Diagnostic tests are the tests that clinicians use to make a diagnosis, i.e., to determine whether or not a person has a particular disease. In gerontology, for example, an array of diagnostic tests have been developed to identify persons with Alzheimer's Disease, and to track the stages of this progressive disease to determine what kinds of intervention are most likely to be helpful. Other tests are used to distinguish between normal cognitive changes and dementia, or to diagnose acute stroke, age-related macular degeneration, age-related cognitive impairment, osteoporosis, aspiration pneumonia in the elderly, coronary artery disease, and delirium among aged hospital patients. Since diagnostic tests are rarely 100% accurate, it is critically important that we know how much faith we can have in the results of a test. The process used to determine the accuracy of a diagnostic test is similar to the process used to determine the utility of a drug. When a test is initially developed it is evaluated in a series of studies, and researchers will publish papers detailing the performance of the test in these studies. As a body of evidence accumulates, ? experts will identify the relevant studies and perform a meta-analysis of the data. If the test performance is consistent across the studies, this analysis would provide a more precise estimate of the common accuracy than any of the studies alone. More often, the test performance will vary from one study to the next, and the goal of the meta-analysis would be to determine what factors are responsible for this variation. This could lead to a finding that the test is accurate in some settings but not others, or with some patient groups but not others. The goal of this project is to develop software for meta-analysis of diagnostic tests. The program, which will include modules for data-entry, analysis, and graphics, will be developed in collaboration with some of the leading experts in this field. ? ? ? ?
