 | ● |  |
|
|
|
| | | Generative AI can help a lot of healthcare workers complete a lot of tasks. But it doesn’t come without pitfalls to avoid. The National Academy of Medicine, aka “NAM,” considers some pertinent ins and outs of the technology in a special publication released this month. The authors of the 15-page report—subtitled “Opportunities and Responsibilities for Transformative Innovation”—pay special attention to GenAI’s emerging role in clinical decision-making, administrative efficiency and patient engagement. The report also offers a side-by-side comparison of GenAI with standard AI—or what the group calls the “predictive/analytical” kind—across five important considerations for adopters. The section also functions as a useful review of similarities and differences: 1. Output evaluation and quality control. - Predictive/analytical AI: These models generate quantitative predictions, “making performance assessment more straightforward through accuracy, precision and recall metrics,” NAM reminds. “The emphasis is on accuracy within defined data parameters rather than subjective quality.”
Generative AI: The primary outputs are new content, such as text, images and audio, “where quality is subjective and context dependent,” the authors note. “Monitoring focuses on coherence, relevance and ensuring ethical content generation, as well as preventing issues like ‘hallucinations’ or factual inaccuracies.”
2. Bias manifestation and detection.- Predictive/analytical AI: Bias checks focus on ensuring that model predictions are fair across different groups, the authors explain. “Monitoring bias in these models often involves fairness audits and statistical checks on outcomes rather than subjective analysis of generated content.”
Generative AI: Bias can appear “in subtle ways, shaping content tone, language or framing,” the authors write. “Monitoring involves detecting biases in generated language or other output media and preventing the spread of misinformation or unintended stereotypes.”
3. Performance degradation and adaptation. - Predictive/analytical AI: Model drift often relates to underlying data shifts, requiring statistical tracking of accuracy and regular retraining. “The process is more data driven and straightforward,” the authors point out, “as performance is measured against historical accuracy benchmarks.”
Generative AI: “Quality degradation may appear as reduced coherence or creativity, requiring frequent content review and adjustments,” NAM explains. “User feedback is often essential in detecting subtle shifts in output quality.” Also, GenAI models are “often designed to evolve over time, learning from new data, so monitoring requires ongoing vigilance to adapt to changes.”
4. Impact on users and society. - Predictive/analytical AI: Impacts are “more directly related to decision making, where inaccurate predictions can affect outcomes in areas like medicine or eligibility for services,” the authors observe. “Monitoring focuses on ensuring reliable decision support and fairness in model applications.”
Generative AI: “The potential for misuse of generative content (e.g., for spreading misinformation) adds a unique layer of impact monitoring, requiring checks on ethical content generation and user satisfaction, NAM states. “Societal impacts include privacy, misinformation and the psychological effect on users.”
5. Compliance and legal considerations. - Predictive/analytical AI: These models often operate in industries with established regulatory frameworks, including healthcare, “so monitoring focuses on meeting interpretability, privacy and compliance requirements within well-defined legal standards,” the authors note.
- Generative AI: “Content produced by generative AI can raise unique compliance issues related to privacy, misinformation and ethical standards,” NAM writes. “Monitoring involves regulatory checks on content generation standards and adherence to ethical guidelines.”
Among the publication’s other noteworthy attributes is a guide for assigning responsibility to individuals according to a 4-point matrix. In order of least to most responsible, NAM’s stakeholder categories are “informed,” “consulted,” “accountable” and “responsible.” Access the publication here. |
| | | |
| |
| | | Are You Prioritizing Clinician Well-Being? In recognition of Health Workforce Well-Being Day, Nabla recently hosted a thought-provoking conversation on what it truly takes to foster a culture of clinician well-being. From going beyond surface-level wellness initiatives to embedding well-being into organizational leadership and strategy, the discussion offered practical, actionable insights for healthcare leaders. Read the key takeaways and watch the full recording: https://www.nabla.com/blog/webinar_clinician_well_being/ |
| | | |
|
| | | Buzzworthy developments of the past few days. - More than half of 4,200+ AI researchers around the world, 57%, believe the technology will improve access to healthcare. Of course, that select slice of the population tends to be more excited about AI than the average person in the street. Still, it’s not nothing that 54% of the surveyed AI scholars expect their favorite technology will deliver more benefits than risks. (A prior survey of U.K. non-expert residents found only 13% on board with that proposition.) The new survey was conducted in the U.K. and is featured in a news item published in Nature. The article summarizes a study yet to be published. Interestingly, the study shows only 21% of AI researchers think artificial general AI is inevitable. By comparison, half of the same 4,200+ don’t believe AGI is even achievable. Nature item here, pre-published study here.
- For their part, physicians are almost evenly split between those who do and don’t believe AI can accurately predict patient outcomes. The sample group is made up of doctors signed up to the global social media platform Sermo. The platform is primarily aimed at physicians and counts 1 million of them among its total base of 1.3 million users. Sermo asked a number of interesting questions of its doctor users. One item asked the docs about their biggest concern regarding AI in healthcare. Some 22% named misuse of, or overreliance on, AI in clinical settings. Sermo comments that medicine “can never be just about data—it requires empathy, ethical reasoning and personalized judgment that AI can’t replicate.” Striking an optimal balance between AI-driven efficiency and human expertise, the company notes, is “key to ensuring patient-centered, high-quality care.” More survey results and analysis here.
- Contrary to popular misgivings, AI may be improving the human touch in healthcare. A healthcare entrepreneur thinks the tide has turned in that direction. He credits ambient AI’s ability to free up physicians for face-to-face time with patients. “This melding of artificial and (human) emotional intelligence—which I call AI+EQ—is unlocking a new golden age in healthcare,” writes Owen Tripp, co-founder of Included Health, in Fast Company. “The very best doctors and nurses have always had a rare mix of experience, expertise, empathy and emotional intelligence. But before AI, that secret sauce was near-impossible to bottle up and scale. That’s exactly what we’re doing now.” Hear him out here.
- Giving ambient AI some love from medicine’s front line is a relieved physician. This one happens to serve his academic medical institution as associate chief of AI and digital health. “We saw that our providers who used Nabla were working less at home,” says Daniel Kortsch, MD, of Denver Health and the University of Colorado. “They had less ‘pajama time.’” That’s a reference to the hours many doctors have logged working at home into the night. “People become doctors not because they want to write notes and fill out paperwork,” Kortsch tells Fox News. “It’s because they want that interaction—and ambient AI gives it to them.”
- Right now, AI is the worst it will ever be. Not because it’s bad but because it’s getting better every day. Given the bright horizons, there’s no time like now for health information management (HIM) professionals to stake a claim in the still-fresh soil. Rose T. Dunn, MBA, notes the opportunity and lays out some ways for HIM people to till, plant and harvest. Forecasting pandemic-like conditions, to name just one example, “has real promise,” Dunn writes at ICD-10 Monitor. “HIM professionals who are involved or wish to be involved in health information exchanges can be on the ground floor of this initiative.” Dunn, chief operating officer with a St. Louis-based healthcare consulting firm, is a past president of the American Health Information Management Association (AHIMA) and recipient of AHIMA’s distinguished member and legacy awards. Get the rest of her AI-leadership tips for HIM professionals here.
- AI has outdone physicians. Again. This time, blinded reviewers rated an algorithm’s recommendations as optimal 77% of the time. Physicians making recommendations for the same symptoms only notched 67% optimal. What’s more, the AI’s outputs were less likely to be rated as potentially harmful than human judgments, 2.8% to 4.6%. The case descriptions concentrated on scenarios common to primary care. Healio has coverage.
- Watch out, da Vinci. There’s a new competitor in your class of surgical robots. It’s Johnson & Johnson’s Ottava. This week word came down that it has successfully completed a gastric bypass operation. The procedure was part of a clinical trial for which Ottava had an investigational device exemption. Along with making the announcement, J&J said it will seek to expand its expected FDA approval to other abdominal surgeries. The da Vinci product line is from market dominator Intuitive Surgical, which has something of a lock on robotic general surgery, as MedTech Dive notes in its coverage of the J&J development.
- AI is helping provider organizations stay on top of accreditation. And finding an appreciative adopter base in Southeast Asia and the broader Asia-Pacific (APAC) region. In Singapore, for example, the Ministry of Health and Health Sciences Authority has issued draft guidelines that stress explainability and data quality to ensure safe AI use in medical settings, TNGlobal reports. “AI systems minimize human errors in documentation and reporting, ensuring accurate and complete data submitted for accreditation,” writes ReHack features editor Zac Amos. “AI [also] enables medical facilities to maintain ongoing compliance rather than preparing for accreditation only when audits are scheduled, making the process less stressful and more predictable.” More here.
- Recent research in the news:
- Notable FDA approvals:
- Funding news of note:
- From AIin.Healthcare’s news partners:
|
| | | |
|
| |
|
|
