New methods for efficient assessment Lead Investigator: Niels Smits Institution : University of Amsterdam E-Mail : n.smits@uva.nl Proposal ID : 815 Proposal Description: The diagnosis of dementia can be defined as the process of determining whether a patient suffers from dementia. A single test is not sufficient because for diagnosing dementia, multiple attributes play a role each of these attributes needs to be assessed. During the neuropsychological examination a set of cognitive test used to evaluate cognitive functions such as memory, concentration, etcetera. The assessment session is long and it may be exhausting, increasing the probability of fallacious answers, random responding or premature termination. To solve this problem, a diagnostic algorithm will be built that optimizes both the number of tests presented to a patient and the number items within each test. With this algorithm an optimal sequence of tests is proposed for each patient. After the administration of one test, the algorithm uses the patient???s previous scores to select the next, most informative test. After each test,information about the uncertainty of the diagnosis is established, and if this uncertainty is low enough the assessment is stopped and the appropriate diagnosis is given. In this study the requested data will be used to develop the algorithm. So new methodological technology will be developed.