Evaluation of a Six Sigma‐Based Dynamic Quality Control Strategy for Hematology Analysis: A Multicenter Study

TL;DR

This study introduces a dynamic quality control strategy for hematology testing using Six Sigma, moving average, and LSTM methods to improve accuracy and reliability across multiple labs.

Contribution

A novel dynamic QC strategy combining Six Sigma, MA, and LSTM for real-time, intelligent hematology testing quality control.

Findings

Hb and WBC achieved world-class performance (σ ≥ 6), while PLT showed lower stability.

The MA–LSTM approach improved error detection sensitivity and reduced false positives compared to traditional QC methods.

The dynamic model enabled real-time monitoring and adaptability across multiple laboratory sites.

Abstract

Quality control (QC) is critical for ensuring the accuracy and reliability of hematology testing. Traditional QC strategies, however, are often limited in their ability to provide timely detection of analytical errors and to adapt to complex, real‐world laboratory conditions.

In this multicenter study, we applied the Six Sigma quality management framework to systematically evaluate the performance of five hematology parameters (Hb, WBC, RBC, HCT, and PLT). To enhance QC monitoring, we established a dynamic quality control strategy that integrates moving average (MA) monitoring with a long short‐term memory (LSTM) predictive model. Patient sample data were incorporated alongside routine QC data to validate clinical adaptability.

Sigma metrics revealed marked performance differences among the parameters, with Hb and WBC achieving world‐class or excellent performance (σ ≥ 6), while PLT showed relatively lower stability. The combined MA–LSTM approach significantly improved sensitivity for error detection while reducing false positives compared with conventional rule‐based QC. The dynamic model demonstrated robust predictive ability, enabling real‐time QC monitoring across multiple laboratory sites.

By combining Six Sigma evaluation, MA monitoring, and LSTM modeling, we propose a dynamic QC strategy that overcomes key limitations of conventional quality control methods. This approach provides laboratories with an intelligent, proactive, and clinically adaptable solution for improving the reliability of hematology testing and ensuring higher quality patient care.

This study employed the Six Sigma quality management framework to evaluate the performance of hematology parameters across multiple centers and established a dynamic quality control strategy by integrating moving average (MA) and long short‐term memory (LSTM) methods. The proposed approach enhanced error detection sensitivity, reduced false positives, and enabled more intelligent, real‐time laboratory quality management.

Figures