King Faisal Specialist Hospital and Research Centre in Riyadh has extended its AI-supported emergency-demand forecasts from 12 hours to 24 hours, it announced on 21 October. The longer window is intended to help teams prepare staffing and capacity before pressure builds. Clinical staff continue to make triage and care decisions.[1]
The useful promise is time to act. Seeing a likely surge while staff can still adjust a shift, prepare a bed or coordinate a discharge is operationally different from spotting congestion after patients are already waiting. The forecast creates an earlier decision point; the hospital still needs people and workable processes to make use of it.
Following the patient beyond the emergency door
KFSH’s command centre dates to September 2021. Its description spans four connected functions: forecasting demand, tracking actual patient flow, monitoring inpatient journeys and coordinating operating-room resources. This is a hospital-wide view of capacity, including the places patients move to after an emergency assessment.[3]
In a separate account on 19 October, the hospital described how its system identifies delays, alerts the responsible team and escalates unresolved problems. It also changed the work around that system: stable patients can enter inpatient care directly from outpatient clinics, morning discharges happen earlier, and teams have accelerated discharge medicines, laboratory services and radiology.[2]
Those details explain why an emergency queue cannot be understood solely as a problem at the front desk. A patient ready for admission may occupy an emergency space while an inpatient bed is unavailable. A delay in discharge medication can therefore affect a different patient who has not yet reached the emergency department. That is the connection a hospital-wide coordination system is trying to make visible.
Several improvements, with different clocks
The hospital reports reductions in both the wait for a physician and the wait for an inpatient bed. They describe different stages of care. In particular, the boarding-time comparison runs from 2023 to mid-2025, preceding this week’s announcement of the longer forecast.[2][1]
What the hospital’s reported measures refer to
| Measure | Reported change | Scope |
|---|---|---|
| Arrival to physician assessment | About 42% shorter | The 21 October release does not give the baseline duration or measurement period.Source 1 |
| Median emergency boarding time | More than 13 hours to 3.2 hours | 2023 to mid-2025; admitted patients waiting for an inpatient bed.Source 2 |
| Average daily emergency visits | About 8.5% higher | Compared with 2022; a measure of demand, separate from waiting times.Source 1 |
The forecast release does not name the model, report prediction errors or compare the two horizons on a common test. The announcements describe operational changes, rather than a controlled estimate of AI’s contribution.[1]
For the hospital, the practical question is whether earlier notice arrives while something can still be changed. Extending that opportunity from half a day to a full day gives the coordination team more time; staffing, beds and discharge decisions determine what it can make of it.
Sources & further reading
- KFSH forecasts emergency demand 24 hours ahead with AI King Faisal Specialist Hospital & Research Centre · 21 Oct 2026 · Accessed 21 Oct 2026
- KFSH reports reduced emergency-department boarding time King Faisal Specialist Hospital & Research Centre · 19 Oct 2026 · Accessed 21 Oct 2026
- Predicting healthcare demand: the Patient Flow & Capacity Command Centre King Faisal Specialist Hospital & Research Centre · Accessed 21 Oct 2026