The pandemic has accelerated advances in artificial intelligence (AI) in remote patient care. Physicians are increasingly using digital patient monitoring to better track health data, identify abnormalities, and provide patient-specific treatment—all without the need for face-to-face interaction. Additionally, emergency departments are adopting remote monitoring solutions to get some patients out of the hospital faster. These transformative technologies are delivering better outcomes for patients and reducing healthcare costs.
The State of Artificial Intelligence in Healthcare
AI use cases continue to grow in healthcare as continuous learning and training of algorithms results in smarter technology and improved patient experience.
Most AI applications in healthcare use “augmented intelligence,” which collates the output of algorithms to give clinicians directions on “where to look” as they get their analysis. It also plays an important quality control role in service delivery. Augmented intelligence focuses on the complementary role of technology, designed to augment rather than replace human intelligence.
Consumer electronics companies such as Apple use artificial intelligence to help individuals keep track of their health status. Heart rate-enabled wrist-worn devices can notify users of abnormal heart rates and provide information to share with their doctors. Doctors are also expanding their ability to monitor patients remotely using FDA-approved technology that runs on AI engines. For example, Current Health’s solution provides predictive vital signs monitoring and health deterioration alerts.
In the medical-grade dynamic cardiac monitoring space, many different company AI is being actively deployed for ECG recording and arrhythmia detection. Leveraging AI can improve patient outcomes compared to traditional techniques such as rule-based or traditional machine learning algorithms (used with Holter monitors). Less sophisticated algorithms often fail to provide doctors with a high enough rate of diagnosis to make a definitive diagnosis without repeated monitoring. Devices using artificial intelligence not only bring the promise of personalized medicine closer to reality, but they also further expand the ability of the healthcare system to serve populations in challenging situations, such as remote areas or situations where medical care may not be available.
How AI and ML can be used to monitor cardiac care
In addition to the benefits to patients, AI frees doctors from administrative back-end work, such as screening and managing large datasets, and enables them to focus on applying clinical skills to patient care.
AI can recognize patterns that humans cannot.E.g, The average heart beat is approx. 1.5 million times in two weeks, and the physician may have to determine the six-second time period that determines the clinical outcome. Finding something clinically meaningful can be a needle in a haystack, and AI ensures greater accuracy at scale.
To achieve this level of reliability, providers, data science teams, and AI need clean data and large amounts of data. The expansion of this data requires sophisticated analytical means, which can be achieved through machine learning and deep learning algorithms. Over the past decade, deep learning (a subset of machine learning) has enabled the development of algorithms to match human performance in many fields of science. Unlike more traditional machine learning methods that predict outcomes based on artificial features, deep learning algorithms using artificial neural networks have the advantage of automatically learning relevant features from raw data. Therefore, they use a large number of annotated examples/data and powerful computing power to build complex models that are able to predict correct results on new inputs with very high accuracy.
Both machine learning and deep learning methods require strict FDA oversight and 510(k) clearance before being deployed in the healthcare space. This license indicates that devices using this technology are safe and effective. As the pace of algorithm innovation and the creation of data volumes accelerates, regulators are proposing frameworks to harmonize best practices and regulatory requirements, while allowing for continuous improvement of equipment at a faster pace than in the past.These efforts were released from the FDA AI/ML-based software as a medical device action plan January 2021, followed by Good Machine Learning Practices for Medical Device Development: Guiding Principles This product was developed in conjunction with Health Canada and the UK Medicines and Healthcare products Regulatory Agency (MHRA).
Using advanced algorithms and massive amounts of data, deep learning achieves expert-level, human-level performance in many applications:
The next phase of AI innovation in healthcare
We are just beginning to see what AI can do in healthcare.Last year, the Biden administration created a Artificial Intelligence Task Force Provide easier access to government data and expand access to key resources and educational tools that will continue to spur AI innovation. The bill builds on 2020 legislation and includes a five-year budget of $250 million.
With nationwide attention and funding for AI innovation, the next frontier in AI and wearables will expand the use of predictive capabilities: shifting the paradigm of clinical insights from retrospective reporting to predicting the risk of future conditions . In healthcare, it will be critical to decide which patient groups to monitor and when to monitor them by identifying and analyzing health risks and ensuring that patients receive appropriate preventive medical care.
AI innovations are transforming healthcare delivery. It can enhance the patient experience, reduce the administrative burden on patients, physicians, and care teams, and potentially improve health outcomes. Further investments and technological advancements will undoubtedly revolutionize remote patient care as we know it. The near-term momentum in telecare adoption and AI applications due to the pandemic will continue as healthcare systems continue to evolve to meet current and future challenges.
Photo: Ismagilov, Getty Images
