ED Throughput and the Boarded Psychiatric Patient
A 420-bed community hospital with a 28-bed ED has averaged 11 psychiatric boarders per day for three weeks, each waiting a median of 31 hours for an inpatient psychiatric bed. Two of the four behavioral-health sitters call out; nursing leadership reports the ED holds admitted medical patients an average of 6.2 hours. The CMO asks you, the ED medical director, to present a plan at the next operations meeting. Administration has offered one lever: fund three additional sitters and a 4-bed secured observation unit, but only if you can show it will not simply relocate the bottleneck. The psychiatry service says its 22 inpatient beds are full and its attending coverage cannot expand before next fiscal year. The hospital is at 94% occupancy and diversion has been triggered 9 times this month.
1. Before endorsing the 4-bed secured observation unit, what data would you pull, and what would each dataset tell you about whether the unit relieves or relocates the bottleneck?
Read the worked answer
I would separate demand, flow, and capacity data.
- Arrival and disposition data (psychiatric arrivals by hour/day, percent admitted vs. discharged from the ED, length of stay by disposition): tells me whether the bottleneck is true inpatient demand or a disposition problem. If most boarders eventually discharge from the ED after long waits, a secured observation unit genuinely helps; if most are admitted, it just creates a holding pen.
- Inpatient psychiatric bed turnover (average LOS, discharge hour distribution, bed-days blocked by placement or legal holds): if discharges cluster in the afternoon, morning admissions stall. This is a scheduling fix, not a bed-count fix.
- Transfer/out-of-county acceptance data: how many boarders could be transferred and how often receiving facilities decline. This sizes the external-capacity lever.
- Medical boarding data (the 6.2-hour holds): if medical holds and psych holds compete for the same physical space and staff, adding psych capacity may free medical flow; if they are independent, expect no change.
- Sitter utilization and acuity mix: three sitters may be inadequate if 1:1 observation is required; quantify how many boarders need continuous observation versus close observation.
The decision rule: the unit is justified only if a meaningful share of boarders are dischargeable within 24 hours and the binding constraint on inpatient psych is discharge timing or staffing, not raw bed count.
2. The psychiatry service refuses to expand attending coverage this year. Draft the operational argument you would make to administration about what is and is not achievable with the offered resources, and what you would ask for instead.
Read the worked answer
I would frame this as a constraint-honest proposal rather than accepting or rejecting the offer wholesale.
What the offered package can achieve:
- A secured observation unit can safely hold low-acuity, likely-dischargeable patients, reducing ED sitter demand and freeing ED beds. It does not add inpatient psychiatric capacity.
- Three sitters address a staffing gap, but only if scheduled to the arrival curve; I would ask for flexible deployment rather than fixed shifts.
What it cannot achieve:
- It will not reduce the 31-hour median for patients who require inpatient admission. With 22 beds full and no coverage expansion, admitted boarders will still wait.
What I would ask for instead or in addition:
- A discharge-before-noon pilot on the psych unit with a target metric (percent discharged by 12:00), which increases effective bed turnover without new beds or attendings.
- A transfer agreement and dedicated transfer coordinator to move appropriate boarders to regional facilities, with data on declines.
- Crisis/telepsychiatry consultation to shorten ED decision time and reduce avoidable admissions.
- A written escalation metric (e.g., psych boarder-hours per day) reported to the operations committee so the constraint stays visible.
I would explicitly state the expected effect size for each element and commit to re-measuring in 60–90 days.
3. Diversion has been triggered 9 times this month. How would you counsel the group on the limits of diversion as a throughput strategy?
Read the worked answer
Diversion is a pressure-relief valve, not a capacity solution, and it carries real costs: ambulances are out of service longer, neighboring EDs absorb the load (and may reciprocate), and community response times worsen. Repeated diversion also signals that the underlying constraint is unaddressed.
I would counsel the group to:
- Treat diversion as a measured, time-limited action with a defined trigger and a defined exit, not a default response to crowding.
- Track whether diversion actually improves internal flow; if ED boarding is driven by admitted patients awaiting inpatient beds, diverting arrivals does little because the constraint is downstream.
- Watch for regulatory and reputational exposure: frequent diversion can draw scrutiny and erodes referral relationships.
- Redirect effort to the real levers: inpatient discharge timing, bed assignment processes, and the psych disposition pathway.
My message: use diversion sparingly and transparently, but do not let it substitute for fixing the boarding problem.