Nobel in medicine honors discovery of ancient DNA

(Christian Charisius/dpa via AP)

The 2022 Nobel Prize in Physiology or Medicine honors a scientist whose discoveries help us understand what makes us human by learning about our ancestors. Svante Pääbo (above) of the Max Planck Institute for Evolutionary Anthropology was recognized “for his discoveries concerning the genomes of extinct hominins and human evolution.” The Swedish scientist recovered ancient DNA from not just Neanderthals but also Denisovans, a previously unknown hominin — both ancient ancestors of humans who populated the Earth when Homo sapiens migrated out of Africa and whose DNA we carry.

Pääbo “accomplished something seemingly impossible,” and also established a new scientific discipline, paleogenomics, the Nobel Assembly said in its announcement from Stockholm this morning about his recovering DNA from 40,000-year-old bones. Learning about how our genes were passed on and how they diverged help scientists understand how our immune response developed, for example, or how Tibetans survive at high altitude. STAT's Megan Molteni has more.

Exclusive: Epic reworks sepsis algorithm flagged for faulty alarms

One-size-fits-all doesn’t work to predict sepsis, a life-threatening complication of infection. Now Epic Systems is revamping its widely criticized sepsis prediction model to improve its accuracy and make its alerts to clinicians more meaningful by recognizing hospitals have different populations and treatment practices. Corporate documents obtained by STAT show that Epic now recommends its model be trained on a hospital’s own data before clinical use, a major shift to ensure predictions are relevant to the actual patient population a hospital treats.

Epic is also changing its definition of sepsis onset to a more commonly accepted standard and reducing its reliance on antibiotics orders to flag the condition. The changes follow investigations by STAT that found an earlier version of Epic’s tool resulted in high rates of false alarms and failed to reliably flag sepsis in advance. STAT’s Casey Ross has more in this exclusive story.

New data strike cautionary tone on monkeypox vaccine protection

Last week the CDC released heartening preliminary data on how much protection people who have had a single dose of monkeypox vaccine may have. But new data published in JAMA strike a more cautionary tone. Researchers from Howard Brown Health in Chicago reported that 90 people who received monkeypox vaccine — Jynneos, made by Bavarian Nordic — through their clinics from late June to September contracted the virus. Most of those people (77%) tested positive in the first 14 days after their first dose, a time during which their immune responses to the vaccine were still developing. But 21 people contracted the virus after 14 days, eight after 28 days, and two contracted monkeypox three weeks or more after receiving both doses of the vaccine. During the study period, 7,339 people received a first dose at the clinics.

Michael Osterholm, director of the University of Minnesota’s Center for Infectious Disease Research and Policy, said people need to hear the message these data convey. “This is not a 14-days-and-you’re-good shot,” said Osterholm, who wasn’t involved in the research. “It’s clearly a cautionary tale.”

Closer look: Facing threats, children's hospitals scrub info on gender-affirming care

Harassment at Boston Children’s Hospital may be the most prominent example, but right-wing pundits have singled out at least 20 other children’s hospitals on social media for providing gender-affirming care. That’s led hospitals to strip information about those services from their websites, sparking fears that transgender adolescents will find it even harder to get care or find reputable sources of information. All but three of the 20 targeted hospitals have modified what they say about gender-affirming care, often removing information about physicians or the services offered, according to a STAT analysis of their websites.

“It’s very concerning,” said Eli Erlick, co-founder of Trans Student Educational Resources and a trans woman who received hormone therapy from a children’s hospital as a minor. “By the different hospitals responding by removing information, the far-right pundits have already succeeded by unfortunately potentially preventing trans youth from accessing necessary medicine.” STAT’s Tara Bannow and Kate Sheridan have more.

The White House wants to widen access to research we fund. What will that look like?

It sounds logical that taxpayers whose dollars fund scientific research should be able to read scientific papers on the fruits of that labor for free — without the current 12-month delay — but it’s not so simple. In August, the White House’s Office of Science and Technology Policy released a memo directing all federal agencies to form plans for doing that by the end of 2025. In practice, that means would-be readers would have to track new papers down in “agency-designated repositories” like PubMed Central, not on journal websites.

For scholarly publishing it means journals that have traditionally relied on charging for access to new research will have to find new ways to make up lost revenue. STAT’s Brittany Trang dives into the waters of open-access journals, preprint archives, and peer review to consider the questions still swirling before the deadline. Read more.

AI model to predict breast cancer risk from MRI scans might prevent unneeded biopsies

It’s the post-breast scan call that patients dread. Their MRI scan shows something that could be cancerous, so a biopsy is recommended to test the suspicious tissue. A benign biopsy result is a relief, but maybe the procedure wasn’t needed in the first place. A new study in Science Translational Medicine suggests that an artificial intelligence tool can predict the probability of breast cancer in MRI scans as well as a panel of board-certified radiologists; a later analysis showed the AI tool could cut the false-positive rate of breast MRIs, reducing unnecessary biopsies by up to 20% for patients whose MRIs show suspicious lesions.

A caveat: Only 6% the images in the data set used to train the tool came from Black women and only 4% from Asian women, so the model might not perform well in all groups. STAT’s Jayne Williamson-Lee has more.

What we're reading

Once known for vaccine skeptics, Marin now tells them ‘you’re not welcome,’ New York Times
Their loved ones are Ukrainian medics — and Russian prisoners of war, Washington Post
Opinion: Will there be a Supreme assault on public health? STAT
The hidden faces of hunger in America, NPR
Your watch says you have a heart condition. Now what? Wall Street Journal