You’re reading the web edition of Health Tech, STAT’s twice-weekly newsletter about how tech is transforming health care and the life sciences. Sign up here to receive it in your inbox every Tuesday and Thursday.
Algorithms rife with racial bias
Algorithms are churning behind the scenes in health care every day, predicting everything from who will develop diabetes to which patients will fail to show up for appointments. But a new report by researchers at the University of Chicago finds that many of them are perpetuating bias along racial and economic lines, infecting daily decisions made about the care of millions of patients. The researchers published a playbook that calls on providers, payers and other businesses to follow a series of steps to identify and eliminate bias in their tools. A key question is whether a more formal — and independent — audit should be required. Casey has the full story.
Popular sepsis tool misses most cases
Speaking of algorithms, a study in JAMA Internal Medicine finds that one developed by Epic to predict sepsis missed two-thirds of cases in the University of Michigan hospitals and deluged clinicians with false alarms. Epic disputed the validity of the findings, asserting the lackluster performance resulted from poor calibration of the algorithm by the researchers. But Casey writes in a new story that the study points to a broader need to suss out performance problems long before algorithms are put into practice — a step that is often skipped to the detriment of physicians and patients.
IPOs and an accelerator
- Pear Therapeutics announced plans this morning to go public in $1.6 billion SPAC merger with Thimble Point Acquisition Corp., Mario reports. That valuation reflects the progress the company has made moving the market forward for digital therapeutics like its reSET app, as well as tremendous optimism from investors and the company about its future potential.
- The physician social networking company Doximity is sharing the details of an upcoming initial public offering. The company says it expects to fetch $20 to $23 per share, raising more than $500 million on a valuation of about $4 billion.
- Amazon is launching an accelerator program for digital health companies through its cloud services division, and will select 10 startups with operations in the U.S. who want to use the company’s cloud technology. Proposals are due by July 23.
One small step for real world data
New research published at the World Microbe Forum offers a window into how real world data can inform the evaluation of drugs. A poster published by Gilead and research partner Aetion uses de-identified insurance claims data to establish a 21% mortality reduction associated with the use of remdesivir in Covid-19 patients. Gilead, of course, is the maker of remdesivir, so choose how much salt to swallow with that finding. But it nonetheless offers a compelling design for using digital data from the real world to help answer pressing questions about effectiveness.
Scant evidence behind digital biomarkers
Pharma companies and health tech startups alike are investing heavily in the promise of digital biomarkers. But the evidence behind novel measures, captured by wearables and other devices, is still far from solid. “Data scientists are trying to interpret and draw conclusions, and the data is really lacking standards. It’s hard to know what’s trustworthy out there,” Christina Manta, head of applied sciences at HumanFirst, told Katie. HumanFirst, which guides pharma companies in the use of digital devices, is aiming to address some of those gaps through a new research collaboration with the Duke Clinical Research Institute. Read more about their newly launched joint effort, called the Digital Measures Evaluation Center, in Katie’s story.
Dropping into new roles…
- Fresh off its IPO, 23andMe has named Valerie Montgomery Rice, the president and dean of Morehouse School of Medicine, to its board. Also new to the board: Peter Taylor, president of the nonprofit ECMC Foundation.
- Andrew Kasarskis has been tapped to serve as chief data officer of Sema4, a genomics and data analytics company spun out of Mount Sinai Health System in New York City. Kasarskis previously served as Mt. Sinai’s chief data officer.