A multi-center research on the establishment and validation of autoverification rules for blood analysis
10.3760/cma.j.issn.1009-9158.2018.08.009
- VernacularTitle: 血液分析自动审核规则建立与验证的多中心研究
- Author:
Wei XU
1
;
Xiaoke HAO
;
Wei CUI
;
Hong JIANG
;
Xuefeng WANG
;
Chenxue QU
;
Lei ZHENG
;
Yandan DU
;
Linlin QU
;
Enliang HU
;
Jianbiao WANG
;
Zhigang MAO
;
Lingling LIU
;
Cuiling ZHENG
;
Dehua SUN
;
Chengwei PU
;
Chunxi BAO
;
Li LING
;
Qiang LI
;
Tan LI
Author Information
1. Department of Clinical Laboratory, the First Hospital of Jinlin University, Changchun 130021, China
- Publication Type:Journal Article
- Keywords:
Autoverification rules;
Blood analysis;
Validation;
Turnaround time;
Miss rate
- From:
Chinese Journal of Laboratory Medicine
2018;41(8):601-607
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To establish a set of rules for autoverification of blood analysis, in order to provide a way to validate autoverification rules for different analytical systems, which can ensure the accuracy of test results as well as shorten turnaround time (TAT) of test reports.
Methods:A total of 34 629 EDTA-K2 anticoagulated blood samples were collected from multicenter cooperative units including the First Hospital of Jinlin University during January 2017 to November 2017. These samples included: 3 478 cases in Autoverification Establishment Group, including 288 cases for Delta check rules; 5 362 cases in Autoverification Validation Group, including 2 494 cases for Delta check; 25 789 cases in Clinical Application Trial Group. All these samples were analyzed for blood routine tests using Sysmex XN series automatic blood analyzers.Blood smears, staining and microscopic examination were done for each sample; then the clinical information, instrument parameters, test results and microscopic results were summarized; screening and determination of autoverification conditions including parameters and cutoff values were done using statistical analysis. The autoverification rules were input into Sysmex Laboman software and undergone stage Ⅰ validation using simulated data, and stage Ⅱ validation for post-analytical samples successively. True negative, false negative, true positive, false positive, autoverification pass rate and passing accuracy were calculated. Autoverification rules were applied to autoverification blood routine results and missed detection rates were validated, and also data of autoverification pass rate and TAT were obtained.
Results:(1)The selected autoverification conditions and cutoff values included 43 rules involving WBC, RBC, PLT, Delta check and abnormal characteristics. (2)Validation of 3 190 cases in Autoverification Establishment Group showed the false negative rate was 1.94%(62/3 190)(P<0.001), autoverification pass rate was 76.74%, passing accuracy was 97.47%; Validation of 2 868 cases in Autoverification Validation Group, the false negative rate was 3.38%(97/2 868)(P=0.002), autoverification pass rate was 42.26%, passing accuracy was 92.00%; Validation of Delta check on 288 cases in Autoverification Establishment Group and 2 494 cases in Autoverification Validation Group showed the false negative rates were respectively 1.39% and 2.61%(P<0.001). (3)Three hospitals adopted these rules of autoverification for 25 789 blood routine samples, and found that the average TAT of blood routine test reports were shortened by 24min, 32min and 7min respectively, the rate of samples reported within 30min were elevated by 33%, 53% and 7%. The autoverification pass rates were 72%-74%.
Conclusions:The application of this set of 43 autoverification rules in blood sample analysis can ensure test quality while shortenTAT and improve work efficiency. It is worth pointing out that for the same analytical systems in this research, validation is necessary before application of this set of rules, and periodic validation is required during application to make necessary adjustment; for different analytical systems, as this research provide a way to establish autoverification rules for blood routine tests.Clinical labs may establish their own suitable autoverification rules on the basis of technological parameters. (Chin J Lab Med, 2018, 41: 601-607)
