criticalaccess.aiAI for rural hospitals

Teaching tool · Live demo

Watch an emergency department bottleneck form.

It is hard to see how staffing, demand, and the wait for an inpatient bed combine into long waits, boarding, and patients who leave without being seen. This simulation lets you change one thing at a time and watch what happens.

three.js r128JavaScriptOne HTML fileNo build step

01 · At a glance

The question, the data, and the terms.

This is a teaching and exploration tool, not a planning or clinical tool. Every patient in it is generated.

The question
How do arrivals, staffing, and inpatient bed waits turn into a bottleneck in the emergency department?
Who it’s for
ED and nursing leaders, operations staff, and students, as a way to teach and explore patient flow.
Data
No real data. All patients and timings are generated. The parameters are illustrative and are not calibrated to any real hospital.
Status
A working teaching tool, with a live demo.
License
No license is stated in the repository.

02 · What it does

A 15-bed department you can push on.

The model follows each patient from arrival through triage, a bed, a nurse, a provider, CT, and then discharge or admission.

  • A

    The department

    2 resuscitation beds, 9 acute care beds, and 4 fast-track beds, plus one triage room and one CT scanner.

  • B

    What you can change

    Arrivals, from 30 to 110 a day. The number of providers and nurses. The wait for an inpatient bed. You can also send a five-patient multi-vehicle crash.

  • C

    What it measures

    Door-to-provider time, length of stay, the rate of patients who leave without being seen, and boarding.

03 · What it shows

An illustration, not a prediction.

With one provider and one nurse, a three-day run of the model gives median door-to-provider times over 100 minutes, and between 5% and 21% of patients leave without being seen. That is an illustration from a model with unfitted parameters. It is not a finding about real emergency departments.

The simulation screen. A row of live measures sits above an angled 3D floor plan of the emergency department. Sliders on the right set arrivals per day, providers, nurses, and the wait for an inpatient bed.
The full screen: live measures across the top, the department in 3D, and the scenario controls. All patients are synthetic.
Top-down view of the floor plan: a waiting room, triage, acute care beds, fast-track beds, resuscitation beds, a CT scanner, and a nurses' station, with beds colored by status.
The top-down view. Bed colors show status: waiting on staff, in treatment, boarding, or needing cleaning.
The simulation in dark mode with most beds occupied and two patients on the waiting list, one of them a multi-vehicle crash trauma patient.
Dark mode, after sending the multi-vehicle crash: five trauma patients arrive by ambulance at once.

04 · Limits

Don’t staff a department from this.

The project’s own documentation calls it “a teaching and exploration tool, not a planning or clinical tool.”

Limits
  • The parameters are plausible defaults, not fitted to data. Do not read the outputs as predictions for a real department.
  • It is simplified. There are no nurse-to-patient ratios, breaks, or shift handovers. There are no consultants, hallway beds, or pediatric or behavioral health areas. There are no deaths or transfers, and walking takes no simulated time.
  • Only CT is modelled as a shared resource.
  • It was checked in headless Chrome on macOS only.
  • The live demo needs an internet connection to load three.js and fonts, and WebGL for the 3D view.

Open it and move a slider.

The demo runs in your browser. The whole thing is one HTML file, with no build step.