AI language models struggle with basic hospital data tasks, study finds
Peer-Reviewed Publication
Updates every hour. Last Updated: 18-May-2026 05:16 ET (18-May-2026 09:16 GMT/UTC)
A new study finds that large language models (LLMs), used with straightforward prompting, perform poorly on routine number-crunching tasks that hospital administrators depend on every day to track patients and allocate resources. The findings were published this week in the open-access journal PLOS Digital Health by Eyal Klang of the Icahn School of Medicine at Mount Sinai, New York, USA, and colleagues.
The TCT Geoffrey O. Hartzler Master Operator Award will be presented to Rebecca T. Hahn, MD, during Transcatheter Cardiovascular Therapeutics (TCT), the annual scientific symposium of the Cardiovascular Research Foundation (CRF). TCT will take place October 31 – November 3, 2026, in San Diego, California. The award is given each year to a physician who has advanced the field of interventional cardiovascular medicine through technical excellence and leadership. Dr. Hahn is the first imager ever to receive this distinction — a testament to how profoundly she has advanced the role of imaging in cardiovascular care.
A large-scale analysis of wildfires in the Western United States shows that U.S. Forest Service fuel-reduction treatments not only curb fire spread and intensity, but also, for each dollar invested, generate more than three times the value in avoided damages. Wildfire activity has intensified dramatically over the last several decades, imposing widespread economic, environmental, and public health damages that amount to hundreds of billions of dollars annually in the United States alone. These risks are expected to grow as climate change and development in fire-prone areas continue. A central driver of worsening wildfire severity is the buildup of combustible vegetation, or “fuel loads,” which were kept in check historically by frequent, low-intensity fires, including those deliberately set through Indigenous land stewardship practices. Fuel-reduction strategies such as prescribed burns and forest thinning aim to restore more resilient conditions and mitigate wildfire-related damages. However, these measures remain underutilized, in part because their economic benefits are delayed and difficult to quantify, and because limited data and complex fire dynamics make it challenging to evaluate their overall effectiveness.
Focusing on the Western United States, where wildfire risk and data are abundant, Frederik Strabo and colleagues compiled a high-resolution dataset encompassing 285 wildfires that intersected with United States Forest Service (USFS) fuel treatment activities across 11 states between 2017 and 2023. By comparing observed fires with modeled scenarios in which no treatments occurred, Strabo et al. estimated the damage avoided due to fuel treatments and assigned the avoided damage an economic value. This allowed the authors to evaluate not only whether fuel treatments work, but also when and where they are most cost-effective. According to the findings, fuel treatments substantially reduced both the spread and intensity of wildfires, likely by reducing flame intensity and making conditions more manageable for suppression crews. In total, treatments reduced total burned area by 36% over the study period compared to scenarios in which no fuel treatments were applied. The authors’ estimates suggest that these interventions prevented roughly $2.7 to $2.8 billion in damages, including reduced property loss, carbon dioxide emissions, and harmful air pollution. Moreover, on average, each dollar invested in fuel reduction yields more than three dollars in avoided damages, with many projects performing even better, suggesting that targeted fuel treatment strategies could further amplify these returns. “But realizing [these strategies’] full potential will require more than scientific consensus – it will demand bold policy reform,” write Strabo et al.
Simon Fraser University (SFU) and Thompson Rivers University (TRU) are forging a new partnership to advance innovation, research, graduate studies and workforce development across British Columbia.
This new partnership – bolstered by a memorandum of understanding signed this month – will lead to collaborations in strategic priority areas for B.C., including artificial intelligence (AI), cybersecurity, health, wildfire management and emergency response, and Indigenous language revitalization.