A simple signaling rule for variable life-adjusted display derived from an equivalent risk-adjusted CUSUM chart

Area of Science:

• Medical Statistics
• Surgical Quality Improvement
• Health Services Research

Background:

• The variable life-adjusted display (VLAD) is a popular graphical tool for monitoring surgeon performance.
• VLAD plots the cumulative sum of expected minus observed deaths, but its increasing variance with sample size can lead to misinterpretation.
• Existing signaling rules for VLAD are often complex, hindering practical application.

Purpose of the Study:

• To establish an equivalence between a VLAD with a V-mask and a risk-adjusted cumulative sum (RA-CUSUM) chart.
• To develop a simple and effective signaling rule for VLAD to improve surgeon performance monitoring.
• To provide a practical design procedure for V-mask parameters using resampling.

Main Methods:

• Establishing an equivalence between VLAD with V-mask and a specific RA-CUSUM chart.
• Utilizing average run length analysis via simulation to compare performance.
• Developing a resampling approach for V-mask parameter estimation from real data.

Main Results:

• The proposed VLAD with V-mask demonstrates performance comparable to established RA-CUSUM charts.
• A simple design procedure for V-mask parameters was developed using resampling.
• The method was illustrated by monitoring a real surgeon's performance.

Conclusions:

• VLAD with a V-mask offers a more reliable and interpretable method for surgeon performance monitoring.
• The developed signaling rule simplifies the use of VLAD, mitigating misinterpretation risks.
• Resampling provides an effective way to determine appropriate V-mask parameters for real-world application.