How this instrument works
NEDOCS is a departmental crowding score for emergency departments, published by Weiss, Derlet, and colleagues as the National ED Overcrowding Study (Academic Emergency Medicine, 2004). It was built from 336 timed site-samplings across eight academic medical centers, comparing objective measurements — bed counts, patient counts, wait times — against staff's own subjective sense of how crowded the department felt at that moment, then fitting a formula to predict that subjective rating from the objective numbers alone.
The formula is exact and published: NEDOCS = 85.8×(C/A) + 600×(F/B) + 13.4×D + 0.93×E + 5.64×G − 20, where A is total staffed ED beds, B is total inpatient hospital beds excluding pediatrics and OB, C is total ED patients right now, D is the count of critical care patients in the ED, E is the longest ED-boarding admit wait in hours, F is the count of ED admits still waiting for an inpatient bed, and G is the most recently placed patient's door-to-bed time in hours. This calculator reproduces those coefficients exactly, corroborated against NEDOCS's own published variable definitions.
This is a department-level operations metric, not a tool for diagnosing, treating, or predicting the outcome of any individual patient. A high score describes system strain — bed pressure, boarding, and throughput — at the moment it's calculated; it says nothing about any one patient's condition or risk. The published score maps to six severity bands: 0-20 not busy, 21-60 busy, 61-100 extremely busy but not overcrowded, 101-140 overcrowded, 141-180 severely overcrowded, and 181 or above dangerously overcrowded.
- Enter Total ED beds (A) — staffed gurneys, chairs, and hallway spots actually in use right now, not the licensed maximum.
- Enter Total inpatient hospital beds (B), excluding pediatric and OB beds.
- Enter Total ED patients (C) currently in the department, including hallway, waiting-room, and boarded-admit patients.
- Enter Critical care patients (D), longest ED-boarding admit wait in hours (E), ED admits still waiting for a bed (F), and the last patient's door-to-bed time in hours (G).
- Read Score and Severity tier — tier 0 is 'not busy,' tier 5 is 'dangerously overcrowded,' matching the published NEDOCS bands.
Worked example — 30 ED beds, 34 patients
30 staffed ED beds (A), 300 inpatient hospital beds (B), 34 ED patients right now (C), 2 critical care patients (D), 3 admits waiting for a bed (F), a 2-hour longest admit wait (E), and a 1-hour last door-to-bed time (G): 85.8×(34/30) = 97.24, plus 600×(3/300) = 6, plus 13.4×2 = 26.8, plus 0.93×2 = 1.86, plus 5.64×1 = 5.64, summing to 137.54, minus 20 gives a NEDOCS score of 117.54 — tier 3, 'overcrowded' (101-140).
The same formula moves a lot with the inputs. A quieter department — 20 beds, 15 patients, no boarders, a 0.5-hour last door-to-bed time — lands around 47, tier 1, 'busy.' A severely strained one — 25 beds but 60 patients, 8 critical care, 12 admits waiting, and a 6-hour longest admit wait — lands around 350, tier 5, 'dangerously overcrowded,' driven mostly by the heavily weighted admits-waiting and critical-care terms.
Questions
Who is this calculator for?
Emergency department charge nurses, physicians, and hospital operations staff tracking real-time departmental strain — not patients trying to assess their own care or risk. NEDOCS describes how crowded a department is as a whole at a given moment; it has no patient-specific inputs and produces no patient-specific output.
What do the six severity tiers actually mean?
They're the published NEDOCS bands: 0-20 is not busy, 21-60 is busy, 61-100 is extremely busy but not yet overcrowded, 101-140 is overcrowded, 141-180 is severely overcrowded, and 181 or higher is dangerously overcrowded. Many departments use crossing into the higher bands as a trigger for surge protocols, such as opening additional beds or diverting non-critical transfers.
Can NEDOCS be used to guide or predict the outcome of an individual patient's care?
No. It was built and validated purely as a department-level crowding metric — it takes no patient-specific clinical inputs (no vitals, diagnosis, or acuity beyond a simple critical-care headcount) and was never designed to estimate any one patient's risk or outcome. Use it for staffing and operations decisions, not clinical decisions about a specific patient.
Why does F, ED admits waiting for a bed, carry such a large coefficient (600)?
Because it's divided by total inpatient hospital beds (B), a much larger number than ED bed count, the raw ratio F/B is small — the large coefficient rescales it back into a meaningful contribution. In practice this term captures hospital-wide boarding pressure: admitted patients stuck in the ED because no inpatient bed is open, a major driver of ED crowding that Weiss et al.'s regression weighted heavily based on its real correlation with staff-perceived crowding.
How was this formula actually derived?
Weiss SJ, Derlet R, and colleagues collected 336 timed site-samplings across eight academic medical centers, pairing objective counts (beds, patients, wait times) with staff's own subjective crowding ratings at that same moment, then fit a regression to predict the subjective rating from the objective numbers. The coefficients in this calculator are exactly the published result of that regression, from Academic Emergency Medicine, 2004.
Where can I double-check these coefficients against another source?
NEDOCS's own published 'NEDOCS Variables and Definitions' documentation (mirrored by Juvare/EMResource, a hospital-status platform that implements the same score) lists the identical formula, variable definitions, and severity bands used here — a link is in the references below alongside the original Weiss et al. paper.
References
- Weiss SJ et al. — Estimating ED Overcrowding: The NEDOCS Study (Acad Emerg Med, 2004)
- Juvare EMResource — NEDOCS Calculation (formula, variables, severity bands)
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.