AI Health

Friday Roundup

The AI Health Friday Roundup highlights the week’s news and publications related to artificial intelligence, data science, public health, and clinical research.

May 30, 2025

In this week’s Duke AI Health Friday Roundup: scientific innovation: slowing or just fine?; oversight for genAI in medical education; HHS abruptly alters COVID vaccine recommendations; survey study looks at AI use in US health systems; funding changes place severe stress on early-career scientists; persistent decline in self-reported mental health for mothers; HHS secretary suggests barring government scientists from publishing in major journals; much more:

AI, STATISTICS & DATA SCIENCE

Two people are illustrated in a warm, cartoon style, one on the left and one on the right. The person on the left, who has their back to the viewer, and is typing on a laptop which is sitting on a table. They are white, their hair is shoulder length and dark, and they are wearing a green t-shirt. The computer screen is dark with rows of coloured squares representing programming. The person on the right looks similar but their hair is now tied back in a pony tail, and they are wearing a white lab coat and safety goggles. They are reaching down to lift up an orange hazard label which is about the size of a book. The label is an orange square with a black exclamation mark in the middle. The person looks like they are being careful as they lift it. Yasmin Dwiputri & Data Hazards Project / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/
Image credit: Yasmin Dwiputri & Data Hazards Project/Better Images of AI (CC BY 4.0)
  • “The intersection of the role of GenAI in medical education and clinical use hinges on issues of governance and regulation. Currently, GenAI is unlikely to become fully autonomous and unreservedly accepted by the wider medical community and by patients due to issues of trustworthiness and the as-yet unknown impacts on the doctor-patient relationship, despite promised gains in efficiency and personalization of outcomes.” A review article by Tran and colleagues, published in NPJ Digital Medicine, examines governance and regulation of generative AI models used in medical education.
  • “Although our analyses suggest that AI firms are filing patents with limited in vivo evidence, we urge caution with respect to prescription. Caution is warranted given our small sample size, the quickly evolving nature of the field, and our inability to fully assess the causal role of AI use….Nonetheless, the dearth of in vivo experiments in patents that appear to be developed primarily using AI is worrying, even if one believes the AI optimist’s view that the models produce drug candidates more likely to succeed in clinical trials.” A new paper published this week in Science by Janet Freilich and Arti Rai examines the intersection of AI, drug development, and patents.
  • “…MIDAS enables the utility of multiple modalities and metadata, and covers a breadth of benign, premalignant, and malignant lesion types….A key strength of MIDAS is that it offers applicability in the evaluation of models intended for diagnostic settings where a broad range of lesions is expected. These include general dermatology clinics and primary care clinics. A research paper published in NEJM AI by Chiou and colleagues introduces MIDAS, a benchmark dermatology image dataset meant to support rigorous evaluations for AI image-recognition models across different settings and patient populations.
  • “All health system respondents reported adoption activities in Ambient Notes, with over half of them reporting a high degree of success in use of AI for Clinical Documentation, virtually all of which uses generative AI. The remainder of the AI use cases assessed showed variable degrees of adoption, with clinical risk stratification models showing moderate adoption and only 38% of respondents reporting a high degree of success. Similarly, while most organizations have deployed AI in imaging and radiology, only 19% of them report a high degree of success in this area.” A research article published in the Journal of the American Medical Informatics Association by Poon and colleagues describes findings from a survey study of more than 40 health systems that solicited information about their adoption and use of AI in clinical settings.
  • “…pathologists are increasingly under strain. Globally, demand is outstripping supply, and many countries are facing shortages. At the same time, pathologists’ jobs have become more demanding. They now involve not only more and more conventional tasks, such as sectioning and staining tissues, then viewing them under a microscope, but also tests that require extra tools and expertise, such as assays for genes and other molecular markers. For Wang and others, one solution to this growing problem could lie in artificial intelligence…” A Nature technology feature by Diana Kwon explores the potential for using AI-assisted image recognition in pathology – as well as some potential pitfalls.

BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH

Picture of a wadded-up ball of white paper next to a dart-style paper airplane against a dark blue background. Image credit: Matt Ridley/Unsplash
Image credit: Matt Ridley/Unsplash
  • “Since the paper by Funk and his colleagues came out, however, critics have argued that its measure of disruption is flawed and, more broadly, that there’s little consensus on how to measure disruptiveness or novelty in a research paper….Despite this, there does seem to be agreement that groundbreaking innovation is getting harder to achieve — but with many contrasting theories as to why. As the debate continues, researchers are starting experiments to try to better quantify groundbreaking work — and suggesting strategies that could promote it. A Nature news feature by David Matthews addresses the vexed question of whether the scientific enterprise as a whole is seeing less innovation and fewer breakthroughs.
  • “…we found a significant association between adherence to the high DI-GM diet and the reduced risk of GI cancers, with a significant linear dose-response relationship. Additionally, higher genetic risk, as quantified by the CPRS, was strongly associated with increased GI cancer risk. Importantly, we observed a significant additive interaction between DI-GM and genetic risk. These findings underscore the importance of considering both dietary factors that influence gut microbiota and genetic risk in GI cancer prevention strategies. ” A research article published in Nutrition Journal by Li and colleagues examines associations among patterns of diet, gut microbial health, and genetics and the implications for cancer risk.
  • “…we found a substantial decline in the self-reported mental health of mothers of US children from 2016 to 2023. Over this period, the prevalence of self-reported poor or fair mental health increased by 3.5 pp, representing a 63.6% relative increase from the baseline prevalence of 5.5% in 2016. Increases in fair or poor mental health occurred across the socioeconomic spectrum….Downward trends in excellent maternal mental and physical health originated before the COVID-19 pandemic, and less than half of the increase in fair/poor mental health was estimated to coincide with the onset of the public health emergency.” A research article published in JAMA Internal Medicine by Daw and colleagues presents findings from a cross-sectional study evaluating the mental health of female parents with children between the ages of newborns up to 17 years old.

COMMUNICATION, Health Equity & Policy

Photograph shows a closed gate secured by a chain and padlock, with a drive/parking lot and buildings out of focus behind it. Image credit: Masaaki Komori/Unsplash
Image credit: Masaaki Komori/Unsplash
  • “Across New England and the country, thousands of budding scientists have awoken to a stark new reality, one they never could have imagined just six months ago. Funding for laboratories that focus on everything from the genetic causes of aging to cancer is drying up. Jobs in biomedicine are vanishing. Medical schools are rescinding offers of admission and once-thriving scientific internship programs are shutting down for lack of money.” A STAT/Boston Globe article by Chris Serres and J. Emory Parker describes the chaos and uncertainty enveloping early-career scientists and trainees as federal funding for research is frozen or clawed back.
  • “Without Medicaid, low-income persons will avoid seeking medical care. This will result in lost opportunities to prevent disease (eg, hypertension treatment to prevent strokes) and to treat serious diseases (eg, cancer) at an earlier stage. Medicaid is vital to beneficiaries and to the health care ecosystem that we all rely on.” A viewpoint article published in JAMA by Mitchell H. Katz examines the likely fallout for the healthcare enterprise from proposed cuts to the Medicaid program.
  • “Health secretary Robert F. Kennedy Jr. announced Tuesday that he has unilaterally struck the recommendation that healthy children and healthy pregnant people get Covid-19 shots — a move that experts say is unprecedented…Kennedy made the announcement…flanked by Jay Bhattacharya, director of the National Institutes of Health, and Marty Makary, the commissioner of the Food and Drug Administration.” STAT News’ Helen Branswell reports on the abrupt unilateral decision by the DHHS secretary to reverse US health agencies’ recommendations for COVID vaccinations for health children and pregnant persons, a decision that triggered strong pushback from the American Academy of Pediatrics and the American College of Obstetricians and Gynecologists.
  • “The NIH attempting to supplant current medical journals with its own in-house journals represents a historical 180-degree turn. Peer review emerged in the United States after World War II, as scientists tried to prevent the government from exerting too much power over the direction of research.” STAT News’ Anil Oza reports that DHHS Secretary Kennedy, accusing several leading medical journals of “corruption,” suggests that federal agency scientists may be barred from publishing in them, and that the department may develop its own journals as an alternative.