Quiet at Night: A Failed Change and a Second Attempt

Six design flaws, each sufficient to explain the null result.

1. Wrong time. The intervention ran 2-4 pm and the outcome is about nighttime. The change could not plausibly move the measure, so a null result tells you nothing about the idea - only about the fit.
2. No theory of cause. Nobody had established what was actually waking patients. The intervention was chosen before the problem was defined.
3. Wrong owners. Day shift designed and ran a change for a night-shift problem, and the two departments named by night staff as the cause (lab, EVS) were never in the room.
4. Weak intervention type. A sign and a huddle reminder is an awareness intervention in a system of competing production pressures. Awareness fades in about two weeks.
5. No measurement. The only measure was a quarterly survey with a small denominator - far too slow and too noisy to tell whether the change did anything. There was no process measure at all.
6. The proposed fix repeats the error, louder. A compliance audit adds pressure on the people who are already trying, does nothing about 4 am labs or the linen cart, and risks turning a quality effort into a policing exercise that night staff will quietly resist.

The framing to use with the manager: "The quiet-time trial wasn't a failure of staff effort; it was a test of the wrong change at the wrong time with no measurement. If we run it again the same way we'll get the same answer, and we'll have spent the unit's credibility."

Define the problem with patients first (weeks 1-2). Interview 10-15 recent patients and 10 night-shift staff: "What woke you? What did you hear?" Reconcile the lists.

Measures - a family, not one number:

• Outcome: quietness top-box (tracked monthly on a p-chart, understood as lagging and noisy); a patient-reported item at discharge ("Did noise keep you awake?").
• Process: noise readings at the desk and outside three rooms at 02:00, 04:00, and 05:30; count of overhead pages 22:00-06:00; count of 4 am draws on stable patients; linen-cart timing log; door-closing compliance.
• Balancing: lab turnaround time, missed or delayed draws, call-light response time, falls, patient satisfaction with staff responsiveness. Any noise fix that delays care or hides staff is not a fix.

Change ideas mapped to causes:

• 4 am labs: review the order set with lab and medicine; default stable patients to 05:30; build an EHR order default rather than relying on individual requests.
• Linen cart: reroute and retime with EVS, or use a quiet-wheel cart and a closed-door staging area.
• Desk noise: a visual cue at the desk, a huddle norm, and confining conversation to a side room after 22:00.
• Alarms: an alarm-management audit with parameter review and a nurse-led default adjustment.
• Overhead paging: convert non-emergent pages to phone or secure text; reserve overhead for true emergencies.

PDSA sequence: start with the highest-yield, lowest-friction change (cart timing and desk noise) on one hallway for two weeks with a stated prediction; then lab timing; then alarms. One change at a time, or you cannot attribute the result. Meet weekly for 15 minutes with the night-shift champion; adapt, adopt, or abandon each cycle.

Convert a complaint into a shared problem with data. Bring lab and EVS leadership the noise readings, the patient quotes, and the specific times - and frame it in their language: sleep is part of healing, delayed draws are part of their quality score, and rerouting a cart costs nothing. Ask them what constraint you don't understand. Departments resist mandates and generally respond to a specific, bounded, evidence-backed request.

Structure:

• A short-term working group with a decision-capable representative from lab, EVS, nursing, and facilities; four meetings, not a permanent committee. Give it a defined scope: nighttime noise sources under each department's control.
• A night-shift champion with paid project time and visible recognition - not an unpaid extra duty. Skepticism drops sharply when the person leading the work is a peer who works those hours.
• Test in one hallway and show the night staff the readings before and after. Their skepticism is rational; evidence is the only thing that moves it.
• Hardwire what works: EHR order defaults, a policy on overhead paging, a written cart route, an orientation module, and inclusion in the charge-nurse checklist. Changes held together by enthusiasm decay in about six weeks.
• Close the loop with staff and patients: report what was tried, what worked, and what was abandoned and why. This is also what earns cooperation for the next cycle.

Sustainment: keep the monthly p-chart and a quarterly re-audit of the process measures. Name a permanent owner rather than relying on the project lead, and re-check at six months - improvement that depends on one champion disappears when the champion changes jobs. If scores drift after the project ends, the answer is the hardwired defaults, not another quiet-time sign.