Virtual health solutions and other healthcare information technology (HIT) components have been set in place for several years, with little urgency—until 2020. Historical adoption of such solutions was previously somewhat low, with only 10% of patients even utilizing telehealth over a 12-month period, per a J.D. Power 2019 study.
Then, once the pandemic struck, the need for virtual healthcare surged as sheltering in place was ordered across the U.S. The event escalated to create numerous setbacks, changed the state of healthcare, and affected patient care in many ways, primarily staff shortages in acuity and long-term care. “The burnout rate is projected to keep rising into the future. A poll from McKinsey & Company reported about 30% of nurses saying they’re considering leaving direct patient care entirely. Just this month, the American Hospital Association linked staffing shortages to delays in patient discharges. And the strain seen in the in-patient setting is compounding the problem” (Hale, 2023). In fact, a deficit of 15 million clinicians in acute care is anticipated by the end of the 2020s.
The development and spread of AI (artificial intelligence) in healthcare is a key answer to lessening clinician shortage nurse-patient communication hindrances, thus creating better patient outcomes.
Medical organizations have felt the need for supplementation as the aging population increases in the wake (and extension) of a global pandemic that escalated any standing obstacles while also staggering new sets of problems due to supply and personnel shortages and a bourgeoning need for more healthcare. But more and more difficulties are arising to impede quality care and outcomes for patients.
AI (artificial intelligence) in healthcare is a method of applying machine learning (ML) algorithms and houses multiple cognitive technologies in medicine. When computers and machines perform the same cognitive tasks as humans, such as learning, thinking, decision making, and doing—that is AI in healthcare.
Healthcare AI represents tools for use in the improvement of medical care safety. Amid staff shortages, AI can supplement onsite clinicians to prevent adverse events. Major adverse events in healthcare can include patient falls, treatment-related infections, drug events, surgical complications, pressure ulcers, venous thromboembolism, falls, decompensation, and diagnostic errors. These preventable events contribute to one of the top ten causes of death and disability worldwide. AI can reduce adverse events and improve patient outcomes.
Chronic conditions in the U.S. are rampant, with a recent study estimating that approximately 60% of adults in the U.S. have at least one, and around 40% have multiple chronic conditions. There is also a mounting number of Americans with chronic and acute mental illness, with 68% of adults with a mental health condition reporting having at least one medical condition. To top off complications of the interconnectivity of medical and mental health issues, there is a triple increase in medical care non-compliance when mental illnesses and/or addictions are present in patients.
The numerous complications in today’s healthcare setting require changes to how patients are diagnosed and treated while working to contain costs and improve employee engagement, productivity, and retention. The use of different types of AI expanded with the onset of the pandemic when telehealth became vital. TripleTree Industry Perspective Q2 2021 report states, “Virtual health companies experienced unprecedented levels of demand during COVID-19 and have become bigger parts of the overall healthcare system.” Now, a transformation is set to occur. “We see an immediate opportunity for a comprehensive reshaping of care delivery to improve the patient experience, drive quality improvement, and reduce the total cost of care by transforming how patients access and navigate care. The traditional provider ecosystem has the opportunity to customize care delivery by seamlessly leveraging a hybrid of virtual/in-person and real-time/asynchronous care modalities while enabling longitudinal and personalized engagement, communication, and monitoring with technology in the home. In addition, a landscape of next-generation virtual primary care, home-based care, and specialty care platforms are rapidly emerging and transforming the care delivery ecosystem.” 6
Screening For Patient Falls
Experts predicted that aside from supporting clinicians with documentation and clinical decisions, AI would play a major role in patient monitoring. Falls and related injuries were major safety and care concerns. As of 2020, about 36 million adults aged 65 and older reported falling yearly in the United States, with 37% requiring medical attention due to injury.
Current screening methods, such as sensing bed pads, have higher rates of false alarms due to extremely sensitive sensors that cannot distinguish between a patient shifting or attempting to leave their bed. This is where virtual patient observation has become a truly effective tool for deterring patient falls and other numerous adverse events. In a HealthcareITNews interview with VirtuSense founder and CTO Deepak Gaddipati, he stated, “An elderly person that falls has a 70% likelihood of dying due to complications from their fall. Statistically, falls will happen, and 20% of those falls will cost the organization financially through direct care costs, staff hours, quality penalties, and insurance claims.”
Gaddipati said, “I most recently came across the 3rd Annual Optum Survey on AI in Health Care report that stated 83% of healthcare executives already have an AI strategy, and another 15% plan to implement one. Fifty-nine percent of the respondents stated they expect to see tangible cost savings from AI – which is a 90% jump compared to those surveyed in 2018.”
ML (machine learning) is one of the most common AI techniques to deliver clinical risk prediction for improving patient safety. Machine learning has many versions and is a statistical technique for fitting models to data and learning’ by training models with data. A common practice of traditional machine learning is precision medicine—predicting what treatment protocols may best work for a patient. Deep learning—neural network models with levels of variables to predict outcomes—is a type of machine learning that is increasingly used for speech recognition.
ML is also a component of AI that can assist in providing alerts for clinicians and patients utilizing algorithms, enhancing safety protocols in healthcare settings.
According to the National Library of Medicine, NLP (natural language processing) is a field of AI that “includes applications such as speech recognition, text analysis, translation, and other goals related to language. In healthcare, the dominant applications of NLP involve creating, understanding, and classifying clinical documentation and published research. NLP systems can analyze unstructured clinical notes on patients, prepare reports (e.g., radiology examinations), transcribe patient interactions, and conduct conversational AI.”5
The Future of AI in Healthcare Information Technology (HIT)
AI models are predicted to assist in narrowed, personalized care for individual patients in a way that can provide specific treatment to cater to genetic history, background, lifestyle, and environment. Experts say that legislation surrounding healthcare information technology AI will also improve to accommodate the vast amount of data continuing to be created and housed by healthcare organizations and that those models will be developed with guidelines and inaccuracy mechanisms to follow regulations to prevent ethical concerns. AI models are anticipated to soon have the capacity for decreasing and removing discrepancies within health systems. AI is shaping up to be the answer for more efficient and accurate care, secure data, patient record efficiency, detail, and cost-effectiveness. It is estimated that by 2026, AI applications will cut U.S. healthcare costs by $150 billion.