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.
January 27, 2023
In today’s Duke AI Health Friday Roundup: heart failure outcomes worst for rural Black men; looking forward to the future of clinical trials; stroke risk algorithms perform worse for Black patients than white ones; avian flu spreads at mink farm; drug manufacturing lapses harm young leukemia patients; trial will assess AI for lung cancer risk prediction; the effect of Twitter tumult on scholarly publishing; much more:
AI, STATISTICS & DATA SCIENCE
- “The team of engineers and doctors found that a feed-forward neural network improved peak finger velocity by 45 percent during control of robotic fingers when compared to traditional algorithms not using neural networks. This overturned an assumption that more complex neural networks, like those used in other fields of machine learning, would be needed to achieve this level of performance improvement.” Recent research by University of Michigan engineers suggests that more complex is not always better, at least when it comes to machine systems mimicking human nervous system functions.
- “Massachusetts General Hospital is launching a prospective trial of an artificial intelligence tool designed to predict patients’ risk of lung cancer, a crucial area of inquiry amid a rising incidence of the disease in never-smokers.” An article at STAT News by Casey Ross (subscription required) describes a trial, now getting underway in Massachusetts, that will evaluate the use of artificial intelligence in predicting lung cancer risk.
- “From its earliest times, science has operated by being open and transparent about methods and evidence, regardless of which technology has been in vogue. Researchers should ask themselves how the transparency and trust-worthiness that the process of generating knowledge relies on can be maintained if they or their colleagues use software that works in a fundamentally opaque manner.” An editorial at Nature draws a line in the sand with regard to the use of AI tools such as ChatGPT in the creation of scientific articles.
- OpenAI has announced an extension of its partnership with Microsoft, one that may see increasing incorporation of AI technologies such as GPT, DALL-E, and Copilot into commercial software offerings.
- “In this retrospective study of predictive accuracy that included 62,482 participants, existing stroke-specific risk prediction models and novel machine learning techniques did not significantly improve discriminative accuracy for new-onset stroke compared with the pooled cohort equations. All algorithms exhibited worse discrimination in Black individuals than in White individuals.” A research article by Hong and colleagues, published this week in JAMA, evaluates algorithms for predicting 10-year stroke risk among adults and finds that across the board, they performed worse for Black patients than for white patients.
BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH
- “…rapid translation from bench to bedside is lagging in most areas of medicine and clinical research remains outpaced. The drug development and clinical trial landscape continues to be expensive for all stakeholders, with a very high failure rate….This, combined with the inherent inefficiencies and deficiencies that plague the healthcare system, is leading to a crisis in clinical research.” In a perspective article for Nature Medicine, MD Anderson’s Vivek Subbiah offers a vision for the future of evidence-based medicine amid rapid technological change (H/T @yuriquintana).
- “Our findings provide no support for the increase in self-injury, suicidality and symptoms of EDs [eating disorders] after the lockdowns. Key limitations are differential attrition and varying age in pre- and post-lockdown measures in the longitudinal data.” A research article by Danielsen and colleagues, published in Nature Human Behaviour, examines mental health outcomes in a cohort of young Danish adults following COVID lockdowns (H/T @tylerblack32).
- “When mink at a big farm in Galicia, a region in northwestern Spain, started to die in October 2022, veterinarians initially thought the culprit might be SARS-CoV-2, which has struck mink farms in several other countries. But lab tests soon revealed something scarier: a deadly avian influenza virus named H5N1. Authorities immediately placed workers on the farm under quarantine restrictions.” Science’s Kai Kupferschmidt reports on growing concern over evidence of mammal-to-mammal transmission of highly pathogenic avian flu (H5N1) at a Spanish mink farm.
- “These findings suggest that actions to facilitate access to buprenorphine-based treatment for opioid use disorder during the COVID-19 pandemic were not associated with an increased proportion of overdose deaths involving buprenorphine…”A study published in JAMA by Tanz and colleagues examined outcomes associated expanded access to buprenorphine for the treatment of opioid use disorder during the COVID-19 pandemic.
COMMUNICATION, Health Equity & Policy
- “At the end of the study period, the researchers found that living in rural America was associated with an increased risk of heart failure among both women and Black men, even after adjustment for other cardiovascular risk factors and socioeconomic status. Overall, the risk of heart failure was about 19% higher in rural residents than their urban counterparts. However, Black men living in rural areas had the highest risk of all — a 34% higher risk of heart failure compared to urban-dwelling Black men. A study funded by the National Institutes of Health and published this week in JAMA Cardiology by Turecamo and colleagues finds that heart failure risks are higher in rural areas than in urban ones – and that that risk is most pronounced for Black men living in rural areas.
- “The complexity of GHPs like GitHub is not amenable to conventional Web archiving techniques. Therefore, the growing use of GHPs in scholarly publications points to an urgent and growing need for dedicated efforts to archive their holdings in order to preserve research code and its scholarly ephemera.” This preprint article by Escamilla and colleagues from August of 2022, available at arXiv and accepted for publication at TPDL 2022, documents the inadequacies of current scholarly information infrastructure for coping with the profusion of GitHub data and metadata.
- “What does the bumpy situation at Twitter portend for scholarly communications? For libraries, professionals, and scholars who rely on information exchange, organizations including publishers who rely on its megaphone, the rocky tenure of Twitter’s new CEO and what many see as his open embrace of disinformation may portend a dramatic shift away from what’s been a decade plus-long bulwark of the industry.” Contributors to Scholarly Kitchen offer their takes on the implications of recent upheavals affecting Twitter.
- “Reporting by the Bureau of Investigative Journalism, in partnership with STAT, reveals that at least a dozen brands of asparaginase have been proven to be poor quality, with 10 still on the market globally. The drug stops cancer cells from dividing and growing. Without it, patients face a dramatically reduced chance of survival. But in some cases brands fell well below the standard needed to treat cancer.” An article by Rosa Furneaux and Laura Margottini at STAT News shines a light on a slow-motion disaster for pediatric leukemia patients, who have been falling victim to substandard drugs sold throughout much of the developing world.
