In today’s Duke AI Health Friday Roundup: war in Ukraine spills over into cyber realm; data shifts spell trouble for clinical AI; Berkeley loses CRISPR patent battle; differences in neighborhood mobility can affect disease risk; racism (not race) as a risk factor; piping digital notebooks directly into manuscripts; grim news from latest climate report; perfect cryptographic secrecy possible; new analyses point to Wuhan market as point of origin for COVID pandemic; who’s keeping track of your location data?; new lemur makes debut at Duke Lemur Center; much more:
- New lemur just dropped: meet Silas, a Coquerel’s sifaka and the Duke Lemur Center’s newest arrival (note: picture above shows a juvenile Coquerel’s sifaka, but not Silas).
- “The results of their analysis, published in Nature on February 2, point to a distinct and previously unexamined network of ecosystems in rainforest trees. More than 60 percent of the 857 species of flies collected, for example, were found above ground level. Many, if not most, likely are new species, Brown says.” An article by National Geographic’s Natasha Day reports on startling new findings recently published in Nature that reveal “another continent” of hitherto unknown insect species living in the Amazonian rain forest canopy.
AI, STATISTICS & DATA SCIENCE
- “Their experiments lift the perfect secrecy inherent in the quantum realm — unknown and uncertain by nature — into the macroscopic world of the everyday. The technology is still too slow to be practical. Nevertheless, even the most paranoid now know it’s possible to communicate with perfect privacy.” At Quanta, Mordechai Rorvig describes several recent experiments that bear out theories of “perfect” privacy achievable with quantum cryptography.
- “Our results demonstrated that the NeuralCVD score can estimate cardiovascular risk trajectories for primary prevention. NeuralCVD learns the transition of predictive information from genotype to phenotype and identifies individuals with high genetic predisposition before developing a severe clinical phenotype.” A research article published in Lancet Digital Health by Steinfeldt and colleagues details a neural-net-based predictive model used to predict 10-year cardiovascular risk in the UK Biobank’s patient cohort (Python package available at GitHub).
- “A novel investigation by STAT and the Massachusetts Institute of Technology found that subtle shifts in data fed into popular health care algorithms — used to warn caregivers of impending medical crises — can cause their accuracy to plummet over time, raising the prospect AI could do more harm than good in many hospitals.” A feature article by STAT News’ Casey Ross documents an unusual scientific-journalistic collaboration that uncovered the potential dangers of “data drift,” in which the performance of clinical algorithms degrades over time as the systems encounter data different from those they were initially trained upon.
- “…health-care providers will increasingly encounter publications and information based on RWE. Learning to incorporate such evidence into clinical practice and appreciating how to assess such studies will be a non-trivial task for practicing medical professionals moving forward. Here, medical journals and professional societies can take a leading role in educating health-care providers about high-quality evidence generation as well as the specific considerations that are appropriate for understanding how RWD and RWE are used broadly and in the evaluation of digital tools in particular.” A viewpoint article published in Lancet Digital Health by Stern and colleagues reports on findings from roundtable discussions on digital health conducted by the German Ministry of Health and the Digital Medicine Society.
- “It’s not the easiest way to write a paper, Bartholdy concedes. It requires computational know-how and a steep learning curve. And flexibility is needed when collaborating with less tech-savvy co-authors. But many argue that the pay-off is worth the investment.” In a feature at Nature, Jeffrey M. Perkel explores an approach to creating data-intensive manuscripts without having to resort to error-prone cutting and pasting by programmatically inserting data from computational research notebooks.
BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH
- An “atlas of human suffering”: The New York Times’ Brad Plumer and Raymond Zhong report on the sobering, if not downright grim, forecasts contained in the latest UN IPCC report on the state of global climate change: “The report by the Intergovernmental Panel on Climate Change, a body of experts convened by the United Nations, is the most detailed look yet at the threats posed by global warming. It concludes that nations aren’t doing nearly enough to protect cities, farms and coastlines from the hazards that climate change has unleashed so far, such as record droughts and rising seas, let alone from the even greater disasters in store as the planet continues to warm.”
- “Unvaccinated students had eight times the incidence of COVID-19 infection compared to vaccinated students in a North Carolina independent school, according to a study by the ABC Science Collaborative appearing online Feb. 22 in the journal Pediatrics.” An article from the ABC Science Collaborative summarizes recent research by the group that was published in Pediatrics. The work by Thakkar and colleagues examines COVID incidence according to vaccination status among middle- and high-school-aged children at a single school in North Carolina.
- STAT News’ Megan Molteni reports (login required) that UC Berkeley has lost a patent battle with the Broad Institute over the ownership of rights to license CRISPR-CAS-9 gene editing technology.
- “These analyses add weight to original suspicions that the pandemic began at the Huanan Seafood Wholesale Market, which many of the people who were infected earliest with SARS-CoV-2 had visited. The preprints contain genetic analyses of coronavirus samples collected from the market and from people infected in December 2019 and January 2020, as well as geolocation analyses connecting many of the samples to a section of the market where live animals were sold.” Nature’s Amy Maxmen distills findings from a trio of preprints that point to a Wuhan wet market as the most likely point of origin for the COVID pandemic.
COMMUNICATIONS & DIGITAL SOCIETY
- “Within three hours, Microsoft threw itself into the middle of a ground war in Europe — from 5,500 miles away. The threat center, north of Seattle, had been on high alert, and it quickly picked apart the malware, named it ‘FoxBlade’ and notified Ukraine’s top cyberdefense authority. Within three hours, Microsoft’s virus detection systems had been updated to block the code, which erases — ‘wipes’ — data on computers in a network.” A New York Times article by David E. Sanger, Julian E. Barnes and Kate Conger examines Microsoft’s role in countering Russian malware unleashed against Ukrainian government IT assets as part of the larger attack on Ukraine (H/T @eperakslis).
- “Without a doubt, the Russian state has sophisticated cyber capabilities with a track record of havoc….Moscow, however, can also unleash an even more expansive, complex, and often opaque web of proxies whose actors are happy to hack and attack on behalf of the regime. The Kremlin’s involvement with these groups varies and may fluctuate over time; it may finance, endorse, ignore, recruit, or use these actors on an ad hoc basis.” On a similar note, Wired’s Justin Sherman takes a somewhat wider-angled look at the diverse array of cyberwar weapons at Russia’s disposal.
- “It seems perverse to starve institutions meant to serve the public and then judge them on how effective they are at surviving on their limited calories. What if we tried feeding everyone instead?…Public education is infrastructure belonging in the same category as our roads, bridges, utilities, municipal services, parks, libraries and everything else we need to have a society that is shared, inhabitable, nondystopian.” In an article for Inside Higher Ed, John Warner assails the notion of “return on investment” as a yardstick for the worth of a college degree.
- “In the next year, researchers should expect to face a sensitive set of questions whenever they send their papers to journals, and when they review or edit manuscripts. More than 50 publishers representing over 15,000 journals globally are preparing to ask scientists about their race or ethnicity — as well as their gender — in an initiative that’s part of a growing effort to analyse researcher diversity around the world.” At Nature, Holly Else and Jeffrey M. Perkel report on an ambitious program designed to provide a better understanding of diversity –or the lack of it – in scientific journal publishing.
- “Disadvantage in a neighborhood’s mobility network has greater impact than its residents’ socioeconomic characteristics. We also find disparities by neighborhood racial/ethnic composition, which can be explained, in part, by residential and mobility-based disadvantage. Neighborhood conditions measured before a pandemic offer substantial predictive power for subsequent incidence, with mobility-based disadvantage playing an important role.” A research article by Levy and colleagues, recently published in Science Advances, examines how neighborhood-level socioeconomic inequality can be used to predict infection rates of COVID-19.
- “Without legislation limiting the location data trade, Apple and Google have become the de facto regulators for keeping your whereabouts private—through shifts in transparency requirements and crackdowns on certain data brokers.” An article by Alfred Ng and Jon Keegan at The Markup examines the murky marketplace for individual location-tracking data (Duke Technology Policy Lab’s Justin Sherman is quoted).
- “…although there is valid debate about the parameters of race-conscious approaches in health, that debate must be placed into context within a question: What is the alternate approach for redressing racial inequities? If there is agreement that racial inequities are persistent and unacceptable, what is the solution? The onus must be on those arguing against race-conscious strategies for providing that alternative. The status quo for Black, Latino, Asian, Pacific Islander, and Indigenous communities cannot be tolerated.” A JAMA Forum article by Chokshi and colleagues provides a blueprint for understanding racism (and not race itself) as a health risk factor.
- “One of many lessons learned from our review was that health equity was not always a priority in model design, participant recruitment and selection, implementation, or evaluation. As a result, some models have not included numbers of underserved beneficiaries proportional to their presence in the general Medicare population. Further, limited and incomplete sociodemographic data has stymied robust monitoring and evaluation of model outcomes for all populations.” In an article in Health Affairs, author Dora Lynn Hughes from the Centers for Medicare and Medicaid Services announces a new program aimed at improving health equity as part of a strategic initiative for improving the quality of US health care.
- “COVID-19 has not only highlighted significant needs and gaps in the delivery of health and social services but also spurred creativity, innovation, and accelerated collaboration with the diverse stakeholders needed to systematically address these needs. These partnerships will remain especially important as public health leaders increasingly look toward addressing COVID-19 as an endemic and seasonal phenomenon.” An article just published in Health Affairs by Huber and colleagues from the Duke-Margolis Center that highlights the role of schools as a critical element of public-health infrastructure in dealing with the COVID pandemic.
- A white paper jointly produced by the National Academy of Sciences and the National Academy of Medicine urges the Biden administration to re-establish the Presidential Bioethics Commission.