ABSTRACT
In this chapter, we focus on assessing diagnostic and screening instruments. Such tools are commonly used in clinical and research studies, ranging from physical health to psychological well-being to social functioning. Since early detection of disease often leads to less suffering and speedier recovery while false diagnosis may unnecessarily expose individuals to potentially harmful treatments (Bach et al., 2007), it is important to assess their accuracies so that informative decisions can be made. When the true status is available, we can assess the accuracy by comparing the instrument directly with the known true status. The receiver operating characteristic (ROC) curve is commonly used when the true status is binary such as presence and absence of a disease. For a continuous or ordinal outcome with a wide range, we may use concordance correlation coefficients.
