You’re studying the online version of STAT Well being Tech, our information to how tech is reworking the life sciences. Signal as much as get this article delivered in your inbox each Tuesday and Thursday.
Medtech is chopping jobs, too
We’ve introduced you information practically each week this 12 months about layoffs hitting the well being tech business — and medtech is feeling the financial stress, too. ResMed CEO Mick Farrell lately described the medtech business to me as not “recession-proof”, however “recession-resistant.” In an financial downturn, folks would possibly forgo the newest shopper tech, however they’re not capable of cross on buying a crucial prescription or medical gadget. Nonetheless, the business has definitely handled appreciable losses, compounded by provide chain stoppages.
Due to this, medical gadget firms are chopping an rising variety of workers unfastened. Robotic surgical procedure firm Vicarious Surgical is chopping 14% of its employees to decrease prices because it prepares a 510(ok) submission for its new robotic system. Different medtech firms which have lately laid off workers embrace:
The soar from Amazon to a medical scribe startup
AI A gaggle of startups have emerged as opponents to Nuance’s DAX, all aiming to avoid wasting physicians time through the use of AI to generate notes of doctor-patient conversations. Considered one of these startups is Abridge, which was spun out of the College of Pittsburgh Medical Middle in 2018. Abridge lately employed Zachary Lipton, an AI researcher at Amazon, as its chief scientific officer.
Lipton believes in Abridge’s capacity to have a concrete impression. “[I’m] seeing the second the place the practice’s leaving the station, pondering, ‘Okay, if I don’t soar in proper now, I’ll remorse it,” he instructed my colleague Brittany Trang. Read here to learn more about Lipton’s leap from Amazon to Abridge, his plans for Abridge’s future, and his background as a pc scientist.
Neural community predicts new child well being dangers
A crew of researchers from Stanford and College of California San Francisco have built a predictive model that makes use of digital well being data to calculate the danger of sepsis, cerebral palsy, and different severe situations in newborns. The crew educated a deep studying mannequin on well being data from greater than 30,000 mother-newborn pairs handled within the Stanford well being system, constructing a neural community that might predict 24 totally different well being outcomes. The researchres, who also published an interactive website for readers to discover the community’s knowledge. mentioned the predictions outperformed currently-used danger scores.
The community additionally confirmed researchers beforehand unknown hyperlinks between maternal situations and new child danger — for instance, continual sickness in moms could result in larger danger of newborns creating necrotizing enterocolitis, a severe GI situation. The predictive mannequin “will doubtlessly function an vital useful resource for clinicians and researchers to look at independently,” the researchers wrote, although additional analysis is required to totally consider the mannequin’s accuracy and impression.
FDA digs into AI/ML for precision medication
Tomorrow, the FDA will host a workshop with the College of Maryland on the methods AI and machine studying can be utilized to develop focused medication and medicines. It’s open to the general public, and can talk about each greatest practices and challenges, similar to disparity of knowledge and susceptibility to bias. Audio system from the FDA, Genentech, MIT, Sage Bionetworks, and others will be there. I, for one, will likely be dropping in.
Business ups and downs
What we’re studying
- How does ChatGPT — and its maker — deal with vaccine conspiracies?, Undark
- Amazon has been making an attempt to interrupt into healthcare for years. Right here’s a take a look at every little thing it’s finished, Business Insider
- With $520 million in hand, two entrepreneurs attempt to upend the biotech funding mannequin, STAT