Wait Time Measurement & Definitions — get a reliable baseline

Good improvement begins with clear measurement. Different teams call "wait time" different things; if you don't agree on the exact start and end points, numbers will mislead. This guide gives practical definitions, recommended core metrics, and measurement methods you can use in EDs, clinics, labs, and imaging.

Key interval definitions (use plain event names your EHR or workflow can record)

• Arrival-to-triage: Time from patient arrival (or check-in) to completion of triage or initial clinical screening.
• Arrival-to-room (or seat): Time from arrival to being taken to the exam room or assigned treatment area.
• Arrival-to-provider (door-to-provider): Time from arrival to the first assessment by a clinician who can make care decisions.
• Room-to-departure (length-of-visit): Time from entering room to physical departure, including documentation and check-out steps.
• Turnaround times for tests/procedures: e.g., order-to-result for labs, arrival-to-scan for imaging.

Core metrics to track (minimum set)

1. Median (and IQR) arrival-to-provider: Prefer median over mean when distributions are skewed.
2. 90th percentile (p90) arrival-to-provider: Shows tail risk—how bad waits get for the slowest 10%.
3. Left without being seen (LWBS) rate: Percent of patients who leave before clinical assessment.
4. No-show and cancellation rate: Affects clinic capacity and effective demand.
5. Throughput by visit type: Separate new vs follow-up, simple vs complex visits for fair comparison.

Measurement methods — pick the simplest accurate method you can sustain

• EHR timestamps: Best when event points are recorded consistently (check-in, triage complete, rooming, provider start). Validate time-stamps with spot checks.
• Manual sampling: If EHR timestamps are unreliable, collect time-stamped samples for representative shifts (random sample across days and clinicians).
• Observer time-motion: Useful for bottleneck analysis when you need rich contextual detail, but expensive to run continuously.

Statistics and sampling guidance

Use median + IQR plus p90 to capture central tendency and tail. Aim for at least 30–50 samples per major shift or clinic session to reduce noise. When in doubt, repeat sampling on different weekdays and times (morning/afternoon) to capture variation.

Common measurement traps

• Mixing visit types (new vs return) hides true performance—segment metrics.
• Counting scheduled time as arrival masks late-arriver effects—use actual arrival timestamps where possible.
• Relying solely on averages hides long tails—always report a tail metric (p90) and LWBS.
• Changing definitions midway ruins trend comparison—freeze definitions until a new baseline is re-measured.

Next step

Use the interactive Wait Time Measurement Audit to collect a baseline that follows these definitions. Then use bottleneck analysis to identify where to run small tests.