Quality Improvement and Patient Safety
Expect methods and measurement: how to prove an improvement is real, how to build a safety culture, and how to respond when harm occurs.
Improvement methods
The Model for Improvement sets an aim, measures and changes, tested in small PDSA cycles. DMAIC suits structured projects, Lean targets waste, and HFMEA prospectively examines a process before change. Measure process, outcome and balancing measures so fixing one thing does not quietly break another.
Reading variation
Run charts and control charts separate common-cause from special-cause variation. Signals include a point beyond a control limit, eight consecutive points on one side of the centerline, six consecutive trending points, and two of three points beyond two sigma.
Safety culture
Just culture separates human error (console), at-risk behavior (coach) and reckless behavior (sanction). Event reporting should be voluntary and nonpunitive and kept separate from the medical record. Root cause analysis follows harm; FMEA precedes change. TeamSTEPPS, SBAR and briefs, huddles and debriefs support communication.
Worked example
Monthly infusion pump PM compliance reads 88, 90, 89, 91, 92, 93, 94, 95, 96 percent. The six consecutive increases form a trend signal, so investigate what changed, such as dedicated PM days, then keep measuring before declaring the gain sustained.
Traps
- Reacting to common-cause noise as if it were a real signal.
- Reporting only averages, hiding a fleet still at 60 percent.
- Treating a rising number of reports as a worsening safety score; it often means a healthier reporting culture.
- Closing an improvement project before verifying effectiveness.