Chunbao Xie, Changrui Sun, Li Xiao, Tao Chen, Fengxi Jia, Fangkui Xiong, Zehong Wang, Yulian Zhou, Hengyong Yu, Huan Ding, Yifeng Huang, Xiaoting Zhan, Guangcheng Luo, Jiawei Zeng, Yi Huang
This multi-center quality-indicator study provides a reproducible, denominator-consistent framework for benchmarking venous blood specimen rejection across multiple centers. Percentile-based benchmarks and standardized reason spectra may support local implementation and transferable Six Sigma monitoring in healthcare systems with heterogeneous resources.
BACKGROUND: Pre-analytical venous blood specimen rejection can delay clinical decision-making and may increase laboratory workload, yet multicenter benchmarking remains limited by inconsistent rejection definitions and denominator selection. This study aimed to support denominator-consistent benchmarking, transferable implementation, and quality monitoring across heterogeneous-resource settings.
METHODS: We conducted a retrospective quality-indicator study across 10 hospitals in Western China using routine venous blood specimen data collected from January 1, 2023, to April 30, 2024. Rejection reasons were standardized with a prespecified mapping dictionary. Center-level rejection rates, defects per million opportunities (DPMO), and Six Sigma values were calculated using monthly venous specimen volume as the primary denominator. Percentile-based benchmarks, pooled estimates, and a month-adjusted random-intercept generalized linear mixed-effects model were used to assess performance and residual between-center heterogeneity.
RESULTS: Among 21,319,200 venous specimens, 42,060 were rejected, yielding a pooled rejection rate of 0.197%. The leading standardized rejection categories were clot/aggregation, insufficient or incorrect volume, and container/tube error. Rejection patterns varied across centers and clinical settings: inpatient and emergency events were dominated by clotting and volume problems, whereas outpatient and physical examination events more often showed order, timing, or identification errors. After temporal adjustment, measurable between-center heterogeneity persisted (intraclass correlation coefficient = 0.045).
CONCLUSION: This multi-center quality-indicator study provides a reproducible, denominator-consistent framework for benchmarking venous blood specimen rejection across multiple centers. Percentile-based benchmarks and standardized reason spectra may support local implementation and transferable Six Sigma monitoring in healthcare systems with heterogeneous resources.