Performance Evaluation of Qualitative Proficiency Testing of Blood Groups Examination Using K-means Clustering Method
Keywords:
qualitative proficiency testing, blood groups examination, K-means clusteringAbstract
Statistical methods to evaluate performance of laboratories participating in qualitative proficiency testing programs are not clearly recommended in ISO 13528: 2015. Assessment by the experts is generally used in many qualitative PT programs. We, therefore, studied a K-means clustering method in conjunction with a scistatcalc program as an alternative tool for assessment PT results of ABO blood group examination. Using the K-means method, 1,102 participating laboratories could be classified into 3 performance levels. The numbers of laboratories with good, satisfied and non-satisfied levels are 953 (86.48%), 74 (6.72%) and 75 (6.81%) laboratories, respectively. In addition, evaluation results as determined by K-means were in parallel with those of the experts (Kappa = 1.00). Furthermore, the K-means method showed that laboratories located in Health Districts 6, 8 and 12 were mostly critical for quality improvement on blood groups examination.
References
ISO/IEC 17043:2010. Conformity assessment-general requirements for proficiency testing. Geneva: International Organization for Standardization; 2010.
ISO 13528:2015. Statistical methods for use in proficiency testing by interlaboratory comparisons. Geneva: International Organization for Standardization; 2015.
Uhlig S, Bläul C, Frost K, Sgorzaly S, Colson B, Simon K. Qualitative PT data analysis with easy-to-interpret scores. Accred Qual Assur 2015; 20(5): 347-53.
Bläul C, Uhlig S. Qualitative PT data analysis with easy-to-interpret scores. [online]. 2014; [cited 2019 Jun 7]; [11 screens]. Available from: URL: https://www.eurachem.org/images/stories/workshops/2014_10_PT/pdf/ECHPT2014_O04_Blaeul_Uhlig.pdf.
Ali HH, Kadhum LE. K-means clustering algorithm applications in data mining and pattern recognition. [online]. 2015; [cited 2019 Jun 7]; [8 screens]. Available from: URL: https://pdfs.semanticscholar.org/a430/da239982e691638b7193ac1947da8d0d241b.pdf.
Cai F, Le-Khac NA, Kechadi MT. Clustering approaches for financial data analysis: a survey. [online]. 2016; [cited 2019 Jun 7]; [7 screens]. Available from: URL: https://pdfs.semanticscholar.org/3a06/1c016599c29e27bc97528c84f6a6e4e54984.pdf.
Chakraborty S, Nagwani NK, Dey L. Weather forecasting using incremental k-means clustering. [online]. 2014; [cited 2019 Jun 7]; [6 screens]. Available from: URL: https://arxiv.org/ftp/arxiv/papers/1406/1406.4756.pdf.
Oyelade OJ, Oladipupo OO, Obagbuwa IC. Application of k-means clustering algorithm for prediction of students’ academic performance. [online]. 2010; [cited 2019 Jun 7]; [4 screens]. Available from: URL: https://arxiv.org/ftp/arxiv/papers/1002/1002.2425.pdf.
Patil R, Deshmukh S, Rajeswari K. Analysis of simple k-means with multiple dimensions using WEKA. [online]. 2015; [cited 2019 Jun 7]; [4 screens]. Available from: URL: https://pdfs.semanticscholar.org/759a/988f38b4ea3465f5747442e6c642fc672598.pdf.
Scistatcalc. K-means clustering calculator. [online]. 2014; [cited 2019 Jun 7]; [1 screens]. Available from: URL: https://scistatcalc.blogspot.com/2014/01/k-means-clustering-calculator.html.
อาริดา กล่อมเกลี้ยง. คู่มือสมาชิกการประเมินคุณภาพการตรวจวิเคราะห์ สาขาธนาคารเลือด ประจำปีงบประมาณ พ.ศ.2561. นนทบุรี: สำนักมาตรฐานห้องปฏิบัติการ กรมวิทยาศาสตร์การแพทย์; 2561.
วีระยุทธ พิมพาภรณ์, พยุง มีสัจ. การเปรียบเทียบประสิทธิภาพการจัดกลุ่มข้อมูล โดยวิธีการเลือกลักษณะสำคัญแบบพลวัตเพื่อเพิ่มประสิทธิภาพของอัลกอริทึม การจัดกลุ่มบนปริภูมิย่อย. วารสารเทคโนโลยีสารสนเทศ 2557; 10(2): 43-51.
GraphPad. Quantify agreement with kappa. [online]. [cited 2019 Jun 7]; [1 screens]. Available from: URL: https://www.graphpad.com/quickcalcs/kappa1/.
Social Science Statistics. Chi-square calculator. [online]. 2019; [cited 2019 Jun 7]; [1 screens]. Available from: URL: https://www.socscistatistics.com/tests/chisquare/default2.aspx.
