Methodologic Issues in Neuropsychological Testing

Abstract

OBJECTIVE: To familiarize athletic trainers with methodologic issues regarding the development and implementation of neuropsychological tests used in programs for monitoring sport-related cerebral concussion. DATA SOURCES: Knowledge base and MEDLINE and PsychLit searches from 1980-2000 using the terms sports, athletes, concussion, and brain. DATA SYNTHESIS: Neuropsychological testing is a proven method for evaluating symptoms of concussion that results from a variety of different causes. These tests have been shown to be effective in evaluating symptoms of subtle cognitive dysfunction in a number of patient groups. Applying these tests in an athletic population has required some procedural modifications, including the use of brief test batteries, collection of preseason baseline data, and evaluation of subtle postconcussive changes in test scores over time. New methods are now being used for improved evaluation of the reliability and validity of neuropsychological tests in athletes. Proper scientific analysis of the psychometric properties of neuropsychological tests and the ultimate value of their use in the sport setting will require years of detailed study on large numbers of athletes with and without symptoms of concussion. CONCLUSIONS/RECOMMENDATIONS: Athletic trainers and related personnel need to be aware of the training and methodologic issues associated with neuropsychological testing. Knowledge of the scientific properties of these tests, their advantages, and current limitations will ultimately enhance the athletic trainer's ability to use information from neuropsychological testing in an effective manner.

Receiver operating characteristic curve for hypothetical tests A and B developed for assessment of concentration disturbance secondary to sport-related concussion. Sensitivity and specificity values for classification of athletes with and without concussion are represented graphically for both tests. The larger area under the curve indicates that test A provides more accurate classification of these 2 athlete groups. This finding can be demonstrated empirically through statistical comparison of area-under-the-curve values for both tests. Such information would be used to establish that test A is a more valid and sensitive measure for assessing the effects of concussion than test B.