Module on Diagnostic Research
Stellenbosch University, South Africa
4 - 7 March 2009
Course Description [Link]
Diagnosis is central to medical practice. It refers to knowing (“gnosis”) about the health of a client. Typically, diagnostic research focuses on estimating the sensitivity and specificity of individual diagnostic tests, their predictive values, and other parameters of interest (such as likelihood ratios, ROC curves, test reliability). Accuracy measures are then used to generate posttest probabilities, given a set of pretest probabilities. This module will cover the basic principles behind diagnostic research, diagnostic study designs, sources of bias, and analysis and interpretation of diagnostic accuracy data. Also, critical appraisal of diagnostic studies, and conventional and advanced methods for systematic reviews (meta-analyses) of diagnostic tests will be presented.
More recently, there is growing appreciation that “test research” focused on sensitivity and specificity is not necessarily the same as “diagnostic research.” Modern methods of diagnostic research focus on a process by which a patient’s test result can and should be considered in the context of other patient characteristics or test results. These methods enable direct estimation of individual probabilities of disease presence based on all diagnostic information and evaluation of the extent to which a particular diagnostic test has added value in the clinical context (i.e. incremental or added value analysis). This course will briefly introduce this multivariable approach to diagnostic research.
Lecture 1: Diagnostic Research - An Introductory Overview [Link]
Lecture 2: Bias in Diagnostic Research [Link]
Lecture 3: Measuring Reliability and Agreement [Link]
Lecture 4: Reporting of Diagnostic Accuracy Studies [Link]
Lecture 5: Diagnostic Research - Multivariable Approach [Link]
Lecture 6: Introduction to Systematic Reviews [Link]
Lecture 7: Systematic Reviews of Diagnostic Test Accuracy [Link]