Predictive Artificial Intelligence (AI) adoption in hospitals has been on the rise, according to data released by the Assistant Secretary for Technology Policy/Office of the National Coordinator for Health IT. However, the data also reveals that small, rural, independent, and critical-access hospitals are lagging behind in adopting this technology.
The analysis shows that while overall use of AI in healthcare delivery increased from 2023 to 2024, there are disparities in adoption rates among different types of hospitals. For example, 86% of hospitals affiliated with health systems reported using predictive AI last year, compared to only 37% of independent facilities. This highlights a persistent digital divide in hospitals’ utilization of predictive AI.
Predictive AI, which leverages machine learning to predict outcomes such as the risk of readmissions, has been utilized in the healthcare sector for years. The adoption of this technology has seen significant growth over the past decade, with 71% of non-federal acute care hospitals integrating predictive AI into their electronic health records last year, up from 66% in 2023.
The use of predictive AI has increased in various healthcare use cases, including simplifying billing procedures, appointment scheduling, and identifying high-risk outpatients for follow-up care. However, the adoption of AI for monitoring health and recommending treatments remains less common, possibly due to the perceived risk of errors.
Despite the benefits of predictive AI, adoption rates vary among hospitals. Critical access hospitals and rural facilities are less likely to use predictive AI compared to non-critical access and urban hospitals. Challenges in adopting AI tools include the need for governance structures, monitoring tools for accuracy and bias, and evaluating their performance post-implementation.
Most hospitals using predictive AI are actively evaluating the tools for accuracy and bias. Task forces, committees, division, and department leaders are commonly involved in the evaluation process. The study also contrasts with other analyses on hospitals’ readiness to implement Generative AI, which creates new original content like texts or images.
In conclusion, while predictive AI adoption is on the rise in healthcare delivery, there are disparities in adoption rates among different types of hospitals. Addressing these disparities and providing support for smaller, rural, and independent facilities to adopt predictive AI can help bridge the digital divide in healthcare delivery.
