Regulations On the horizon: FDA's digital health crackdown
The agency is making good on its promise last year to more closely regulate the software embedded in medical devices — including by publicly censuring heart pump company Abiomed for failing to submit data for approval for a patient monitoring software before selling it, Lizzy Lawrence reports.
The agency issued a warning to the company — bought by Johnson & Johnson last year for $16.6 billion — contending that it should have submitted its technology for approval, and also alleging that Abiomed hasn't been forthright with regulators about other issues with its heart pumps.
It's only the latest example of regulatory tightening: In June, the agency sent a warning letter to heart monitoring company iRhythm for tweaking algorithms purportedly designed to detect heart abnormalities. Taken together with the latest public reprimand, the agency's sending a strong signal that it considers clinical decision software within its regulatory purview, Lizzy writes.
These steps come about a year after the agency established plans to regulate AI — a stance heavily opposed by the medical device industry. Read more from Lizzy.
What else is on FDA's regulatory docket?
Also from the agency, its Center for Devices and Radiological Health plans to issue final guidance on how companies should manage changes in their AI, and draft guidance on premarket submissions and life cycle management for companies selling devices embedded with AI and machine-learning, during the 2024 fiscal year.
Deals and data
Evidation scores contract to detect flu-like illness
San Mateo, Calif.-based Evidation won a $20 million federal contract with the Biomedical Advanced Research and Development Authority to aid some studies on detecting flu-like illnesses including Covid, flu and RSV, the startup told STAT. Typically, Evidation is paid by researchers and pharma companies to help them gather real world data directly from participants, including through wearable devices like Apple Watches or Fitbits. Wearables collect data about wearers' heart rate, sleep and exercise, among other biomarkers.
The contract is part of the federal "Project NextGen," a BARDA and National Institute of Allergy and Infectious Diseases collaboration to develop vaccines, drug treatments and other tools to stave off future public health crises.
Evidation plans to examine how useful machine learning models are in using data from wearable devices to detect respiratory viral infections; eventually the goal is to build models capable to detecting the infections before symptoms manifest. It's the third such contract it's scored from BARDA; the other two were for research examining how useful data from wearables are for flu monitoring, and later, in 2020, how useful that data might be for Covid-19 detection.
A closer look at A16z-backed Waymark's tech in action
Public benefit corporation Waymark, whose investors include Andreessen Horowitz and the New Enterprise Associates, says a partnership with the UnitedHealthcare Community Plan of Washington has produced early data suggesting that predicting Medicaid patients' needs and tapping a network of community-based workers to address them meaningfully reduces non-emergent emergency department visits and in-patient admissions. In a white paper, the health plan said it found a 20% reduction in non-emergent ED visits and a 10% drop in in-patient admissions across about 15,000 Medicaid patients in January of 2023.
A rare Silicon Valley tech company focused on Medicaid, Waymark's secret sauce involves proprietary algorithms drawing on de-identified Medicaid claims to predict patients' needs, and then dispatching community health workers, care coordinators and mental health workers to the people it believes needs the services most.
No comments