A market failure is crushing health AI innovation: the most broadly used AI products are mediocre in quality and healthcare organizations are largely unable to assess the ROI of their purchases.
Mid-market and low-market health systems don’t have data scientists or data engineers to evaluate AI solutions prior to implementation or to perform post-implementation monitoring. On the other hand, a handful of mature organizations have sophisticated internal capabilities to develop, locally evaluate, integrate, and monitor growing portfolios of AI solutions. But these “best in class” AI solutions used within mature organizations are either not commercially available or struggle to penetrate the market as point-solutions.
How can we begin to eliminate this digital divide? At Health AI Partnership, we bring together a network of healthcare organizations that represent the most mature AI organizations in the United States. These sites contribute to best practices, share their learnings through workshops and interviews, and serve as mentors to community and rural settings who are navigating the AI product lifecycle. We also bring together Practice Network sites that represent community health centers and community hospitals seeking to utilize off-the-shelf commercial AI solutions.
Health AI Partnerships is halfway through a demonstration project of AI technical assistance serving four federally qualified health centers (FQHCs) and a community hospital. Within this project, we try to:
- Break down silos of expertise to equip community and rural settings to make the best use of AI.
- Hold vendors and health IT suppliers accountable to serve the best interests of safety net organizations and not to exploit information asymmetries.
- Discourage purchase and use of AI products whose performance diverges significantly from vendor claims.
With over 1,600 FQHCs and 6,000 hospitals in the United States, our efforts at Health AI Partnership can have a huge impact on the healthcare delivery system in the U.S. We do our best to raise awareness of the dire situation in rural FQHCs, but our preliminary efforts are a drop in the bucket.
In the past, the federal government coordinated step-function changes in the adoption of emerging technologies in healthcare to support the private sector. For example:
- HITECH not only funded electronic health record adoption, but also funded regional extension centers across the country to provide technical assistance for community and rural settings.
- HRSA funds telehealth resource centers and telehealth centers of excellence to support telemedicine use in low-resource settings.
We need immediate action at the federal level to scale technical assistance for AI use in healthcare. All healthcare organizations should have access to the “best in class” AI solutions and should be able to ensure optimal ROI from AI purchases. If the federal government fails to act now, we are likely to witness significant waste and inefficiencies as investments in AI increase dramatically and few realize a ROI.
Connect with Mark here.
