Doctor having headache and tired from work while wearing PPE suit for protect coronavirus disease. … [+]
The United States is projected to experience a shortage of up to 124,000 physicians by 2034 as the demand for physicians outstrips supply. Physician burnout contributes significantly to this challenge, causing many to leave the workforce. One of the most frequently identified reasons for this burnout in the medical field is the growing burden of clerical tasks related to electronic health records that clinicians must handle.
Major tech corporations are turning to generative AI as a primary approach to tackle clinician burnout. However, despite their strong optimism and faith in generative AI’s potential, healthcare CIOs are not plunging in as rapidly as expected.
Clinical Buy-In Falls Short
Clinicians have not yet adopted Generative AI for clinical decisions. Most use cases and press releases reflect the big tech’s efforts to automate administrative tasks.
The American Medical Association plans to formulate principles and recommendations addressing the benefits and pitfalls of generative artificial intelligence in healthcare. Besides creating recommendations, the AMA intends to collaborate with the federal government and other relevant organizations to shield patients from false or misleading AI-generated medical advice. Furthermore, the AMA has initiated and will continue to motivate physicians to educate their patients about AI engagement’s potential benefits and risks.
Jesse Ehrenfeld, president of the American Medical Association was not as hyped on Generative AI for clinical decision-making. He said, “The algorithms are great for solving a textbook patient or a very narrow clinical question, but patients, they’re not a standardized question stem. They’re humans with thoughts, with emotions, with complex medical, social, psychiatric backgrounds. And I’ll tell you, they rarely follow the textbooks.”
Dr. Ehrenfeld believes that generative AI will reduce administrative tasks and considers it a valuable asset.
Privacy And Security Risk
Users of generative AI must address privacy issues, security, intellectual property protection and the evolving legal rules surrounding AI development and use. Currently, we have several lawsuits from comedians, authors and more to come against OpenAI, Meta and other tech companies on copyright infringement. We may have a similar scenario with healthcare data access.
Jake Dorst, chief information and innovation officer at Tahoe Forest Hospital District, is also bullish on generative AI and said, “Generative AI has the potential to revolutionize the healthcare space in several ways. It can aid in medical image analysis, drug discovery, personalized treatment plans, and generating synthetic data for research.” He said, “Data privacy and security concerns are significant, as sensitive patient information needs protection.”
Big tech companies are in the initial phases of voluntary AI commitments, yet healthcare decision-makers must witness substantial progress before fully committing.
What Is Next?
While healthcare CIOs maintain a cautiously optimistic stance on generative AI, it’s worth highlighting the latest announcements from big tech regarding their newest AI innovations. These significant advancements illustrate the rapid pace at which this technology is progressing, despite the challenges and complexities that it presents.
Amazon Web Services (AWS)
AWS announced HealthScribe, a new HIPAA-eligible service that allows healthcare software providers to develop clinical applications. Using speech recognition and generative AI, these applications produce clinical documentation, thereby conserving clinicians’ time.
Healthscribe is built on Amazon Bedrock and currently supports general medicine and orthopedics. Medical coding and transcription companies 3M and ScribeEMR are early adopters of HealthScribe and will partner with AWS to use this service.
Microsoft
Microsoft took an early lead in conversational AI for generating clinical documentation by acquiring Nuance. In April 2022, Microsoft and EHR vendor Epic declared their partnership to incorporate Azure OpenAI Services. Recently, Teledoc, a virtual health provider, announced its intention to partner with Microsoft to integrate OpenAI services, automating the creation of clinical documentation during virtual exams.
Google is actively developing its large language model, Med-PaLM 2, which it anticipates will excel at healthcare discussions over general-purpose algorithms, given its training on questions and answers from medical licensing exams. They are collaborating with Mayo Clinic and other health systems and partnering with the healthcare technology vendor, CareCloud.
Every press release currently focuses on conversational Generative AI as a heralded solution to tackle burnout and automate administrative tasks. However, the question still lingers: is this reality or simply hype?
Punit Soni, CEO and founder of Suki said, “Generative AI is currently going through a hype cycle and getting a lot of attention. In six months’ time, it’s likely that the hype will settle. While gen AI will still have value, it will become clear that for these solutions to truly deliver meaningful impact, they will need to invest in strong foundational elements like EHR interoperability, secure and private infrastructure, an intuitive and flexible user experience that allows users to easily complete their tasks in whatever way works best for them, and an innovative mindset that builds quickly and at the cutting edge of what the tech can do. Without these, the impact of these solutions will be limited to niche use cases.”
While AI may take some time to replace medical professionals, the current focus on alleviating administrative burdens is certainly a step in the right direction.