AI Health

Friday Roundup

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

December 3, 2021

In today’s Roundup: federated learning on the Internet of Things; recognizing cells via barcodes; favoritism in scientific publishing; calling for better BIPOC representation in health data; building an evidence base to fight health misinformation; ethical complications for large population genetics datasets; survey indicates growing burnout among scientists; closing the global gap in COVID vaccination; much more:

Deep Breaths

“If a frog sees a skinny object moving parallel to its long axis — like how a worm travels along the ground — it sees dinner. But if a frog sees a similar shape moving perpendicular its long axis — very unlike a worm — it sees a threat to flee from. Scientists call this latter movement the anti-worm stimulus, and it strikes fear into the hearts of frogs.” We’re searching for a Michigan J. Frog joke to accompany this story about the finer points of frog behavior by the New York Times’ Sabrina Imbler, but so far it just won’t quite come together…
“As we explore in more detail in our related article, countries tend to follow a predictable development in forest cover, a U-shaped curve. They lose forests as populations grow and demand for agricultural land and fuel increases, but eventually they reach the so-called ‘forest transition point’ where they begin to regrow more forests than they lose.” All is not lost, especially forests: Our World in Data shows how deforestation, once thought to be a more or less permanent outcome of human activity, can be reversed over time.

AI, STATISTICS & DATA SCIENCE

“In a bid to promote transparency, the Defense Innovation Unit, which awards DoD contracts to companies, has released what it calls “responsible artificial intelligence” guidelines that it will require third-party developers to use when building AI for the military, whether that AI is for an HR system or target recognition.” MIT Technology Review’s Will Douglas Heaven reports on a set of ethics guidelines released by the US Department of Defense that are meant to shape the development of military AI applications.
“Most data being used to address Covid-19, and public health in general, are missing information from BIPOC communities. Despite federal requirements from the Office of Management and Budget, race and ethnicity data are often incompletely collected or misclassified. In 2020, more than half of U.S. health departments did not report data about all racial and ethnic groups. This lack of representation in data is leading to systemic erasure of already vulnerable populations, as well as making it difficult to assess and mitigate the impact of Covid-19 in these communities.” An opinion article at STAT News by Duke’s Warren Kibbe and UNC’s Giselle Corbie-Smith presents a call to action for engaging with communities to improve representation of BIPOC persons in datasets used to inform public health research.
“With the help of FL, the decentralized paradigm in IoFT exploits edge compute resources in order to enable devices to collaboratively extract knowledge and build smart analytics/models while keeping their personal data stored locally. This paradigm shift not only reduces privacy concerns but also sets forth many intrinsic advantages including cost efficiency, diversity, and reduced computation, amongst many others…” A preprint by Kontar and colleagues available at arXiv provides a sweeping overview of the potential for federated learning via the computing resources represented by the “Internet of Things.” (H/T @kdpsinghlab)
“…we must first recognize that the ‘advances’ in AI celebrated over the past decade were not due to fundamental scientific breakthroughs in AI techniques. They were and are primarily the product of significantly concentrated data and compute resources that reside in the hands of a few large tech corporations. Modern AI is fundamentally dependent on corporate resources and business practices, and our increasing reliance on such AI cedes inordinate power over our lives and institutions to a handful of tech firms.” An essay at Interactions by AI expert Meredith Whittaker details the perils of concentrated power in the AI and information technology sectors.

BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH

“Some people are better at fighting off seasonal flu when the strain of influenza virus is similar to the first one they encountered in childhood — a phenomenon evocatively dubbed ‘original antigenic sin’, or OAS. Now, there is increasing evidence that people’s immune responses to COVID-19 could be shaped in a similar way by previous infections with common-cold coronaviruses.” Nature’s Rachel Brazil explores the contentious question of whether prior infection by ubiquitous viral diseases could in some cases confer a protective effect against diseases such as COVID.
“The method works by marking individual cells with a stamp that is passed on to all of a cell’s progeny. Researchers can look at a cell, note its bar code and trace its lineage back to its parents, grandparents, great-grandparents — all the way back to its origins — because each cell that arose from the original bar coded cell has the same stamp.” At the New York Times, Gina Kolata examines the practice of “bar coding” cells to shed light on mutation and disease development.
“In many cases, especially in the late twentieth century, samples have been collected from people (including prisoners) without adequate consent or any record of consent, then shared across research groups or deposited in public databases. In others, participants seem to have given some kind of consent, but it is unclear whether they understood exactly what their DNA would be used for. From two interviews with geneticists, we even learnt that, in some medical studies, various incentives were offered to Roma people — a practice considered unacceptable by most human geneticists.” A comment article at Nature by Lipphardt and colleagues turns a spotlight on ethically dubious genetics research practices that disproportionately impact marginalized minority communities in Europe.
“Although many of the ethical concerns related to incentives and payments in explanatory trials pertain to pragmatic clinical trials, the pragmatic features may introduce additional challenges. These include those related to the risk of incentives and payments undermining the scientific validity and social value of pragmatic clinical trials, the sources of data used in pragmatic clinical trials, and when the pragmatic clinical trials are conducted under waivers of consent.” A new paper published in Clinical Trials by Garland and colleagues explores some ethical complexities that may emerge from incentives designed to encourage participation in pragmatic clinical trials.
“As we prepare for the next pandemic, we need social and behavioral scientists and public health experts to work together to advance the behavioral science of infodemic response. Through the Mercury Project, the Social Science Research Council will fund researchers to build on the existing knowledge base and discover new, evidence-based, data-driven tools, methods, and interventions to counter mis- and disinformation and to support the spread and uptake of accurate health information.” Writing at Health Affairs Blog, Anna Harvey describes the nascent efforts of the Mercury Project, a collaborative effort to employ a behavioral science perspective in building an evidence base related to the causes of and responses to health-related mis- and disinformation.

COMMUNICATIONS & DIGITAL SOCIETY

“Worries over the rise in the acrid tone and harmful and manipulative interactions in some online spaces, and concerns over the role of technology firms in all of this, have spawned efforts by tech activists to try to redesign online spaces in ways that facilitate debate, enhance civility and provide personal security.” A report, the result of a joint effort between the Pew Research Center and Elon University, polls a series of experts to provide a glimpse of what the future may hold for an increasingly problematic public media sphere.
“It is human nature to want a definitive answer to questions. But science is not about certainty, and it seems likely, for many reasons, that we will never have a definite answer about the lab-leak theory. Science can allow us to approach the truth, but always within a range of likelihoods and probabilities — a reality that has been obscured by the need for headlines and social media likes from politicians, journalists, industry players and others…” In an essay for Knowable magazine, Retraction Watch’s Ivan Oransky discusses the “lab leak” hypothesis for the origin of COVID, the problem of conspiracy theories, and the “alignment trap” that can waylay persons on any side of an argument.
“Many clickbait farms today now monetize with both Instant Articles and AdSense, receiving payouts from both companies. And because Facebook’s and YouTube’s algorithms boost whatever is engaging to users, they’ve created an information ecosystem where content that goes viral on one platform will often be recycled on the other to maximize distribution and revenue.” At MIT Technology Review, reporter Karen Hao explores how social media companies find their interests intertwined with the propagation of misinformation, even as they tout recent efforts to curtail its spread.
“For younger generations of men and women, the opening act of their adult lives is likely to follow quite different scripts. It’s a difference likely to echo through their later years, leading today’s young adults to make different romantic choices than earlier generations, to choose different family forms, and to make career decisions that will fundamentally reshape the economy. The rising gender gap in higher education might turn out to be one of the most transformative trends of our time.” An essay by Justin Wolfers at the New York Times examines some of the possible ramifications of a growing trend toward gender imbalance in college campuses and in postgraduate and professional schools.
“…while a majority of journals carried publications distributed across a large number of authors, five percent of journals had a single, highly prolific author that was responsible for at least 11 percent of the published articles in the journal. Furthermore, in a random sample of this subset of journals, the highly published author was a member of the editorial board 61 percent of the time and their papers were accepted in a median time of three weeks after submission…” The Scientist interviews Clara Locher, lead author of a recent publication in PLOS Biology that scrutinizes the extent and effects of favoritism in scientific publishing.

POLICY

“Unprecedented scientific achievements have led to the development of highly effective and safe vaccines, promising therapies, and other critical interventions such as diagnostics, yet the world continues to struggle to enact a coordinated, effective, and equitable response. The widening gap between vaccine haves and have-nots around the world has prolonged the pandemic, worsened global inequity, and risks the emergence of additional variants that could pierce vaccine immunity.” The COVID GAP project (a joint program involving multiple Duke institutes and the COVID Collaborative has published a report that outlines the pressing need for coordinated action against COVID that encompasses a truly global scale
“The most commonly cited barriers were technology, reimbursement for services, and facilitating acceptance of the telehealth among school staff, clinicians, parents, and students.” A qualitative study by Fox and colleagues published in the Journal of School Health evaluates the factors potentially affecting the extension of telehealth services to rural schools (H/T Rural Health Research and Policy Centers).
A perspective article by Sachs and Bagley published in the New England Journal of Medicine addresses the impact that Medicare coverage decisions regarding the controversial Alzheimer therapy Aducanumab could have on state healthcare budgets.
“Less than 60% of respondents to the sixth Salary and Job Satisfaction survey reported being satisfied with their positions. That’s about 10 percentage points less than in earlier satisfaction surveys….As satisfaction wanes, mental health seems to be a growing concern. Forty-two per cent of respondents said they had sought help or wanted to seek help for job-related anxiety or depression, a rise of six percentage points from 2018.” A career feature essay at Nature by Chris Woolston delves into data from a recent survey study that indicates scientists, although generally motivated and enthusiastic about their work, are suffering from alarmingly high levels of exhaustion and burnout.