Performance Improvement
Domain 9 - Performance Improvement
This domain tests whether you can turn surveillance data into measurable improvement and interpret variation correctly.
Core rules
- Model for Improvement: a clear aim, measures (outcome, process, balancing), and changes tested through PDSA cycles. Change one element at a time so the effect can be attributed.
- Bundles work because they are evidence-based, small, and measured together (CLABSI: hand hygiene, maximal barrier precautions, chlorhexidine skin prep, optimal site selection, daily line necessity review).
- FMEA is prospective - analyze a process before failure; RCA is retrospective - analyze an event that already occurred. Both should fix systems, not blame individuals.
- Statistical process control: distinguish common-cause from special-cause variation. Signals include one point beyond 3 sigma, a shift of 6 or more consecutive points on one side of the centerline, a trend of 5 points up or down, and repeating patterns.
- Benchmarking with NHSN SIR/SPR gives risk-adjusted comparison; interpret with confidence intervals and data-quality caveats.
Worked example
CLABSI rates rise from 0.8 to 2.1 per 1,000 line-days over two months. Reasoning: before changing practice, check the data - definition changes, more blood cultures ordered, or new case-finding can create an artifact. Then plot a control chart; three consecutive points above the upper control limit indicate special cause. Only then investigate line maintenance, dressing compliance, and staffing.
Traps
Reacting to a single above-average month (overreacting to common-cause variation); confounding a PDSA result by launching several changes at once; running an RCA without frontline staff; quoting an SIR without its confidence interval or using unstable low-volume rates; declaring success before gains are sustained over several months.