What's Happening?
A recent data brief from the American Hospital Association (AHA) and the ASTP/ONC highlights the growing adoption of predictive AI in hospitals, with 71% of hospitals reporting its use in 2024, up from 66% in 2023. Predictive AI is primarily used to forecast health trajectories or risks for inpatients, but its fastest-growing applications are administrative, such as simplifying billing and scheduling. Despite this growth, a digital divide persists, with small, rural, independent, and government-owned hospitals lagging behind larger, urban, and system-affiliated counterparts. In 2024, 86% of multi-hospital system members used predictive AI, compared to only 37% of independent hospitals. Hospitals source their AI models from electronic health record developers, third-party developers, and self-developed solutions.
Why It's Important?
The increasing use of predictive AI in hospitals signifies a shift towards more data-driven healthcare, potentially improving patient outcomes and operational efficiency. However, the digital divide highlights disparities in healthcare technology access, which could exacerbate existing inequalities in healthcare delivery. Larger hospitals with more resources are better positioned to leverage AI, potentially widening the gap in healthcare quality between urban and rural areas. The focus on administrative applications suggests a prioritization of cost-saving measures, which could impact how resources are allocated within healthcare systems.
What's Next?
As predictive AI becomes more prevalent, hospitals will need to address challenges related to the evaluation and governance of these technologies. The report indicates that most hospitals are already evaluating AI models for accuracy and bias, with a growing number doing so comprehensively. The responsibility for these evaluations is often shared among multiple entities within hospitals, suggesting a need for coordinated efforts to ensure AI tools are used effectively and ethically. The continued development of AI for clinical applications remains a potential area for growth, though it is currently limited by concerns over accuracy and risk.
Beyond the Headlines
The adoption of predictive AI in healthcare raises ethical and legal questions about data privacy, bias in AI algorithms, and the potential for AI to replace human decision-making in critical healthcare scenarios. As hospitals increasingly rely on AI, there will be a need for robust frameworks to ensure these technologies are used responsibly and equitably. The digital divide also underscores the importance of policy interventions to support smaller and rural hospitals in accessing and implementing advanced technologies.
