A new optimization combination used in the exponentially weighted moving average quality control of electrolyte programs

Objective

This study aimed to optimize and validate the EWMA QC procedure for potassium (K), sodium (Na), and chloride (Cl) by selecting optimal exponential weighting factors (λ), quality control limits (CLs), and truncation limits (TLs).

Methods

Data from 150,000 serum tests for K, Na, and Cl were obtained from the Laboratory Medicine Department of the First Affiliated Hospital of Ningbo University. The dataset included inpatient, outpatient, and physical examination populations. Using testing times, 400 simulated days with 375 tests each were created. The EWMA QC process was optimized and verified across the three analytes using false positive rate (FPR), true positive rate (TPR), and the median number of patient results affected before error detection (MNped).

Result

The EWMA QC procedure with λ = 0.2 outperformed other λ values for Na and Cl, achieving higher sensitivity for errors exceeding the total allowable error (TEa) and better performance with λ = 0.5 for errors below 1/2TEa in K. All three programs maintained FPRs below 5 %. The optimal QC (λ = 0.2) detected 99.93 % of Na errors and 86.30 % of Cl errors. However, K’s QC showed lower sensitivity, with poor performance even at 2TEa.

Conclusion

The EWMA QC method demonstrated high sensitivity and specificity for Na and Cl programs with an optimal λ of 0.2. However, its performance for K was suboptimal, indicating limited applicability in this analyte.