Survey: A third of U.S. hospitals are early adopters of genAI

In 2024, 31.5% of U.S. hospitals were early adopters of generative AI, or GenAI, while 43.7% were delayed adopters, according to a new study.

Published in JAMA Network Open, the study examined the adoption of GenAI integrated with the EHR among acute care hospitals. The researchers analyzed data from the 2024 American Hospital Association (AHA) Information Technology (IT) Supplement survey. Representatives from 2,174 U.S. hospitals responded to survey questions on AI.

The survey results show that 31.5% of surveyed hospitals were early adopters of GenAI, which means their organization used a large language model (LLM) integrated with their EHR; 24.7% were fast followers, meaning that their organization planned to use EHR-integrated LLMs in the next year; and 43.7% were delayed adopters, that is, they either reported plans to use LLMs in the next five years, had no plans to use LLMs or were unsure.

Among delayed adopters, 32% reported their hospital did not use or indicate plans to use GenAI, 6.7% indicated plans in the next five years and 5% did not know whether their hospital used GenAI.

The study also shows that fewer respondents indicated their hospital used GenAI compared to predictive AI (52.2%) or other predictive models (54.8%) integrated with the EHR. Notably,  hospital use of predictive AI was strongly associated with use of GenAI. For example, 52.8% of hospitals that used predictive AI were early GenAI adopters, and 72.6% of hospitals that did not use any predictive AI were delayed adopters. 

Additionally, the choice of EHR vendor impacted the use of GenAI among U.S. hospitals. Hospitals that used Epic were 21.9 percentage points more likely than those using Oracle to be early adopters or fast followers.

Not only that, but independent hospitals and critical access hospitals were less likely to be early adopters or fast followers of GenAI, while major teaching hospitals and hospitals with high operating margins were more likely to be early adopters.

Researchers wrote in the study that, "AI adoption may be driven more strongly by the hospital's primary EHR developer and other resources to purchase AI tools rather than their capacity to locally evaluate and monitor AI to ensure its accuracy, fairness, and ongoing effectiveness."

They concluded that combining GenAI adoption with best practices to "effectively evaluate, monitor, and improve generative AI will likely be essential to ensuring long-term value from the technology."

Prior research also indicates that healthcare organizations face a 'readiness gap' with regard to GenAI adoption. According to a survey of 300 healthcare professionals, only 18% knew about formal policies on GenAI use in their organizations, and only one in five were mandated to participate in structured training.

Anuja Vaidya has covered the healthcare industry since 2012. She currently covers the virtual healthcare landscape, including telehealth, remote patient monitoring and digital therapeutics.