por
Gus Iversen, Editor in Chief | July 06, 2026
A study published in the American Journal of Neuroradiology found that hospitals' adoption of artificial intelligence software for detecting large vessel occlusion in acute ischemic stroke was driven largely by institutional resources and infrastructure rather than patient characteristics.
Researchers from the Harvey L. Neiman Health Policy Institute and Northwell Health analyzed 2,116 Medicare inpatient stroke episodes treated at 1,076 hospitals between October 2020 and December 2023. The study examined use of AI software billed through Medicare's New Technology Add-On Payment (NTAP) program, which provided temporary supplemental reimbursement for qualifying technologies.
Overall, NTAP-billed AI was used in 14.8% of eligible stroke cases during the study period. Adoption increased each year after the reimbursement pathway became available, reaching 21% of eligible cases in 2022 before declining in 2023 as the temporary NTAP code began to phase out.
Ad Statistics
Times Displayed: 365866
Times Visited: 21120 MIT labs, experts in Multi-Vendor component level repair of: MRI Coils, RF amplifiers, Gradient Amplifiers Contrast Media Injectors. System repairs, sub-assembly repairs, component level repairs, refurbish/calibrate. info@mitlabsusa.com/+1 (305) 470-8013
The researchers found no significant differences in AI use based on patients' age, sex, race or ethnicity, or measures of stroke severity. Instead, hospital characteristics were the strongest predictors of adoption.
Comprehensive stroke centers were significantly more likely to use the technology, as were hospitals located in the Stroke Belt. In contrast, facilities serving communities with greater socioeconomic deprivation were significantly less likely to bill for AI use, suggesting that access to AI-assisted stroke evaluation may be more limited in underserved areas.
Lead author Casey Pelzl, principal research scientist at the Harvey L. Neiman Health Policy Institute, said the findings indicate that patients' access to AI-assisted stroke detection depends more on where they receive care than on their clinical needs.
The authors noted that while reimbursement can encourage adoption, it may not be sufficient for smaller hospitals that lack the infrastructure needed to integrate AI into clinical workflows. They suggested that alternative deployment models, including centralized or shared AI services, could help expand access in lower-resourced settings.
The study was supported by an American Heart Association grant.
