The Food Truck Response to a Collapsing Engagement Score

Test the measurement first, then triangulate.

Measurement checks:

• Size and stability. At n=138 and a proportion near 0.69, the 95% confidence interval is roughly ±8 points; an 11-point drop is at the edge of meaningful and a 7-point drop at n=210 (about ±6) is not clearly outside noise on its own. Check three years, not two points.
• Who answered. A 11-point drop can be produced by a change in respondents. Compare response rate by unit, job class, and shift; check whether agency or travel staff were included this year but not last.
• Instrument changes. Confirm no wording, scale, vendor, or reporting-threshold change. Check whether the unit was reconfigured or merged, which changes the comparison group.
• Anonymity effects. In small units, staff may doubt confidentiality; look for a pattern of neutral answers rather than genuinely negative ones.

Triangulation (this is where the real answer lives):

• Turnover and vacancy by role and shift, first-year turnover, and exit-interview themes.
• Sick time, overtime, agency use, and schedule-variance data.
• Safety event reports, including near misses and whether staff feel safe speaking up.
• Patient complaints and comments that mention staff behavior, wait times, or call-light response - responsiveness is a staff-capacity symptom as much as a courtesy one.
• Free-text comments coded by theme, plus 15-20 skip-level conversations.

Then decide: if turnover, vacancy, and qualitative data all point the same direction, the survey is telling the truth and the driver is likely workload, voice, and respect. If the survey moved but nothing else did, treat it as a measurement question and fix the instrument before spending money.

What the proposals get right and wrong.

• Resilience app: places the burden of fixing a system problem on individual coping. It can be useful as a small adjunct but will read as "manage your own distress" when the distress is 25% turnover and short staffing. Low cost, low yield, some risk of resentment.
• Food-truck Friday: a morale gesture with no mechanism. In a unit where nurses cannot reliably take a break, a food truck can be actively irritating. Also inequitable across shifts - night shift never sees it.
• Engagement question in huddle: the best of the three if it feeds a closed loop; harmful if it is added to a huddle that already feels like a checklist and nothing comes back. One question is fine only if answers get responses within a defined time.

Alternative, aimed at the likely drivers:

1. Leader rounding with follow-through - a trained, scripted-but-genuine visit to every staff member quarterly, with issues logged and closed within 30 days, and visible reporting of what closed.
2. Voice infrastructure - unit-based shared governance or a staff council with real agenda control and a defined escalation path to the COO; publish a "you raised / we did" board.
3. Schedule and workload control - self-scheduling where feasible, shift-length options, break-relief coverage as a protected standard, and a hard look at vacancy and agency mix in the two units.
4. Safety and respect - visible executive response to the top three safety concerns raised, and a zero-tolerance standard for disruptive behavior, including from physicians.
5. Peer support and first-year onboarding - preceptor continuity and a structured check-in at 30/60/90 days, since first-year turnover is usually the largest single lever.

Cost logic: the app and food trucks are recurring spend with no measured mechanism; rounding and shared governance cost leader time, and break-relief coverage costs real dollars but is the item most directly connected to turnover. A single RN replacement is commonly estimated at tens of thousands of dollars, so avoiding a handful of departures in a 100-nurse unit often funds the coverage. Present it that way - cost per avoided departure, not cost per program.

Measure at 90 days with leading indicators: rounding completion and issues closed, break-relief rate, vacancy, first-year turnover, and a three-item pulse - not the annual survey.

The failure is in plan design, not staff motivation.

• Co-create, don't cascade. Plans written by the manager alone are the manager's plan. Require that at least half the actions come from staff, with staff named as owners. This is the single change most predictive of follow-through.
• One or two priorities, not nine. Most weak plans are long lists. Pick the top driver the unit can actually influence in 90 days.
• 90-day cycles with a named owner and a date. Quarterly review with the unit's executive sponsor present.
• Visible progress. A physical or digital board on the unit, updated every two weeks, showing the action, the owner, and the status. Visibility to the unit matters more than visibility to the C-suite.
• Accountability framed as support. The sponsor's question is "what do you need to finish this?" rather than "why isn't this done?" Blame converts the next survey into a marketing exercise.
• Kill the paperwork. If the plan is a template filed in a shared drive, it is compliance theater. Replace it with a one-page charter and a 15-minute monthly check.
• Close the loop publicly. Report back what happened to every idea raised, including the ones rejected, with reasons. That is what makes the next round of staff input honest.
• Measure the process, not just the outcome: percentage of staff-raised issues closed within 30 days, rounding completion, and participation in shared governance. The annual engagement score is too slow and too noisy to steer by.