Four Pressure Injuries in One Quarter

Problem statement (what, where, when, how much, compared to what): "Between October and December, four patients on the 28-bed medical unit developed hospital-acquired pressure injuries (two stage 2, two unstageable), against zero in the preceding quarter. Concurrently, Braden reassessment after a change in condition occurred in 41% of audited charts and repositioning was documented for 58% of high-risk patients."

Aim statement (SMART): "By June 30, reduce hospital-acquired pressure injuries on the medical unit from 4 per quarter to 0–1 per quarter, and raise reassessment-after-change-in-condition from 41% to ≥90%."

What I deliberately left out:

1. Any mention of the refresher in-service. Naming the solution in the problem statement guarantees the team debates the solution instead of the problem, and the first meeting becomes a fight about whether education works.
2. A single-cause explanation. "New grads caused it" is tempting given the turnover and would be both unfair and wrong - 22% turnover affects the whole unit's continuity, not just the new hires. Blaming a group kills the reporting that produces the data.
3. A target of zero with no baseline honesty. Zero is the right aspiration; setting it as a hard target without acknowledging that pressure injuries are multifactorial invites under-reporting, which is the one outcome I most want to avoid.

Also deliberately included: the process measures in the same breath as the outcome. Without them, the team cannot tell whether a good quarter means the work improved or the reporting changed.

Selection logic: match the intervention to the mechanism the data reveal. The audit points at two distinct failure mechanisms and they need different fixes:

1. Reassessment failure (41%) - a reliability problem, not a knowledge problem. The fix is a trigger: mandatory Braden reassessment tied to defined events (transfer, new sedating medication, surgery, condition change) embedded in the workflow, not a memory task. Education cannot fix a prompt that does not exist.
2. Repositioning failure (58%) - this is where the evidence is strongest and most specific: support surfaces and structured repositioning for high-risk patients reduce incidence; but documented turning ≠ actual turning. I would pair any turning protocol with a method that validates it (rounding with a visible cue, or skin inspection at each shift handoff), because documentation compliance is exactly what the chart already overstates.

Appraising what to adopt: for each candidate intervention I ask three things - (a) is there evidence it works in a population like ours, (b) does it address one of our two mechanisms, and (c) what does it cost in nursing time that something else currently uses?

On the in-service: nutrition screening and skin assessment education have modest, real effects; a general refresher does not. If I use education at all, I would target it narrowly at new-grad onboarding and at why reassessment timing matters - not as the unit-wide fix.

Delivery method: one mechanism at a time, PDSA cycle, measurable process target (reassessment ≥90% within one month), so a failure is small and diagnosable.

What I would say: one quarter of one injury is encouraging but not yet a result, and "stop the project" is the highest-risk moment in any improvement effort. The gains were produced by attention - auditing, reminders, visible focus. Attention decays. If the improvement depends on the project being active, it will regress the month the project ends.

Data that would settle it:

1. A run chart or control chart across all six months, not a quarter-to-quarter comparison. Common-cause variation and a real shift look identical in two data points. I would look for a shift (six or more consecutive points below the baseline centerline) or a sustained level change, and check the standard rules before claiming special-cause improvement.
2. The process measures held at target without the audit pressure - reassessment ≥90% and repositioning documented ≥90% over the last two months. If the process slipped while the outcome stayed good, the outcome is probably noise or under-detection.
3. A case-finding check. Four injuries to one is exactly the pattern that can also mean less detection. I would verify that skin inspections at admission and discharge are still being done, so I am not celebrating a surveillance failure.
4. The balancing measures - nursing time, restraint use, patient-reported comfort, and whether documentation burden displaced something else.

What I would recommend instead of stopping: convert the interventions into standard work - embed the reassessment triggers in the EHR, put skin assessment in the shift handoff template, add the two process measures to the unit dashboard, and move the project to quarterly review. The goal is not to run a project forever; it is to make the improvement stop needing one.

The honest caveat: with an outcome this rare, one quarter genuinely cannot distinguish a good system from a lucky one. Saying so is not pessimism - it is the difference between a claim I can defend and a claim I will have to walk back.