Emergency Department Flow Constraint Without Capital
A 55,000-visit emergency department has a median door-to-provider time of 68 minutes; the target is 30. The left-without-being-seen rate is 4.8%, up from 3.1% a year ago. Laboratory turnaround median is 92 minutes (target 60); CT turnaround is 138 minutes. Inpatient boarding averages 6.4 hours for admitted patients. The ED has 18 beds plus 6 hallway spaces. Triage uses a 5-level system; documentation is on paper for the first 20 minutes, then entered into the EHR. Two rapid-improvement events focused on triage failed because nurses were not involved in design and the medical director changed triage criteria without training. Administration has no capital for new beds but can fund 1.0 FTE for a six-month process analyst.
1. Construct a high-level current-state value stream for an uncomplicated abdominal pain patient and identify the most likely bottleneck. Explain what data you would collect to confirm it.
Read the worked answer
Current-state value stream: arrival, triage (paper, 5-level), registration, room or hallway, provider assessment, orders (labs, CT), lab collection/result, CT transport/scan/read, reassessment, and disposition (admit or discharge). Waiting occurs between most steps.
Likely bottleneck: not triage alone. With 18 beds plus 6 hallway spaces and 6.4-hour boarding, the constraint is likely inflow-to-bed matching and diagnostic turnaround, especially CT and lab. Door-to-provider of 68 minutes is a symptom; hallway spaces and boarding reduce effective capacity.
Data to confirm: time stamps for each step; queue counts every 15 minutes; bed occupancy; provider staffing by hour; arrival pattern; lab collection-to-result and CT order-to-read; transport delays; boarding hours; LWBS by hour; walk-through observation. Use a Pareto of delay minutes and a process map with wait times. Without these data, triage redesign may optimize a non-constraint.
2. Propose three redesign changes that do not require new beds or capital, and sequence them.
Read the worked answer
Changes: (1) Split flow: low-acuity abdominal pain patients go to a vertical zone with chair seating and a dedicated provider/MA pair; higher-acuity patients use beds. (2) Provider-in-triage or triage-based order entry: a provider or protocol nurse orders labs and CT early for stable patients, reducing serial waiting. (3) Stat lab and CT fast track: dedicated ED lab runner and CT slot protocol; results callback standard.
Sequence: first map and baseline (2 weeks); then test split flow on one shift with PDSA (weeks 3-6); add early ordering only after triage criteria are stable and staff trained (weeks 5-10); then negotiate CT slots with radiology (weeks 8-12). Use the 1.0 FTE process analyst to facilitate, not to own. Measures: door-to-provider, decision-to-discharge, LWBS, ED length, CT/lab turnaround, hallway use, and staff experience. Risk: split flow can worsen inequity if assignment is biased; audit by language and acuity.
3. How will you sustain the improvements and avoid the previous failure?
Read the worked answer
Sustainment: co-design with nurses and physicians; do not change triage criteria without training and sign-off. Assign a process owner for each change; use daily huddles and visual boards; review run charts weekly, not monthly. Build standard work for split flow and early ordering; train new staff during onboarding. Add balancing measures: LWBS, ED length, patient experience, staff burnout, and hallway boarding. Use a control chart to distinguish common-cause variation from real change. Escalate to administration if boarding consumes gains; ED flow cannot be fixed inside the ED alone. Avoid prior failure: the previous events failed because nurses were excluded and criteria changed without training. This time, include frontline staff in design, test on one shift, document decisions, and use a 'stop rule' if safety or experience worsens. Celebrate small wins but do not declare success until 3 months of stable data.