Artificial Intelligence Changing the Delivery of HealthcareDec 30, 2021 11:49AM ● By Susan Meyer
Photo by Jon Pearson
Better, faster, smarter, more efficient—most people want this in every aspect of their lives. Turns out, artificial intelligence is making it happen through facial recognition, Siri, Alexas, voice-to-text, security surveillance, online fraud protection, and sales.
Today, almost every industry is using AI to improve performance and make better business decisions. One of those industries is health care—and the impact is profound.
According to Michael Ash, M.D., executive vice president and chief transformation officer at Nebraska Medicine, AI is making big changes in how health care is delivered in a number of ways, including quality, safety and delivery of care, provider and patient satisfaction, costs, regulatory requirements, research, education, and security.
Artificial intelligence refers to the use of machines to carry out tasks in a way that mimics human learning and thinking. Machine learning is a part of AI that applies the simulation through the use of data to develop algorithms to help predict outcomes and/or find patterns and associations. Also referred to as “smart” systems, machines are programmed or automated to process data to achieve a specific task, such as predicting the risk of certain conditions or diseases. The more data they receive, the smarter and more accurate they become over time, growing their knowledge and capabilities so that they are essentially learning for themselves. By adding a feedback loop, the machine can learn by being told whether its decisions are right or wrong, allowing it to adjust its methodology in the future.
Nebraska Medicine is realizing the critical role AI can play in improving quality and safety of care through the use of clinical decision support models and predictive modeling (using data to develop algorithms to help predict outcomes and/or find patterns and associations).
“These tools are streamlining the clinical process by quickly getting large amounts of information to our providers at the point of decision-making so they can make more informed decisions,” Ash said. “This is huge for doctors, who must constantly sift through large mounds of data when making health care decisions about a patient.”
He explained that one of the big challenges in health care is that, while at times a patient can have symptoms that are predictive of a disease, a lot of times patients have subtle changes that are difficult to detect. Artificial intelligence and predictive modeling can help detect those subtler changes and take proactive measures.
Case in point: sepsis. Nebraska Medicine has been using this tool to lower their sepsis rates—the number one cause of hospitalized deaths worldwide. And the results have been dramatic.
“Sepsis can be difficult to detect and a delay in therapy can be deadly,” Ash said.
“With the use of clinical decision support, we can more effectively identify people at risk for sepsis and begin interventions sooner to prevent it or reduce its severity. This has allowed us to significantly lower our rate of sepsis mortality, propelling us to a ranking that puts us among the top 15% of academic medical centers in the country in terms of safety.”
Early determination of deterioration is another area the hospital is actively working to improve with the help of AI and predictive analytics. This entails identifying which hospitalized patients are at greatest risk for developing life-threatening incidents like arrhythmias, heart attack or pulmonary embolism. Armed with this information, providers can begin preventive interventions earlier to thwart such events.
“We’ve been working on it for several years,” Ash said. “Developing these types of predictive models can take years and years depending on the complexity of the condition. It’s a collaboration of data scientists, physicians, pharmacists, and analytic experts using data and algorithms to find patterns and gain insights so we can develop predictive models. These are then tested repeatedly for safety, accuracy, and making sure interventions are effective.”
Nebraska Medicine recently became one of the first hospitals in the country to adopt the voice recognition technology called Dragon Ambient Experience to transcribe doctors’ notes into electronic health records as they evaluate and visit with patients.
“This technology improves patient care, as it allows the doctor to focus on the patient, instead of a computer screen,” Ash said. “It also saves our doctors an enormous amount of time that they normally would have to spend entering and organizing their notes. This system does it for them, parsing out what’s most clinically relevant. Eventually, the goal is that this technology will also be able to make suggestions for diagnosis and treatments and provide providers with the most cost-effective choices.”
In research, decision support models are helping providers identify and match patients with the most appropriate clinical trials, resulting in a 20% jump in clinical trial participation. In education, AI is helping professors develop smarter tests that do a better job of determining clinical competency in specific areas as well as the development of better teaching models and approaches to improve learning.
Cybersecurity is another area in which AI is making waves. “Instead of waiting for the cyberattack to happen, using predictive modeling, we can look at patterns of activity that might be suggestive of a virus or cyberattack and intercept it much earlier,” Ash said.
The use of AI to drive new and exciting rehabilitative technologies at Omaha’s QLI, an intensive inpatient rehabilitation center for catastrophic injuries, also has been a game changer.
“The rehabilitation process for someone who has sustained a catastrophic injury can take weeks to months to even years,” said Megan Potter, P.T., DPT, coordinator of Physical and Occupational Therapy at QLI. “Over the past five years, the synergy of robotics and AI technologies combined with our clinical expertise is providing us more efficiencies and capabilities, and is allowing people to regain function more quickly and restore quality of life.”
One of those technologies is the Eksobionics exoskeleton, a robotic-powered device designed to help individuals with lower-limb extremity paralysis or weakness due to spinal cord injury, traumatic brain injury, or stroke. The device is worn like a backpack with an attached body suit and helps an individual relearn the physical movement patterns necessary to walk. Artificial intelligence assists by providing predictive information about the person’s step length and step height once they begin walking and then determines how much support they need to maintain optimal gait mechanics. The exoskeleton technology is also helping those who have impairments that cause unpredictable and erratic movements by working to normalize it.
“These technologies have drastically changed the rehabilitation process,” Potter said. “Repetition is the bedrock of getting someone back to independence. More repetitions result in better outcomes such as greater functional ability, and ultimately, independence. Before the exoskeleton and similar rehab technology, gait training was very laborious. It could take up to three therapists plus a body weight support system to hold a person up and help them take a step. With the exoskeleton, the volume of steps we can achieve in gait training can be 500 to 1,000 steps per session compared to 30 steps without it. This allows us to get our patients back on their feet much sooner and is helping reduce length of stays.”
For patients with upper extremity impairments, QLI employs the ArmeoSpring, a sensor-based technology that provides support and resistance to promote motor relearning while also building strength. The device allows patients to complete tasks utilizing a game on its computer screen to target movement in all planes. It encourages movements that simulate everyday tasks like eating, brushing teeth, and putting on clothes, moving them closer toward independence.
Another new technology system called functional electrical stimulation (FES) assists in the recovery process by applying electrical stimulation to a muscle or muscle group to prompt it to move throughout functional movements like cycling or stepping.
“Using traditional electrical stimulation, we could only stimulate two muscle groups at a time,” Potter explained. “Now, with FES, we are able to stimulate up to 12 muscle groups paired to functional movements. By stimulating certain muscles, we can spur neuromuscular reeducation and help someone complete more difficult tasks like walking up the stairs.”
QLI has big plans when it comes to the future of AI and they will be employing these new “futuristic” type technologies in “smart” apartments currently being constructed on the rehabilitation campus.
“These innovative apartments will incorporate smart-home technology and voice recognition applications,” said Scott Meyers, chief technology officer at QLI. “Our clients will also be able to explore the use of technologies like virtual and augmented reality as well as robotics.”
“This new level of care is truly cutting-edge within the sphere of catastrophic injury rehabilitation,” he says.
“The capabilities artificial intelligence can bring to health care are limitless,” Ash said. “We are not even close to realizing its potential. In the future, it could lead to diagnosing conditions we didn’t even know existed before, achieving earlier detection of diseases, identifying causality and correlation between patterns so we can intervene sooner or even delay the onset of certain conditions. It stands to have a very profound impact on health care.”
This article originally appeared in the January 2022 issue of Omaha Magazine. To receive the magazine, click here to subscribe.