Forge AI Health Friday Roundup - October 1, 2021

October 1, 2021

In today’s Roundup: dissecting Epic’s sepsis algorithm; EUA sought for antiviral COVID therapy; wearables to detect viral infections; trial gets a handle on statin side effects; US leads peers in drop in life expectancy; “colonization” of health equity research; home security robot whiffs with critics; the care and feeding of machine learning systems; vaccine myocarditis paper withdrawn after arithmetic error revealed; the ethical challenges of therapeutic chimeras; rural US suffering disproportionately from COVID deaths; teen’s “side-hustle” advice hobbles popular survey research platform; much more:

Castle-shaped sculpture built out of sugar cubes. Image credit: Tengyart/Unsplash
Image credit: Tengyart/Unsplash


  • “…this article argues that Byzantine literature and culture were more lively than they are given credit for and that—by medieval standards—Byzantine authors were quite open to outside influence. Moreover, it emphasizes that the subsequent use of sugar in Byzantine daily medical practice constituted a significant investment in health, especially bearing in mind the high cost involved in the cultivation, production, and transportation of sugar.”  A spoonful of sugar really does make the medicine go down: A fascinating article published in the journal Speculum by Petros Bouras-Vallianatos traces the diffusion of medieval Arabic medical knowledge into the Latin-speaking world as mediated by Greek-speaking Byzantium – and one of the key traces is the use of sugar in medicinal concoctions.


  • “This cohort study suggests that the use of a noninvasive, wrist-worn wearable device to predict an individual’s response to viral exposure prior to symptoms is feasible. Harnessing this technology would support early interventions to limit presymptomatic spread of viral respiratory infections…” A research article by Grzesiak and colleagues, just published in JAMA Network Open, examines the feasibility of using wearable devices to detect cold and flu infections before they become symptomatic.
  • “Sure enough, when the team investigated over a four-month period, it found clear evidence of the biases flagged by Madland and other Twitter users: The algorithm tended to favor women over men and white people over Black people, including comparisons by gender demographic.” Morning Brew’s Hayden Field investigates Twitter’s decision to open up its image-cropping algorithm to public scrutiny for possible bias.
  • “Using a systematic evaluation by more than 50 expert meteorologists, we show that our generative model ranked first for its accuracy and usefulness in 89% of cases against two competitive methods. When verified quantitatively, these nowcasts are skillful without resorting to blurring. We show that generative nowcasting can provide probabilistic predictions that improve forecast value and support operational utility, and at resolutions and lead times where alternative methods struggle.” A research article published in Nature by Ravuri and colleagues from the DeepMind AI project details the application of a deep generative model to provide accurate “nowcasting” – the ability to predict meteorological events such as rainfall totals in specific geographic locations with only a few hours’ lead time.
  • “Artificial intelligence may seem like some amorphous, all-knowing entity that could outperform humans at even the most complex of tasks. But behind the scenes, humans must spend countless hours cleaning data and teaching these algorithms to ‘think.’” STAT News’ Hyacinth Empinado brings us a really nifty video explainer created by researchers at Emory University, who use a combination of narration and animation to describe the laborious process of training machine learning systems to recognize clinical relevant changes in images.
  • “In marketing materials and internal documents, national electronic health record vendor Epic Systems has touted the ability of its sepsis algorithm to crunch dozens of variables to detect the life-threatening condition in advance, enabling doctors to expedite the delivery of potentially lifesaving antibiotics….But STAT has learned it is using a curious piece of data to make its prediction: whether a doctor has already ordered antibiotics.” STAT News’ Casey Ross returns to an ongoing story examining the real-world performance of predictive software created by electronic health record vendor Epic that has been touted as being able to alert clinicians to the likelihood of sepsis in a patient before the condition would otherwise be apparent.


Empty, desert landscape with mountains in the background; in middle distance, a lone house; in foreground, an apparently abandoned roadside “arrow” marquee sign. Image credit: Dustin Belt/Unsplash
Image credit: Dustin Belt/Unsplash
  • “Rural Americans are dying of covid at more than twice the rate of their urban counterparts — a divide that health experts say is likely to widen as access to medical care shrinks for a population that tends to be older, sicker, heavier, poorer and less vaccinated….Since the pandemic began, about 1 in 434 rural Americans have died of covid, compared with roughly 1 in 513 urban Americans, the institute’s data shows. And though vaccines have reduced overall covid death rates since the winter peak, rural mortality rates are now more than double urban rates — and accelerating quickly.” Kaiser Health News relays a sobering statistic from the Rural Policy Research Institute: COVID is claiming rural lives at twice the rate of those who live in metropolitan areas.
  • “Despite having permanently abandoned statin tablets because of intolerable side effects, most participants could nevertheless complete a 12-month multiple-crossover protocol intended to verify these side effects and identify their origins. These side effects predominantly arose from taking a tablet, rather than from the statin within it.” An innovative study, the SAMSON trial, reported by Howard and colleagues in the Journal of American College of Cardiology, sheds new light on the phenomenon of patients discontinuing statin therapy due to side effects such as muscle aches and weakness.
  • “…soon after the study’s high-profile launch on 24 August, people with autism and some ASD researchers expressed concern that it had gone ahead without meaningfully consulting the autism community.” Nature’s Katherine Sanderson investigates the halting of a high-profile UK autism research project.
  • “…the Covid-19 pandemic saw life expectancy fall across most of Europe and the USA in 2020, on a scale not seen since the World War Two, according to research from Oxford University. And experts say further reductions may be seen in the next year or so, before life expectancy starts to recover.” The BBC reports on a recent study showing that while the COVID pandemic has had a global impact on life expectancies, the harm has not been evenly distributed – and the United States has seen the worst drop in both male and female life expectancy out of 29 peer nations.
  • “Pfizer and BioNTech are another step closer to seeking authorization for young children to receive the COVID-19 coronavirus vaccine, submitting data to the Food and Drug Administration that shows a ‘robust’ antibody response and ‘favorable’ safety outcomes in kids ages 5 to 11 who received the two-dose regimen in clinical trials.” NPR is one of many national outlets this week reporting the news that Pfizer and partner BioNtech have formally submitted data to regulators at FDA from clinical trials of its COVID vaccine in children aged 5-11 years.
  • “…in the best case, a CETP inhibitor might eventually make it as a niche drug, several steps back from the front line, which is not exactly what anyone was picturing 20 years ago. It's hard to even estimate the amount of time, effort, and money that has gone into this field over the years, most of which has gone into showing us that we don't know as much about human lipoproteins as we thought we did. It's been a hard lesson.” At In the Pipeline, Derek Lowe uses a recent article surveying the current (and disappointing) state of play for a once-promising family cholesterol therapies as an example of just how difficult the development of clinical therapeutics can be – even when the underlying hypothesis seems relatively straightforward.
  • “Friday’s results came from a 1,500-patient Phase III trial that randomized newly diagnosed patients to receive the drug or placebo and tracked their outcomes over the next 29 days, although Merck said its independent monitoring board halted the study early after a far smaller sample. Twenty-eight out of 385 patients were hospitalized after receiving molnupiravir, compared to 53 out of 337 patients on placebo….None of the patients who received molnupiravir died, compared to eight on placebo.” Endpoints’ Jason Mast reports that Merck will be filing for an Emergency Use Authorization from the FDA for its oral antiviral drug molnupiravir, which has been undergoing phase 3 testing as a COVID treatment. The data released this week is from an interim analysis of the study.

Small wooden robot sculpture with mismatched features made from screws and washers. Image credit: Nina Mercado/Unsplash
Image credit: Nina Mercado/Unsplash


  • “One person who worked on the robot, which was announced today, reportedly called it a ‘disaster that’s not ready for release,’ while another said it was “terrible,” and the idea that it’d be a useful accessibility device (part of Amazon’s pitch is that Astro can help with elder care) was ‘absurdist nonsense.’ The sources also mentioned that the bot doesn’t handle stairs well…” At The Verge, Mitchell Clark reports on the debut of Amazon’s new home security/surveillance robot, which according to internal communications is not attracting universal admiration, even from some of the people who helped build it.
  • “Health equity researchers say they welcome new interest — and white allies — in their area, which focuses on finding solutions for poorer health outcomes in people from different races, ethnicities, genders, sexual identities, or income levels. But many are troubled by ‘health equity tourists’ — some seen as well-meaning and motivated by their new awareness of racism, others as opportunistic scientific carpetbaggers — parachuting in to ‘discover’ a field that dates back more than a century.” A trenchant piece in STAT News by Usha Lee McFarling examines who has recently been stampeding into the long-neglected field of health equity – and who is being elbowed aside during the inrush.
  • “…overall only about 50% of respondents can correctly identify penicillin as an antibiotic and less than 50% of museum visitors view microbes as beneficial. The results described here suggest that we are perhaps off target with our educational efforts in this area and that a major shift in approach toward more basic microbial topics is warranted in our educational efforts.” A survey study conducted from a kiosk in the American Museum of Natural History and reported in PLOS One by Zichello and colleagues plumbs the public’s knowledge of microbes.
  • “To interpret speech, voice assistants typically convert voice commands into text and compare that text to recognizable words in a database. Many databases historically have not contained reference data collected from those with different speech patterns like slurred sounds and word repetitions. Mr. Rudzicz said that many companies have tried to ‘reach 80 percent of people with 20 percent of the effort,’ using a ‘default voice.’…In other words, companies have rarely prioritized those of us whose speech doesn’t match what engineers assume to be the norm.” A New York Times essay by Char Adams highlights the persistent shortcomings of widely used technologies that ideally should be increasing access for users with disabilities – but aren’t.
  • The Ottawa National Post reports that a preprint article - one that gained international attention when it appeared to show a higher than expected risk of myocarditis associated with COVID vaccination - has been withdrawn after a catastrophic error in ascertaining the study denominator came to light (H/T @RetractionWatch).
  • The Verge’s Rafi Letzter reports on how a teenager’s viral TikTok video on “side hustle” opportunities wound up inadvertently derailing thousands of survey studies that relied on a popular online survey platform: “Prolific, a tool for scientists conducting behavioral research, had no free screening tools in place to make sure that it delivered representative population samples to each study. Suddenly, scientists used to getting a wide mix of subjects for their Prolific studies saw their surveys flooded with responses from young women around Frank’s age….For researchers who rely on representative samples of the US population, that demographic shift was a major problem with no obvious cause and no immediately clear way to fix.” (H/T @RetractionWatch)
  • “…our experiment demonstrates that even if a platform has no partisan bias, the social networks and activities of its users may still create an environment in which unbiased agents end up in echo chambers with constant exposure to partisan, inauthentic, and misleading content. In addition, we observe a net bias whereby the drifters are drawn toward the political Right.” A research article published in Nature Communications by Chen and colleagues describes the use of “neutral bots” to reveal systemic bias in the social media ecosystem.


Two colored (white/purple) rose showing genetic chimerism. Image credit: Raquel Baranow/Wikipedia (CC BY-SA 4.0)
Image credit: Raquel Baranow/Wikipedia (CC BY-SA 4.0)
  • “Scientists have grafted tissues from one species into another for centuries to improve biological understanding. Many people walking around today are able to do so only because their hearts contain valves taken from sheep or cows. But interspecies stem-cell transplants are a new frontier. In some instances, these cells proliferate and integrate with the host species to yield what could be described as new types of animal.” Duke’s Nita Farahany is one of the experts quoted in this Nature article by Liam Drew that explores some of the ethical complexities of developing chimeras – animals genetically engineered to produce human tissues and even organs for transplantation.
  • “…we report on another approach: the creation of an ad hoc public–private consortium that was able to rapidly design, develop and implement a high-throughput diagnostic platform for SARS-CoV-2, enabling testing on a massive scale in the Netherlands. This consortium may provide a model for other countries seeking to rapidly build capacity in diagnostic testing for COVID-19 and for other infectious diseases.” A paper by Krijger and colleagues published in Nature Biotechnology advocates for a public-private partnership approach to developing COVID-19 diagnostic tests.
  • “The likelihood of a prescriber prescribing a drug for an off-label indication may depend on the specific drug and its potential adverse effects. For Vascepa, a side effect of concern is a small increase in the risk of atrial fibrillation, but this is not likely to dissuade a physician from prescribing either the brand-name drug or generic counterpart off-label for cardiovascular prevention, especially if the generic formulation is significantly less expensive than the innovator formulation.” This post by Gregory Curfman at the Harvard Petrie-Flom School’s Bill of Health Blog takes readers down a winding path of clinical research, patent battles, and the relative merits (and longer-term ramifications) of a so-called “skinny label” for a generic competitor.
  • “Eligibility for Medicare at age 65 years, which has been associated with an immediate and substantial reduction in the uninsurance rate and improvements in measures of access to care…was not associated with mortality during the COVID-19 pandemic. These null findings may reflect the influence of state and federal policies that directed payments to hospitals for COVID-19 treatment and eliminated cost sharing for COVID-19 testing…” A research letter by Wallace and colleagues published in JAMA’s Health Forum examined whether Medicare eligibility was associated with differences in mortality for populations on just either side of the eligibility cut-off age.
  • A research article by Ganguli and colleagues published in JAMA Internal Medicine presents findings from a cohort study that sought to better characterize the nature and extent to which so-called low value care – medical interventions that provide little or no benefit to patients but create cost burdens – is being provided at US health care systems.
  • “The FDA’s decision to approve Biogen’s Aduhelm based on the same, less-decisive criteria has sparked intense criticism of the agency from Alzheimer’s experts, scrutiny from members of Congress, and an investigation by the Department of Health and Human Services’ Office of Inspector General….But the accelerated approval of Aduhelm may have also set a precedent that will compel the FDA to review and approve other treatments for Alzheimer’s using the same relaxed standards — further inflaming the debate.” STAT News’ Adam Feuerstein reports that, as some had predicted, the FDA’s approval of the Alzheimer therapy aducanumab (trade name Aduhelm) on the basis of surrogate outcomes (in this case, reduction in amyloid plaque) has led to other drug developers coming forward with therapeutic candidates on the same basis, seeking similar accelerated approval – and the foyer is already starting to look a little crowded.