The Nurse Who Keeps Missing Sepsis Screens: A Systems Fix
You are the CNS for a 40-bed pediatric inpatient unit and a 6-bed pediatric ICU. Over the last quarter, the sepsis screening tool was completed within 1 hour of admission on only 62% of admissions (target 95%). Two patients had delayed recognition of sepsis; one required transfer to the PICU. Chart audit shows the screen is a manual field in the EHR, the admission workflow has 14 required clicks before it, and night-shift staffing is 1 nurse to 6 patients on the floor. Nurses tell you the screen 'fires for everyone' and they 'click through it.' The medical director wants mandatory re-education and a competency fair. Nursing leadership is worried about morale. You have 8 weeks before the next quality report.
1. What is your diagnosis of the problem, and how do you respond to the medical director's proposed fix?
Read the worked answer
This is a workflow and design problem, not a knowledge problem.
Evidence for that diagnosis:
- 62% completion with a manual field buried behind 14 clicks - the tool is competing with the admission workflow, not supporting it.
- 'Fires for everyone' / 'click through' - alert fatigue from poor specificity; the screen is not discriminating, so it trains staff to ignore it.
- 1:6 night staffing - the highest-risk time has the least slack; any tool that depends on spare attention will fail at night.
- Two delayed recognitions - the failure mode is exactly what you'd predict from the above.
Response to the medical director:
- Agree that the outcome is unacceptable; disagree that education is the lever. Re-education on a tool people already know how to use but can't complete in workflow will not move 62% to 95% - it will move it briefly and then regress, and it will cost morale.
- Propose a two-track plan: (1) fix the workflow/EHR design (primary), and (2) targeted education only where a genuine knowledge gap is demonstrated by audit.
- Bring data: show the click-count, the alert specificity, and the night-staffing pattern. Frame it as 'we agree on the goal; here is the evidence for the mechanism.'
Framing for leadership: this is a just-culture issue - the nurses are not careless; the system is set up so that the correct action is the hardest one at the busiest time.
2. Design the intervention. What do you change, in what order, and how do you know it worked?
Read the worked answer
Phase 1 (weeks 1–2): Make the right action the easy action.
- Automate the screen - pull vitals, temp, and risk factors from the EHR into a passive calculation; the nurse confirms rather than re-enters.
- Reduce clicks - move the screen to the top of the admission navigator, or trigger it automatically on admission and on any new set of vitals meeting criteria.
- Improve specificity - audit the false-positive rate; tune thresholds so the alert means something. An alert that fires for everyone is an alert that fires for no one.
- Hard-stop / soft-stop - a soft stop at admission that requires acknowledgment (not a free-text override) and a hard stop before certain order sets.
Phase 2 (weeks 2–4): Support the night.
- Charge nurse and rapid response as backstop - the screen result routes to the charge nurse if unacknowledged in 30 minutes.
- Staffing review - escalate 1:6 night ratio to leadership with the sepsis data; this is a legitimate safety concern, not a scheduling complaint.
- Just-in-time coaching at the bedside, not a classroom.
Phase 3 (weeks 4–8): Sustain and measure.
- Metrics: screen completion within 1 hour (primary), time-to-antibiotics in positive screens, false-positive rate, override reasons, and PICU transfers for sepsis.
- Run charts, not single numbers - 8 weeks is enough for a signal, not a verdict.
- Weekly huddle feedback to staff: 'here's what the data show, here's what we changed.'
- Listen to overrides - every free-text override is a design suggestion.
Success criteria: ≥90% completion by week 8 with a falling false-positive rate and no increase in override volume - and staff reporting the screen is easier, not just mandatory.
3. How do you handle the morale concern and the two nurses whose charts showed the delays?
Read the worked answer
Separate the system from the individual, explicitly and publicly.
On morale:
- Name it in the open. Tell staff: 'We asked you to do a task that was designed to fail at 3 a.m. We're fixing the task, not blaming you.' That single sentence does more for morale than any pizza.
- Invite nurses into the redesign - a staff nurse from nights, a charge nurse, and a newer graduate on the workgroup. Ownership beats compliance.
- Protect the feedback loop - share the run chart and credit the improvement to the unit.
On the two nurses:
- Do not scapegoat. Review their charts in context: were they on 1:6 nights? Was the screen buried? Did they override with a reason that was clinically reasonable?
- Distinguish system failure from genuine negligence. If there is evidence of reckless practice (e.g., ignoring a documented abnormal vital sign with no action), that goes to the manager through the proper just-culture route - privately, fairly, with due process.
- If it was system-driven, the response is support and coaching, not discipline. Disciplining people for a design flaw teaches the unit to hide errors.
- Close the loop with the affected patients' families per policy, and use the cases (de-identified) as learning, not as cautionary tales about individuals.
The CNS throughline: psychological safety is not soft - it is the mechanism by which you find out about the next near miss before it becomes a harm.