NEWS

New AI Tool May Help Patients and Nurses Get Extra Rest at Night

physicians checking patient's vitals in ICU

andresr / Getty Images

Key Takeaways

  • A new AI tool may help doctors decide if their stable patients need to be woken up in the middle of the night for vital sign monitoring.
  • On average, doctors and nurses check a patient’s vital signs every four hours throughout the day and night.
  • Implementing ways to limit the number of times a patient’s vital signs get checked overnight may also help nurses to better manage their time.

There's plenty of evidence supporting the benefits of a good night’s sleep. Some studies suggest it can boost memory function in the brain, increase your energy, keep your immune system strong and even help maintain healthy glowing skin. However, when you’re a patient staying overnight at a hospital, your sleep efforts most likely go right out the window, especially with nurses and doctors checking your vital signs throughout the night. 

One study found that patients get their vital signs collected every four hours, during the day and night—not the best environment for a restful night of sleep.

But researchers at The Feinstein Institutes for Medical Research are hoping to change that routine by using a new artificial intelligence (AI) tool that is intended to predict which patients don’t need to be woken up for vital monitoring. 

“The goal of monitoring vital signs is to best care for the patients and to intervene therapeutically to rescue patients if and when their vital signs become unstable,” Jamie Hirsch, MD, director, Quality Informatics & Data Science at Northwell Health at Northwell Health and co-senior author of the Feinstein study, tells Verywell. “But the frequency of monitoring is also a result of habit and culture. There is no evidence-based standard dictating how frequently patients need to be monitored.”

The study, published on November 13 in Nature Partner Journals Digital Medicine, identifies which patients could safely forgo overnight vital sign monitoring to improve sleep and recovery. 

To prove this type of assessment would work, Hirsch and a team led by co-senior author Theodoros Zanos, PhD, used a predictive model that analyzed 2.13 million patient visits from several Northwell Health hospitals in New York over the course of seven years. The model uses an algorithm that incorporates patient data like respiratory rate, heart rate, systolic blood pressure, body temperature, patient age, etc. to determine which stable patients can safely avoid overnight vital sign checks. It also uses a risk score known as the Modified Early Warning Score (MEWS), which gives doctors a snapshot of how normal the patient’s vitals are overall. 

Zanos says the AI tool was designed to be extremely accurate—as their results seem to show. Out of 10,000 cases, the algorithm only misclassified two patient-nights.

“Even for those few misclassifications, the vital signs during those patients barely deviate from normal,” Zanos, assistant professor at the Feinstein Institutes for Medical Research at Northwell Health, tells VeryWell. “Because of the fact that priority is given on patient safety, the algorithm can suggest to forgo more than half of the patient nights.”

The State of Hospital Sleep Environments

Hospitals are not traditionally known for being cozy sleep havens for people—think of all those bright fluorescent lights and loud, beeping machines. This less-than-tranquil environment is just one of the reasons why nurses and physicians welcome new efforts to improve the quality and quantity of sleep for their hospitalized patients.

Ernest Patti, DO, an emergency medicine physician at St. Barnabas Hospital in New York, tells Verywell that the emergency department in a hospital is actually quite similar to a casino. 

“The lights are always on; there are no windows that you can see from where the patient care is, so you lose your sense of whether it's night or day outside,” Patti says. “Through the night shifts, we're constantly checking people’s vital signs. They can help us determine what the next steps are going to be: will the patient need to go to the OR? Are they going to need another CAT scan or imaging study? Do they need more blood tests? Vital signs also give us an idea if a patient is improving or worsening.”

While sleep is important for general health, it’s even more crucial for those fighting off illness or trying to recoup after surgery.

Theresa Walsh, RN, works nights in the operating room (OR) at the Jersey Shore University Medical Center. She says she can see how limiting the number of times nurses check vital signs at night could benefit patients.

“People who are ill and hospitalized are already dealing with difficulties sleeping," Walsh tells Verywell. "If we eliminate unnecessary disturbance of sleep, I believe we would decrease patient’s anxiety, exhaustion, and perceptions of wellness."

What This Means For You

A hospital room isn’t the first place people think of as a comfortable sleep oasis. With nurses and physicians checking your vital signs every few hours, it’s easy to see how sleep patterns can suffer during important times of recovery. But scientists have now come up with an AI tool that determines which patients are stable enough to skip overnight vital sign checks—meaning in the future, if you or a loved one has to stay overnight at a hospital, you may be able to sleep through the night if this system is utilized in your healthcare system.

A Major Improvement For Staff

Before this new AI tool was developed, Hirsch says most hospitals had very few ways to help improve the sleep of their patients without the use of medications.

“Many hospitals have dimmed hallway lights or encouraged nurses and other staff to maintain a quieter environment. And there have been attempts to shift some diagnostic testing to the daytime period," he says. "But avoiding vital signs altogether has not been tried in a systematic manner.”

Doctors and nurses may also like the fact that an AI tool like this could help healthcare systems maximize their time more efficiently. A 2018 study found nurses spend between 19 and 35% of their time documenting vital signs. As hospitals around the world struggle to keep up with the influx of coronavirus patients, time-saving measures like this could prove to be a viable and critical support tool.

Patti believes this kind of model could also be beneficial as the medical community becomes more immersed in virtual medicine.

“I would still maintain a healthy level of concern for patients that I was worried about, where maybe I would still do the extra vital signs on those smaller number of patients," he says. "But if it was my stable patients, I think I would give [this tool] a try. It seems to be where we’re headed; AI is supposed to help us.”

Zanos and his team are planning for a pilot implementation of the AI tool in the first quarter of 2021, with possibly more rollouts later in the year.

 

Was this page helpful?
7 Sources
Verywell Health uses only high-quality sources, including peer-reviewed studies, to support the facts within our articles. Read our editorial process to learn more about how we fact-check and keep our content accurate, reliable, and trustworthy.
  1. Abel T, Havekes R, Saletin JM, Walker MP. Sleep, plasticity and memory from molecules to whole-brain networks. Curr Biol. 2013 Sep 9;23(17):R774-88. doi:10.1016/j.cub.2013.07.025

  2. Scharf MT, Naidoo N, Zimmerman JE, Pack AI. The energy hypothesis of sleep revisited. Prog Neurobiol. 2008 Nov;86(3):264-80. doi:10.1016/j.pneurobio.2008.08.003. Epub 2008 Sep 3.

  3. Besedovsky L, Lange T, Born J. Sleep and immune functionPflugers Arch. 2012;463(1):121-137. doi:10.1007/s00424-011-1044-0

  4. Chang, J., Garva, R., Pickard, A. et al. Circadian control of the secretory pathway maintains collagen homeostasisNat Cell Biol 22, 74–86 (2020). doi:10.1038/s41556-019-0441-z

  5. Yoder JC, Yuen TC, Churpek MM, Arora VM, Edelson DP. A prospective study of nighttime vital sign monitoring frequency and risk of clinical deteriorationJAMA Intern Med. 2013;173(16):1554-1555. doi:10.1001/jamainternmed.2013.7791

  6. Tóth V, Meytlis M, Barnaby DP, et al. Let Sleeping Patients Lie, avoiding unnecessary overnight vitals monitoring using a clinically based deep-learning modelnpj Digit. Med. 3, 149 (2020). doi:10.1038/s41746-020-00355-7

  7. Collins S, Couture B, Kang MJ, Dykes P, Schnock K, Knaplund C, Chang F, Cato K. Quantifying and visualizing nursing flowsheet documentation burden in acute and critical care. AMIA Annu Symp Proc. 2018 Dec 5;2018:348-357.