In today’s Duke AI Health Friday Roundup: digital biomarkers for disease surveillance; weak electrical current for countering memory loss; racial disparities in prostate cancer diagnosis persist in affluent neighborhoods; proposing a new approach for managing journal retractions; cybersecurity primer for healthcare; a boom in rare kidney disease research; segregation, redlining, and firearm violence in Baltimore; “touchless” sensing for detecting Parkinson disease; much more:
AI, STATISTICS & DATA SCIENCE
- “The potential exposure of health care infrastructure to cyberattack presents a grave threat to clinical systems and to patient well-being — a threat that is becoming increasingly severe. However, by educating and training clinical staff in cyber-incident response and preparing them to participate actively in countering a cyberattack, hospitals, clinics, and health systems will be able to actively mitigate and reduce the harmful effects of any acute cyber event.” Duke Clinical Research Institute’s Eric Perakslis provides a primer on healthcare cybersecurity in a New England Journal of Medicine
- “Our rigorous model and parameter search uncovered the optimal time periods and aggregate metrics for monitoring continuous digital biomarkers to increase the positivity rate of COVID-19 diagnostic testing. We found that resting heart rate (RHR) features distinguished between COVID-19-positive and -negative cases earlier in the course of the infection than steps features, as early as 10 and 5 days prior to the diagnostic test, respectively.” A paper just published in NPJ Digital Medicine by Shandhi and colleagues describes the use of “digital biomarkers” gathered from wearable technology as part of the CovIdentify study for targeted COVID surveillance testing.
- “…the model can assess [Parkinson disease] in the home setting in a touchless manner, by extracting breathing from radio waves that bounce off a person’s body during sleep. Our study demonstrates the feasibility of objective, noninvasive, at-home assessment of PD, and also provides initial evidence that this AI model may be useful for risk assessment before clinical diagnosis.” An article published in Nature Medicine by Yang and colleagues describes an AI model for detecting Parkinson disease by monitoring nighttime breathing patterns.
- “Learning, rather than occurring intentionally, happens automatically as a side effect of processing. In other words, you don’t decide what is stored in long-term memory. Instead, the architecture determines what is learned based on whatever you do think about. This can yield learning of new facts you are exposed to or new skills that you attempt. It can also yield refinements to existing facts and skills.” An article at The Conversation by Rosenbloom. Lebiere, and Laird makes the case for a multidisciplinary approach to pursuing the goal of a workable model of artificial general intelligence (H/T @MichiganAI).
- “Here we present FedDis to collaboratively train an unsupervised deep convolutional autoencoder on 1,532 healthy magnetic resonance scans from four different institutions, and evaluate its performance in identifying pathologies such as multiple sclerosis, vascular lesions, and low- and high-grade tumours/glioblastoma on a total of 538 volumes from six different institutions.” In article published in Nature Machine Intelligence, Bercea and colleagues describe a collaborative method for training an autoencoder to detect brain anomalies on MRI images from multiple institutions.
BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH
- “Sending weak electrical current into the brain for 20 minutes a day for four days in a row reversed declines in working and long-term memory that come with aging, scientists reported Monday in Nature Neuroscience. The researchers found that the effects lingered even after the electricity was turned off. When they tested subjects a month later, many of the improvements from the brief sessions of brain stimulation remained.” At STAT News, Megan Molteni reports on a study that examined the use of electrical stimulation of the brain to counter age-related memory loss.
- “Nephrology in general poked along, underfunded, as other research fields boomed. A full 50 years after Berger described the unusual condition, IgAN remained relatively unknown and incurable. Most people with the illness have been diagnosed by chance, upon seeing a doctor about an upper respiratory tract infection or finding blood in their urine…” At STAT News, Isabella Cueto reports on new interest in a long-neglected rare disease.
- “As students and parents brace for the start of the school year, many are entering North Carolina classrooms without the protections of required vaccines…During the COVID-19 pandemic, children fell behind on their vaccination schedules, and while some students are catching up, they still have a ways to go, said Zack Moore, state epidemiologist and the epidemiology section chief in the Division of Public Health of the N.C. Department of Health and Human Services.” North Carolina Health News’ Kate Martin reports on the COVID pandemic ripple effects that have seen NC’s childhood vaccination rates slipping in recent years.
- “…although insurance coverage was not a significant predictor of disease severity in this study, Black men were 2 times more likely than White men to be diagnosed at later stage of disease and were 1.8 times more likely to be diagnosed before the age of 60 years. The inverse was true for White men who were 1.8 times more likely to be diagnosed after the age of 70 years. These findings suggest that economic determinants may only be partially responsible for the racial inequities associated with prostate cancer.” A paper published in Research and Reports in Urology by Nemesure and colleagues shows a complex picture of racial disparities in diagnosis of prostate cancer.
- “Poor-quality or fraudulent medicines are causing extensive humanitarian and economic harm, particularly on the African continent. These products are a global problem that requires a coordinated cross-border response, including enhanced regulatory capacities and surveillance, reporting and education.” A commentary by Pyzick and Abubakar in Nature Reviews Disease Primers examines the ongoing toll of degraded or outright fake medications – a toll that disproportionately affects people in African countries.
COMMUNICATION, Health Equity & Policy
- “…the invention of these four ghosts…points to an intriguing development in the author function in science: In some cases the name of the author’s institution has become as important and possibly more important than the name of the author. The name of the author is still there but does not refer to either an author or a person. It begins, instead, to work as a vector, as an intermediary that connects the publication to the name of the institution – not to Loarte’s and Lopes’ actual institutions, but those of the coauthors they have made up.” In an article for Social Studies in Science, Mario Biagioli casts a wide net – one that ranges from Foucault to the Council of Trent – to interrogate recent evolutions in dubious authorship practices.
- “The researchers found that psychologically “inoculating” internet users against lies and conspiracy theories — by pre-emptively showing them videos about the tactics behind misinformation — made people more skeptical of falsehoods afterward, according to an academic paper published in the journal Science Advances on Wednesday. But effective educational tools still may not be enough to reach people with hardened political beliefs, the researchers found.” The New York Times’ Nico Grant and Tiffany Hsu report on recent research into methods of countering online misinformation by “psychological inoculation” against false news.
- “In the first stage, the hope is that the universities would be willing to do more (and more quickly) if the personnel issues were separated from the determination of the validity of the paper.” Retraction Watch interviews Science Editor-in-Chief Holden Thorp on the need for a new approach to retracting papers that would make the trigger for a journal retraction hinge on a determination of the validity of the findings, while determinations about potential misconduct would then be left to universities and institutions to sort out.
- “For many decades, lending institutions redlined low-income neighborhoods, which often had large populations of people of color, classifying them as investment risks. The practice resulted in neighborhoods with lower home values and fewer services and resources. Combined racial and economic segregation, also known as racialized economic segregation, measures current factors, including race and income.” A study by public health researchers at Johns Hopkins reveals associations between excess firearm violence and segregation and redlining practices in Baltimore neighborhoods.