An interactive website for analytical method comparison and bias estimation

• Published in: Clinical biochemistry Vol. 50; no. 18; pp. 1025 - 1029
• Main Authors: Bahar, Burak, Tuncel, Ayse F., Holmes, Earle W., Holmes, Daniel T.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.12.2017
• Subjects: Bias; Chemistry Techniques, Analytical - methods; Clinical chemistry; Clinical Laboratory Services; Data Accuracy; Humans; Internet; Method comparison; Programming Languages; R programming; Regression Analysis; Reproducibility of Results; Shiny; Software; Website; Method comparison; Website; Clinical chemistry; Shiny; R programming
• DOI: 10.1016/j.clinbiochem.2017.08.008
• ISSN: 0009-9120

Regulatory standards mandate laboratories to perform studies to ensure accuracy and reliability of their test results. Method comparison and bias estimation are important components of these studies. We developed an interactive website for evaluating the relative performance of two analytical methods using R programming language tools. The website can be accessed at https://bahar.shinyapps.io/method_compare/. The site has an easy-to-use interface that allows both copy-pasting and manual entry of data. It also allows selection of a regression model and creation of regression and difference plots. Available regression models include Ordinary Least Squares, Weighted-Ordinary Least Squares, Deming, Weighted-Deming, Passing-Bablok and Passing-Bablok for large datasets. The server processes the data and generates downloadable reports in PDF or HTML format. Our website provides clinical laboratories a practical way to assess the relative performance of two analytical methods. •We built a website (https://bahar.shinyapps.io/method_compare/) by R language for method comparison and bias estimation.•Available regression models: Ordinary Least Squares, Deming and Passing-Bablok•Available plots: scatter plot, coefficient plot and Bland-Altman plot•Generated reports can be download in PDF or HTML format.