🚑 Emergency Department Overcrowding Diversion Simulator

The ED at Steady State — Where Overcrowding Begins

Emergency department overcrowding is rarely caused by too many patients arriving at the front door. The dominant driver, confirmed across decades of health-services research, is output failure: patients who are ready to leave the ED but cannot, because there is nowhere to put them.

• ~145M: US ED visits / year (CDC NHAMCS estimate)
• ~90%: EDs reporting routine crowding (academic/urban centers)
• 4–6 h: Median ED LOS, admitted patient (US average, pre-pandemic baseline)
• 3 domains: Input vs throughput vs output (Asplin conceptual model, 2003)

The input-throughput-output model of ED crowding

Asplin et al.'s widely-used conceptual model (Annals of Emergency Medicine, 2003) frames crowding as a systems problem with three domains:

Input factors: community demand (influenza season, mass-casualty events), primary care access gaps that funnel non-emergent visits to the ED, ambulance volume.

Throughput factors: ED efficiency itself — triage speed, provider staffing ratios, lab/imaging turnaround time, internal process design.

Output factors: the ED's ability to move admitted patients upstairs to inpatient beds. This is overwhelmingly the dominant driver of severe crowding — an ED can be extremely efficient internally and still be overwhelmed if the hospital's inpatient beds are full.

The practical implication: interventions that only speed up ED throughput (more nurses, faster triage) have limited effect on severe crowding if hospital-wide bed capacity and discharge processes aren't addressed simultaneously.

Boarding — Admitted Patients With Nowhere to Go

"Boarding" refers to a patient who has been formally admitted to the hospital but remains physically in an ED bed because no inpatient bed is available. Every boarding patient occupies a bed and consumes nursing resources that could otherwise be used for new arrivals.

• >24 h: Boarding time, definition of "extreme" (a widely used crowding flag)
• <4 h: ACEP boarding time target (professional society goal, rarely met)
• Increased with boarding >6h: Mortality association (multiple observational studies)
• 10–30%: Boarding as % of total ED census (during high-crowding periods)

Why boarding cascades into worse crowding

Boarding creates a compounding feedback loop: each boarded patient occupies an ED bed and nursing attention indefinitely, reducing the number of beds available for new triage-to-bed placement. New patients then wait longer in the waiting room or hallway, delaying time-sensitive care (antibiotics for sepsis, reperfusion for STEMI/stroke) and increasing left-without-being-seen rates.

Root causes of boarding are almost always upstream of the ED itself: inpatient bed availability constrained by staffing (not physical beds), delayed discharges (waiting on transport, skilled nursing facility placement, family logistics), and elective surgical schedules that are prioritized over ED-driven admissions for hospital revenue reasons.

Evidence-based mitigation strategies include: full-capacity protocols (temporarily boarding admitted patients in inpatient hallways instead of the ED, which distributes the burden hospital-wide rather than concentrating it in one department), early discharge rounding to clear beds before noon, and surgical schedule smoothing to reduce weekday elective-case bed competition.

Multiple studies (including a large Canadian cohort, Chan et al.) have linked ED boarding time beyond 6 hours to increased inpatient mortality — not because boarding itself causes harm directly, but because it is a marker of a system too overwhelmed to deliver timely care at every subsequent step.

Quantifying Crowding — the NEDOCS Score

The National ED Overcrowding Study (NEDOCS) score is the most widely validated composite metric for real-time ED crowding, converting multiple raw operational variables into a single 0–200+ scale used to trigger internal escalation protocols and diversion decisions.

• 7: NEDOCS variables (beds, admits, LOS, wait time, ventilators, etc.)
• Severe: Score 100–139 (NEDOCS interpretation band)
• Dangerously overcrowded: Score ≥140 (NEDOCS interpretation band)
• Weiss et al. 2004: Original validation (Academic Emergency Medicine)

How the score is computed and used operationally

NEDOCS combines: total ED beds, total hospital beds, number of ED patients, number of admitted (boarding) patients, longest admit boarding time, waiting room time, and number of patients on ventilators — into a single weighted formula producing a score typically in the 0–200 range (though it is unbounded upward in extreme cases).

Interpretation bands (as originally validated): <20 not busy, 20–60 busy, 60–100 crowded, 100–140 severely crowded, 140–180 dangerously overcrowded, >180 hazardous.

Many hospitals use NEDOCS (or a locally adapted equivalent) as an automated, real-time dashboard metric that triggers a stepwise internal escalation policy — surge staffing, calling in charge-nurse huddles, activating full-capacity protocols — well before the more drastic step of formal ambulance diversion is considered.

Ambulance Diversion — the Last-Resort Lever

When internal escalation fails to relieve crowding, a hospital may request "ambulance diversion" status: asking EMS dispatch to route new incoming patients to other nearby facilities. This is a blunt instrument with a poor evidence base for actually solving the underlying problem.

• >50%: Hospitals reporting diversion (peak years) (annually, US urban EDs, mid-2000s)
• Massachusetts (2009), others: US jurisdictions banning routine diversion (statewide policy)
• 5–20 min: Added transport time when diverted (typical, more in rural areas)
• Often null or negative: Effect on system-wide wait time (multiple studies)

Why diversion often fails to help — and can make things worse

The core problem with ambulance diversion is that it treats a symptom (too many patients at Hospital A) without addressing the cause (not enough discharge-ready inpatient beds across the whole region). If neighboring hospitals are also boarding and crowded — which is common, since crowding drivers like flu season or a regional bed shortage affect an entire area simultaneously — diverting patients from Hospital A to Hospital B simply shifts the same problem to Hospital B's door, and delays care for the diverted patient with an added 5-20 minute (or longer) transport time.

Multiple studies, including a landmark analysis after Massachusetts eliminated routine ambulance diversion in 2009, found no increase in ED wait times or ambulance turnaround times statewide — the anticipated harm from banning diversion did not materialize, suggesting the practice provided limited real system-wide benefit while adding real delay and coordination cost for diverted patients.

As a result, many regions and the CMS Conditions of Participation framework now discourage or prohibit routine diversion, instead mandating internal full-capacity protocols and requiring hospitals to accept and manage their own crowding rather than externalizing it to their neighbors.

When an Entire Region Goes on Diversion Simultaneously

The most dangerous scenario in EMS systems science is "all-hospital diversion" — when every ED in a region is simultaneously overcrowded and requests diversion, leaving EMS with no legal or practical destination for a new critical patient.

• Documented in most major metros: All-hospital diversion events (periodic occurrence, flu season peaks)
• Nearest ED regardless of status: EMS response, all-diversion scenario (default safety override in most protocols)
• MCI/EMS coordination centers: Regional coordination bodies (monitor real-time regional bed status)
• Diversion elimination or strict capping: Policy trend (growing number of US states/regions)

System-level solutions beyond any single ED

When simultaneous regional overcrowding is anticipated (flu season, a mass gathering event, a mass-casualty incident), most EMS systems have a standing override rule: diversion status is automatically suspended for all facilities, and ambulances transport to the geographically or clinically most appropriate ED regardless of stated crowding, because a moving patient with an unstable condition cannot safely be left without a destination.

Longer-term system-level solutions that address the root cause rather than shuffling patients between crowded EDs include: regional bed-status dashboards shared in real time across hospitals and EMS dispatch, load-balancing agreements that proactively route non-critical transports to less-crowded facilities before any single ED reaches crisis level, and hospital-wide (not just ED) capacity management — smoothing elective surgical admissions, expanding discharge-planning staff on weekends, and investing in skilled-nursing/post-acute placement capacity so that boarding is resolved at its source rather than displaced onto neighboring EDs.

The strongest available evidence suggests that regional, coordinated capacity management — not ambulance diversion between individual competing hospitals — is what actually reduces system-wide ED crowding and patient harm.