10 ways NC health care providers are harnessing AI

Written by Emily Vespa and Michelle Crouch

Co-published with charlotte leisure

Facing a worsening health care worker shortage and health care worker burnout, some health care leaders in North Carolina say they see great promise in artificial intelligence systems.

The state’s health system has been an early adopter of AI tools that shape the patient and provider experience, both in ways that are both obvious and invisible. Across North Carolina, AI is helping healthcare providers predict health risks, communicate with patients, and manage administrative tasks.

Here are 10 ways North Carolina healthcare providers are leveraging AI.

1. AI helps doctors diagnose lung cancer

Some doctors at Atrium Health are leveraging AI for early detection of lung cancer.

When a pulmonologist finds a lung nodule on a scan, they consider factors such as the patient’s medical history to assess the risk of lung cancer. Doctors may recommend a lung biopsy for high-risk patients, but avoid unnecessary procedures if the patient is at low risk.

At Atrium Health Wake Forest Baptist, doctors use AI tools to determine the likelihood that a nodule is cancerous. The tool, called Virtual Nodule Clinic, scores the nodules selected in the scan from 1 to 10. 10 is the highest cancer risk.

Travis Dotson, a pulmonologist at Wake Forest Baptist, said the virtual clinic has helped him and his patients decide next steps when a lung nodule appears to pose a moderate cancer risk. Ta.

“This is another level of support that adds to the clinician’s sentiment and overview of how they feel about the nodule,” Dotson said. “The right thing to do is to be conservative. If the patient is very worried and there is uncertainty about what this nodule is, you can imagine that that is quite difficult for the patient.”

Dotson said the tool recently assigned a high score to a nodule that, by other calculations, had a fairly low risk of cancer. This score helped the patient decide to undergo a biopsy, which revealed that the nodule was indeed malignant.

Although the tool does not replace clinician judgment, Dotson said they have found that patients are highly receptive to the use of AI to enhance clinical decision-making. Wake Forest Baptist will be the first academic medical center in the nation to begin using the technology in March 2023, the institute announced. news release.

2. AI checks surgical patients

At OrthoCarolina in Charlotte, an AI-based digital assistant called Medical Brain follows up with patients to see how they’re recovering after hip or knee replacement surgery.

According to Brian Krenzel, a hip and knee surgeon and OrthoCarolina’s chief quality and development officer, the smartphone app “interacts with patients, asks them questions about their recovery, and provides information about what to expect.” That’s what it means.

Patients can ask questions and get answers right away.

Krenzel said that in the four months since OrthoCarolina began piloting the technology, it has interacted with about 200 patients, averaging 30 to 60 messages per patient.

He said the medical team manually reviews all interactions and the system has not made any mistakes so far. This tool relies on responses that OrthoCarolina pre-creates and loads onto the system. If you don’t know the answer, instruct the patient to call the clinic’s triage line.

Krenzel said Medical Brain has reduced the amount of traditional messages and phone calls the clinic receives after surgery by about 70%. OrthoCarolina is rolling out this technology to all hip and knee patients and adapting it for use in spine surgery patients as well.

3. AI finds dangers in patient X-rays and scans

Across the state, AI is already changing the face of medical screening by providing faster and potentially more accurate results so doctors can focus on the most urgent cases.

For example, in Novant Health’s emergency rooms, AI scans images of patients to identify serious conditions such as neck fractures, brain bleeds, and blood clots, and then treats them, according to Novant Health spokeswoman Caroline Arey. They say they are making sure patients are treated first.

The technology acts as “a second set of eyes for the radiologist,” Alley wrote in an email. She said this would allow the most acute cases to be identified first and alleviate the burnout “often experienced in a race against time to ensure prompt diagnosis and treatment.” Ta.

Another AI platform, called Viz.ai, specifically analyzes CT scans of patients with suspected strokes, looking for particularly dangerous types of blockages in brain arteries, Alley said.

The tool sends results to the stroke specialist’s smartphone within seconds. This is important because millions of brain cells die every minute if a stroke is not treated, increasing the risk of permanent brain damage.

4. AI could draft the messages it receives from doctors.

Like many health systems across the country, Atrium Health and WakeMed turned to AI to help their clinical teams manage the high volume of messages they receive through their patient portals.

Number of messages to doctors since the pandemic skyrocketedadding to clinicians’ already burdened workloads and increasing stress.

WakeMed reduces 12 to 15 patient portal messages per day per provider by using AI to craft answers, filter out unnecessary messages, and route some messages to other staff members Reduced. article Becker’s Hospital Reviews.

At Atrium, AI “drafts the initial response to a patient’s message, which the team then edits before sending,” a spokesperson said in an email.

According to one study, AI-generated messages Performance that outperforms that written by humans They were more than twice as likely to be considered empathetic in terms of understandability and tone. However, the same study also found that AI responses were 38% longer and more likely to use complex language.

critics are worried Relying too much on AI to communicate with patients can lead to nuances of patient concerns being overlooked and the doctor-patient relationship being disrupted.

5. AI helps clinicians decide who will benefit from drugs and treatments

At Wake Forest University School of Medicine, developers have created an AI-based screening tool to identify patients who may have cognitive impairment. The measure, called the Electronic Cognitive Health Index, helps flag patients who might benefit from programs that provide specialized care and medication to people at risk of developing Alzheimer’s disease.

6. AI makes sure you never miss a follow-up

North Carolina health system uses AI to identify patients who are overdue for follow-up visits or routine procedures. Eric Poon, Duke Health’s chief medical information officer, said the system uses AI to identify patients who are overdue for a mammogram.

At Wake Forest Baptist, a virtual nodule clinic flags patients who miss their prescribed follow-up lung scans. Dodson said these scans are “one of the most important aspects of pulmonary nodule management.”

7. AI finds patients with sepsis symptoms

Poon said Duke Health also collects sufficient patient care data through research and daily operations and uses AI to evaluate the data to inform future clinical situations.

For example, Duke Health’s emergency room is inputting data into algorithms trained to detect patients at high risk of developing a life-threatening condition called sepsis, Poon said. The system first introduced a tool called Sepsis Watch to emergency departments in 2018.

Prominent AI sepsis prediction tool built by health software giant Epic in 2021 faced intense scrutiny After the researcher It turns out that the false alarm rate is high.But Poon said Duke Health thoroughly vets AI tools before deploying them to ensure they are safe, effective and fair. According to one report, Duke’s Sepsis Watch reduced the system’s sepsis mortality rate by 31%. news release.

Sepsis Watch was trained on data from more than 42,000 patient encounters. According to Duke. Use information such as patient vital signs and medical history to identify high risk of sepsis and alert rapid response teams.

“The rapid response team will take another look at the chart and discuss with the patient’s primary care team to determine whether the primary care team should consider sepsis as a symptom and whether they have taken all appropriate steps to rule out sepsis. ‘ or an underlying infection,’ Poon said.

The health organization said health care workers used Sepsis Watch to identify more than 3,000 patients with suspected infections. Duke Institute Health Innovation Report.

UNC Health also said it has developed its own sepsis detection model that is more accurate than models built into electronic medical record systems. Presentation shared by system leaders At the 2024 Medical Compliance Association Conference.

8. AI flags patients at high risk for suicide and other health risks

At Novant Health, machine learning systems analyze patient medical record data in real time to identify patients at risk for suicide, a spokesperson said.

The Behavioral Health Acuity Risk model creates a simple, color-coded risk assessment that healthcare providers see when they retrieve a patient’s electronic health record. This makes it easier to act quickly if a patient is at high risk of suicidal thoughts.

The model was built by Novant experts in mental health, emergency medicine, and psychiatry.

Duke Health uses an algorithm to identify oncology patients who are at high risk of needing hospitalization, Poon said. This is “almost invisible to the patient, but it really helps the clinical team be smarter, allowing them to focus on the patients who need the most attention,” he said.

9. AI helps plan operating room schedules

Duke Health uses AI to make operating room scheduling more efficient. An AI model developed by Duke Health researchers and deployed to system hospitals in 2023 predicts how long surgeries will take in the operating room. researchers Found to be 13% accurate than human schedulers.

With operating room costs estimated to range from $22 to $133 per minute, seamless surgical scheduling can reduce costs such as overtime. Discovered in Duke research.

10. AI chatbot guides UNC Health staff to reference tools

UNC Health uses generative AI to help healthcare providers spend less time on administrative tasks. An internal chat bot can answer questions related to UNC Health and direct users to resources in the system’s training library. UNC Health said in a news release:.

“We’re thrilled to be working with UNC Health,” said David McSwain, a pediatric critical care physician at UNC Children’s Hospital and director of UNC Health. “This is just one example of how we can innovatively use this technology to help our teammates spend less time in front of their computers and more time with patients.” Medical information officer, according to a news release.

UNC Health was one of several health systems. Partnering with Epic in 2023 To pilot a new generative AI tool developed with Microsoft.

This article is produced as part of a partnership between The Charlotte Ledger and North Carolina Health News. original healthcare report Mainly in the Charlotte area.

You can support this effort with a tax-deductible donation.