Quality, Safety, and Performance Improvement
Measure, analyze, change, re-measure
Quality work follows a loop: define the process, measure it, analyze variation, test a change, and confirm the result. Know the structure-process-outcome framework, and the tools: PDSA cycles, root cause analysis for a sentinel event, failure mode and effects analysis for a process that has not yet failed, flowcharts, Pareto charts, and run charts. Distinguish common-cause variation, the normal noise of a stable process, from special-cause variation, which signals that something changed.
In a revenue-cycle context, know the indicators and what they reveal: clean claim rate, denial rate broken down by reason, days in accounts receivable, first-pass resolution rate, coding accuracy from internal audits, and patient collection ratios. Understand benchmarking and why comparisons must be made between like populations and like service mixes.
Traps
- Blaming an individual for a system failure; a just culture separates human error, at-risk behavior, and reckless behavior.
- Confusing a process measure (was the check done?) with an outcome measure (did harm occur?).
- Reacting to one or two data points as if they were a trend.
- Auditing only the coders you already trust, or only the claims that paid.
- Choosing a solution before completing the analysis.
- Reporting a safety event verbally and never entering it in the reporting system.
Rule of thumb
A run of six or more points on one side of the median, or an obvious sustained shift, is your signal to investigate. One outlier is not a trend.
Worked example: Denials for missing authorization rise from 4 percent to 11 percent over five consecutive months. Rather than retraining the whole staff, you stratify by payer and find the increase is concentrated in two plans that changed their portal requirements. That is a special-cause finding pointing to a process fix, not a people problem.