[Accuracy evaluation of diagnostic test in laboratory]

Area of Science:

• Biostatistics
• Medical Diagnostics
• Immunology

Context:

• Hepatitis B virus (HBV) infection diagnosis relies on accurate serological markers.
• Evaluating the performance of diagnostic tests is crucial for clinical decision-making.
• Receiver Operating Characteristic (ROC) curve analysis is a standard method for assessing diagnostic accuracy.

Purpose:

• To introduce the fundamental theory and practical applications of ROC curves.
• To detail the calculation and interpretation of key ROC-related indices: area under the curve (AUC), partial area under the curve (pAUC), and sensitivity at a specific false-positive rate [Se(FPR=e)].
• To apply these indices to real-world data from five hepatitis B surface antigen (HBsAg) detection kits, employing the bootstrap method for robust statistical inference.

Summary:

• The study provides a concise overview of ROC curve theory and its applications in diagnostic test evaluation.
• It demonstrates the practical calculation and statistical inference of AUC, pAUC, and Se(FPR=e) using real data from HBsAg kits.
• The bootstrap method is utilized to ensure the reliability of the statistical analysis of these performance indices.

Impact:

• Offers a clear methodology for comparing the diagnostic performance of hepatitis B testing kits.
• Provides valuable insights into the utility of AUC, pAUC, and Se(FPR=e) for evaluating diagnostic assays.
• Enhances understanding of statistical inference techniques, such as bootstrapping, in the context of diagnostic accuracy assessment.