Search

Objectives: Preanalytical processes in pediatric patients are generally manual and associated with a higher risk of error. The optimized delta check rules for detecting misidentified children samples are examined.Methods: Relative difference and absolute different delta check limits were applied on original and reshuffled (to simulate sample mislabeling/mix-up) paired deidentified pediatric results of 57 laboratory tests. The sensitivity, specificity, and accuracy of a range of delta check limits were determined. The delta check limit associated with the highest accuracy was considered optimal.Results: In general, the delta check limits had poor to moderate accuracy (0.50-0.81) in detecting misidentified patient samples. The sensitivity (rule out misidentified sample) quickly deteriorated at increasing delta check limits. At the same time, the specificity (rule in misidentified sample) of the delta check limit was also low. The performance of the relative difference and absolute difference delta check rules was similar.Conclusions: Our findings showed poor delta check performance in the pediatric population. The high false-positive flag rate may lead to wasteful resource-intensive investigations and delay in result reporting. In addition, we observed that the optimized pediatric delta check correlated strongly with within-subject biologic variation, whereas delta check accuracy correlated poorly with index of individuality. [ABSTRACT FROM AUTHOR]

Objectives: An unfortunate reality of laboratory medicine is that blood specimens collected from one patient occasionally get mislabeled with identifiers from a different patient, resulting in so-called "wrong blood in tube" (WBIT) errors and potential patient harm. Here, we sought to develop a machine learning-based, multianalyte delta check algorithm to detect WBIT errors and mitigate patient harm.Methods: We simulated WBIT errors within sets of routine inpatient chemistry test results to develop, train, and evaluate five machine learning-based WBIT detection algorithms.Results: The best-performing WBIT detection algorithm we developed was based on a support vector machine and incorporated changes in test results between consecutive collections across 11 analytes. This algorithm achieved an area under the curve of 0.97 and considerably outperformed traditional single-analyte delta checks.Conclusions: Machine learning-based multianalyte delta checks may offer a practical strategy to identify WBIT errors prior to test reporting and improve patient safety. [ABSTRACT FROM AUTHOR]

Objectives: To assess the utility of reference change values (RCVs) as delta check limits.Methods: A total of 1,650,518 paired results for 23 general chemistry test results from June 1, 2014, to October 31, 2016, were analyzed. The RCVs for each analyte were calculated from the analytical imprecision and within-subject biological variation. The percent differences between two consecutive results in one patient were categorized into one of four groups: outpatients, inpatients, emergency care, and general health care. For each, 2.5th and 97.5th percentile values were computed and compared with their RCVs. The distributions were assessed for normality using the Kolmogorov-Smirnov test.Results: Most of the estimated limits were larger than the corresponding RCVs and, furthermore, with notable differences across the groups. Patients in the emergency care group usually demonstrated larger delta percent values than those in the other groups. None of the distributions of the percent differences passed tests of normality when subjected to Kolmogorov-Smirnov analysis.Conclusions: Comparison of estimated RCVs and real-world patient data revealed the pitfalls of applying RCVs in clinical laboratories. Laboratory managers should be aware of the limitations of RCVs and exercise caution when using them. [ABSTRACT FROM AUTHOR]