AI in healthcare: the technology is there, the users are not
Since the start of the year, there has been a significant increase in healthcare plans, healthcare providers and analytics companies using AI to change the way healthcare is delivered and the way whose patients can be more involved in their care.
Today, AI applications in healthcare are ubiquitous. Data from my company’s digital health information database, DamoIntelTM, identified a significant increase in the launch of AI use cases in clinical and administrative fields in 2020. An analysis of AI / ML applications deployed by the top 50 healthcare systems in the United States indicates that AI-enabled solutions fall into several technological categories. : machine learning, natural language processing (NLP), conversational interfaces such as chatbots and robotic process automation (RPA). Use cases related to COVID in clinical and administrative fields have contributed to the growth in the adoption of new technologies such as chatbots in healthcare.
A focus on real-time responses with AI-powered point-of-care solutions
The biggest challenge for AI-powered care is providing real-time insight into clinical workflow at the point of care. For example, speech recognition technologies are effective for lower level tasks such as drafting doctor-patient meetings. However, they have yet to evolve into decision support systems that provide additional point-of-care information for diagnostic and treatment decisions.
On the other hand, solutions capable of delivering real-time information have yet to reach large scale and wide adoption. An example is that of Stanford University smart watch based COVID diagnostic app, in partnership with Amazon, which analyzes elevated heart rates and other abnormalities to send real-time alerts to patients with suspected COVID infection. Dr. Michael Snyder, Professor and President of Genetics, is working to scale up the solution with the goal of creating a framework for continuous monitoring of health indicators at the individual level. Its goal is to cover anyone, anywhere, who owns a smartwatch. Amazon offered millions in cloud computing credits for similar diagnostic solutions for digital health innovators around the world.
Data collaborations to generate advanced analyzes in real time
If there’s a new trend this year, it’s data collaborations. Truveta, a consortium of 14 healthcare systems launched in February, aims to pool patient data from all member systems to generate advanced analytics to improve healthcare outcomes. Google has announced a series of partnerships with healthcare companies including Mayo Clinic, Ascension Health and Highmark. Use cases include, but are not limited to, data analysis for quality metrics, benchmarking, and administrative reporting. In addition to its partnership with Google, Mayo Clinic has launched new data collaboration initiatives with AI startups, targeting data from remote monitoring devices. Highmark, a leading Pennsylvania-based health plan, has formed a 10-year partnership with Christiana Care in Delaware to pool medical data and claims for better results. Expect to see more consortia as large payers and providers pool their datasets to improve efficiency through advanced analytical insights.
Other trends that will drive the future of healthcare through AI
- Increased innovation in AI-enabled applications following the CMS Final Rule that allows patients to access and share their medical information with developers looking to create new digital health products and services.
- Hospital rooms of the future that will integrate superior experiences through AI-enabled digital interactions between caregivers, patients and their families. An example is the $ 1.5 billion investment by Penn Medicine in Philadelphia. Titled Pavilion, this 500-bed facility features patient rooms with 75-foot interactive monitors on the walls. John Donohue, VP of Entity Services for Medicine, has been closely involved in aspects of the technology empowerment of the patient room of the future. It references the Disney-inspired user experience design as part of the 6-year project being developed.
- Analyzes from remote monitoring devices. AI-enabled applications that ingest and analyze large amounts of data from home monitoring devices and sensors will lead to the next step in the evolution of healthcare. As healthcare moves from hospital to home, expect heavy investments in analyzing data from remote sensors and monitoring devices. Amazon’s recently launched Amazon Care offering includes home care in addition to virtual care services as part of the overall package. Large healthcare systems such as Kaiser Permanente and Mayo Clinic have also entered the game. They announced investments in Medically Home, a tech company primarily focused on home care.
Is the patient ready now – or is she?
As the technologies and IT infrastructure for AI-enabled care have matured, the adoption of AI-enabled care is driven by the varying levels of readiness of incumbents in the current healthcare ecosystem and concerns about the safety of AI-enabled care, especially for complex care. clinical conditions.
Patients are also unsure of AI-enabled care: a recent study points out that patients find chatbots intrusive and are reluctant to seek the advice of a robot. Administrative use cases of AI-enabled applications can provide a better return on investment in the short term. Sachin Patel, CEO of Apixio, a health analytics company acquired by Centene in 2020, attests to a 4x to 7x return on AI applications in financial operations such as risk adjustments.
My company’s research indicates that more than half of all hospitals across the country continue to use electronic health record (EHR) systems as their primary point-of-care tool. New cloud-based and AI-based solutions continue to face challenges in integrating seamlessly into the clinical workflow at the point of care. Interoperability issues and challenges related to standardization and standardization of health data will continue to be a major challenge for AI-enabled applications. In addition, standards such as CIM, SNOMED and FHIR continue to evolve, representing an ongoing demand for authoritative code change management and data standardization solutions validated by subject matter experts. New and emerging data sources, such as genomic data, will require additional ethical and privacy safeguards before their use in AI applications.
A final concern about AI in healthcare is the lack of visibility into how algorithms are trained to work in healthcare, exacerbated by the systemic bias inherent in many AI applications. Despite advances in AI techniques, algorithms trained on one dataset cannot be easily transferred to another dataset, especially since the role of operational data and social determinants of health in the population health risk assessment increases. As cloud platforms become the dominant data repositories for the development of AI-powered solutions, concerns about protecting data privacy will boost the trust and consent needed to advance the adoption of business tools. ‘IA.
One bright spot for AI in healthcare is the rapid pace of adoption of AI in administrative functions. Healthcare systems leaders need to broaden the reach of these applications to cover new operational areas, including patient access and engagement, to increase efficiency and improve the quality of the experience. Clinical leaders must continue to develop the use of AI applications with caution and focusing on operational areas that do not necessarily seek to replace human intuition and judgment. An example of this is the use of AI to optimize chemotherapy schedules at Penn Medicine.
As healthcare leaders seek to accelerate the adoption of AI, they must also carefully weigh the costs and benefits of the efforts involved in developing and deploying AI solutions. The question always comes back to what we can do with the information we get from AI applications. If we cannot move the needle on the basis of knowledge and information, clinical leaders must question the value of the program and the energy required to produce the knowledge in the first place. The key is to invest in areas where we can see demonstrated results and build them from there. We are still several years away from widespread use of AI in key clinical aspects of healthcare. Until then, we just keep pushing the boundaries.
Copyright © 2021 IDG Communications, Inc.