WASHINGTON, D.C. — Children’s National Hospital is using an artificial intelligence forecast in its pediatric emergency department to spot crowding risk early and call in extra staff before families leave without care in Washington.
Emergency physician Kenneth McKinley said the tool flags periods when more than two high-acuity patients are predicted to leave unseen over the next eight hours, triggering automatic alerts so leaders can mobilize backup doctors and nurses.
The platform updates every hour and weighs patterns such as day of week, time of day, holidays, arrivals, census, wait times, boarding, and standard overcrowding scores. It does not decide how to treat a child; it guides staffing.
Hospital leaders say wait times improved by more than 25 percent in September compared with the same month in 2023 and 2024 on days when the ED treated more than 200 patients. Federal figures still put average D.C.-area emergency stays above five hours.
Open research from Children’s National with Virginia Tech partners reviewed more than half a million ED encounters spanning 2018 through 2024. Gradient boosting and temporal-fusion systems performed best at spotting windows when several seriously ill children were likely to walk out, especially in the noon-to-evening peak.
Children’s National began deploying the AI approach in fall 2024. Leaders say the next step is moving forecasts earlier in the day so surge teams can act sooner while keeping the sickest children from walking out.
Section: More Washington DC business news
Filed under: Washington DC Business News
