A unicorn in women’s health & machine learning skepticism – STAT – STAT

A Unicorn In Womens Health Machine Learning Skepticism Stat Stat

You’re reading the web edition of STAT Health Tech, our guide to how tech is transforming the life sciences. Sign up to get this newsletter delivered in your inbox every Tuesday and Thursday. 

A $1 billion valuation in women’s health

The women’s and family health provider Maven Clinic nailed down a $110 million Series D round led by Dragoneer Investment Group and Lux Capital. Normally we wouldn’t spend more than a sentence on such news at a time when nine-figure fundraising rounds are so common. But the deal puts Maven’s valuation over $1 billion in a category of business that until recently has been given short shrift. Erin has the full story.

advertisement

An emerging trend in machine learning: skepticism

This year’s gathering of HIMSS featured its usual mix of case studies and corresponding hype over the use of machine learning in health care. But several speakers also called for more careful review and regulation of models now being used to help diagnose and treat patients. In one session, informatics experts even suggested that a failure to better govern its use may cause another AI winter, in which disillusionment with the technology leads to a loss of interest and investment. Read Casey’s full story here.

advertisement

The problem with opioid misuse algorithms 

A recent story in Wired detailed how a machine learning algorithm to determine the risk of overdose — called NarxCare — has flagged some low-risk patients, leading doctors to deny them necessary prescriptions. A study in JAMIA this week surfaces the opposite problem: the bias that could lead an algorithm to miss patients at high risk of opioid misuse, keeping them from receiving education and treatment options. In an assessment of a natural language processing algorithm trained on data from Loyola University Medical Center, Black patients at high risk of opioid misuse were almost twice as likely to be missed as white patients. Those that fell in the false negative group also had higher mortality risk.

Two hands up if you love VR demos!

A Unicorn In Womens Health Machine Learning Skepticism Stat Stat
Casey tries a VR tool at HIMSS.

This image is proof that exhibitors can get Casey to do anything at HIMSS. In this case, he agreed to take a VR demo guided by Penumbra, which launched a new product called the Real i-Series. It is meant to support care for conditions such as anxiety and depression by immersing users in novel environments, like an underwater dive into a tropical paradise. The company also makes a product that aims to improve upper extremity movement and cognition for patients in a rehabilitation setting —  which Casey is trying here — that received clearance from the FDA last year. As for the i-Series, the next step is to get clinical deployments and build evidence that the product can improve outcomes for patients.

Look who’s raising money

  • The Chinese AI-for-drug-discovery company XtalPi has landed $400 million in a Series D round led by OrbiMed and Hopu Investment The startup, valued at roughly $2 billion, uses its AI to identify molecular candidates to treat illnesses targeted by pharmaceutical companies.
  • Labcorp is acquiring the digital women’s health company Ovia Health, which uses interactive software to support family planning discussions and pregnancy. Financial terms were not disclosed.
  • ClosedLoop.ai has locked down a $34 million Series B round led by Telstra Ventures. The company provides a machine learning platform to help health care companies build software to address a variety of problems.
  • Ultromics, which uses AI to help diagnose cardiovascular disease, raised $33 million in a Series B round led by Blue Venture Fund. The company uses its AI to analyze echocardiograms to assess heart function and identify abnormalities.

A direct attack on misinformation

Despite increasingly stringent policies to cut down on misinformation, falsehoods about vaccines are still circulating widely on social platforms as the country struggles to increase vaccination rates. At HIMSS, health executives from Facebook and YouTube joined a panel to discuss the challenge — but wound up focusing most on the platforms’ efforts to promote accurate information, Katie wrote. “One of the best ways of addressing these issues is actually just helping people get those questions answered directly,” said Kang-Xing Jin, head of health at Facebook. “Just removing this information alone isn’t going to address that need.”

Watch out for the “worried well”

Makers of wearable technology often target healthy consumers with a penchant for data. But sometimes, a new case study illuminates, self-monitoring can create new, harmful health problems. In Cardiovascular Digital Health Journal, a cardiologist at the University of North Carolina at Chapel Hill describes a 70-year-old patient who developed health anxiety after using her smartwatch to keep track of potential heart problems. Over the course of a year, she began to respond to smartwatch notifications by taking more and more electrocardiograms with the watch — 916 total — resulting in 12 visits to the doctor and emergency room.

“Her constant worry and frequent health care visits had a profoundly negative impact on her mental health, relationships, and quality of life,” the authors wrote. “Our patient may represent the tip of an iceberg.”

Movers & shakers

  • Andy Palan joined mental health startup Happify Health as its first CIO, where he will drive the company’s data strategy and engineering initiatives. Palan previously served as CTO at senior home care company AdaptHealth and biotech firm Intarcia Therapeutics.
  • Theator, a company that applies computer vision to the analysis of surgical videos, named Kavi Vyas, most recently at AI-based stroke detection firm Viz.ai, as its new CCO.

What we’re reading