Statistical Methods for Evaluation of Biomarkers

Eunice Kennedy Shriver National Institute Of Child Health & Human Development

Investigators

Linked publications, trials & patents

Abstract

In the past year investigators and staff on this project proposed a novel robust-efficient method for evaluation of biomarkers based on group testing data. They addressed a very challenging problem of estimating the diagnostic accuracy of a continuous biomarker (e.g., C-reactive protein) in predicting a disease (e.g., chlamydia) when the disease status are only available at group levels, either from a group testing design or from grouping of individual disease status to protect confidentiality. The problem is challenging in that individual disease statuses are not observed and testing results are often subject to misclassification, with further complication that the misclassification may be differential as the group size and the number of the diseased individuals in the group vary. Dr. Liu and his collaborators showed that the nonparametric likelihood function is not identifiable with respect to the unknown prevalence and distributions, and went on to propose a shape-restricted two-step procedure to construct nonparametric estimation of the distribution of the biomarker in the diseased and non-diseased population and obtain its asymptotic properties. Utilizing the estimated distributions, they obtained estimators for the ROC curve and its area, and established their asymptotic distributions. Under various design considerations concerning group sizes and classification errors, they demonstrated that the distribution and ROC curve estimators constructed from group-based data are more efficient than that constructed from individual testing data. The method is exemplified with data from the National Health and Nutrition Examination Survey study to estimate the distribution and diagnostic accuracy of C-reactive protein in blood samples in predicting chlamydia incidence, and is highly expected to be useful in studying predictive biomarkers for COVID-19 infection and immunity.