Analysis and Improvement: From Event to Durable System Change
This domain tests whether you can investigate harm rigorously, choose the right analytic method, and convert findings into actions strong enough to prevent recurrence. Expect items that contrast reactive and proactive methods and that rank corrective actions.
Core rules and concepts
- RCA² is for events that reached the patient and caused serious harm: multidisciplinary, systems-focused, completed within roughly 45 days, ending in an action plan with named owners, due dates, and follow-up measurement.
- FMEA is prospective: map the process, list failure modes, score severity × occurrence × detection for a risk priority number, and act on the highest risks before harm occurs.
- Model for Improvement: three questions (What are we trying to accomplish? How will we know a change is an improvement? What changes can we make?) plus PDSA cycles. Test small, then implement and spread.
- Action hierarchy: strong (forcing functions, standardization, simplification, automation), intermediate (checklists, memory aids, double checks, staffing), weak (education, warnings, new policy).
- Supporting tools: fishbone, 5 whys, Pareto, driver diagram, process map, control chart.
- Safety-I learns from things that go wrong; Safety-II studies why things usually go right. Resilience needs both.
Worked example
A wrong-site nerve block: the RCA finds the consent form listed left, the site mark was made in the holding area and then hidden by drapes, and the time-out occurred before the final imaging was displayed. Weak actions are retraining staff and reissuing policy. Strong actions: mark the site after final imaging with the patient awake and involved, require imaging to be displayed at time-out, add a scheduling hard stop when laterality fields conflict, and audit time-out fidelity monthly on a control chart.
Traps
- RCA reports that end with staff will be re-educated.
- Reaching for FMEA after an event - that is RCA's job; FMEA is proactive.
- Action plans with no owner, due date, or measurement of effect.
- Hunting a single root cause when harm is almost always multifactorial.
- Blaming individuals for behavior the system shaped and rewarded.