need systems that will also free up mental space so nurses can focus on patients and real problems,” Khanna said.
To create such a system, Aasvang said, he and Meyhoff turned to AI with a goal of building a patient monitoring alert system that provided nurses with the same level of sophistication an anesthesiologist would have in the OR.
Powered by AI, WARD-CSS uses wireless sensors to collect and integrate patient data — such as blood pressure, heart rate, respiration rate, and oxygen levels — then sends that data, along with an overview of the patient’s condition, to an app on a nurse’s smartphone with notifications of significant deviations in these measurements.
The system’s AI algorithm interprets the data so that what the nurse receives are alerts showing severe distress, not normal changes.
Routine Fluctuations
As Aasvang explained, without AI modifications, vital sign “alerts” are often benign. For example, fluctuations in blood pressure, heart rate, or respiration are often normal and do not require immediate attention.
This is why, in Aasvang’s view, current early warning systems that use scores to assess patient status are inadequate.
“Consider yourself during the day, you have these fluctuations up and down,” Aasvang said. “The problem is that if these happen while you’re being monitored, there’s an alert going off, because the system detects something is outside the norm, the resting norm, and that results on average of around 200 alerts per patient per day. This renders the system totally useless in the general ward, because how can one nurse monitor all that?”
One of the most important vital signs to monitor, according to Khanna, is respiration rate. However, it’s also one in which fluctuations are common.
WARD’s ability to increase the accuracy of respiration measurements and reduce false alarms, Khanna said, means nurses are more likely to know when a patient is in distress due to an infection, sepsis, or blood loss as opposed to normal stress, such as getting out of bed or walking around.
“This is one of the biggest, hidden culprits that often goes unrecognized and definitely needs more attention,” Khanna said, adding that increasing the accuracy in blood pressure alerts is another potential benefit WARD offers.
Spikes and drops in blood pressure are both common and can happen for various reasons, but both can also signal a heart attack, Khanna said, especially a “silent” heart attack, or myocardial injury.
“These are exceptionally dangerous because patients do not feel chest pain or have typically EKG changes,” he said, “but the risk of death is as much as a typical myocardial infarction.”
WARD’s AI algorithm sorts out benign fluctuations in vital signs, reducing the number of alerts to one that nurses can reasonably handle. For example, low blood pressure or a rapid heartbeat alone may not be an issue of concern. But when presenting together, as Aasvang explained, they could be.
Such fluctuations in data are what the WARD algorithms can analyze and sift through before sending an alert.
Various pilot and clinical studies, according to Aasvang, have shown WARD-CSS to be highly effective in reducing false alerts, and a new pilot, he added, is set to begin at the Cleveland Clinic.
Additionally, the company said the FDA is reviewing an earlier non-AI iteration of the system with approval expected in March. The company said it plans to submit the newer AI version, which builds and expands upon the current system, to the FDA next year.
The status quo for patient monitoring, Aasvang said, is like a pilot only opening his eyes every six or eight hours in the air.
“Maybe he gets lucky and opens his eyes when there’s a mountain or an oncoming plane,” he said. “Well, this is the reality of patients in general wards. We don’t look at them during those six to eight hours except for intermediate checks. That’s not good enough.”