Abstract
BACKGROUND In concussed athletes, decline in neuropsychological test performance from preseason baseline levels may be cited as evidence that the athlete's brain function has not returned to normal after the injury. Despite the pivotal medical management role that neuropsychological testing can play in such circumstances, the psychometric properties such tests require for accurate detection of cognitive impairments are not well understood. PURPOSE To examine the psychometric properties required for accurate detection of post-concussion impairments in neuropsychological test performance. METHODS 427 athletes performed computerized neuropsychological testing prior to the 2002 Australian football season. Change from baseline to post-concussion in 10 athletes who were symptomatic at the time of post-injury testing are reported. Reliable change indices (RCI) were calculated using a reaction time task requiring 75 responses per athlete. These scores were then re-calculated after selecting 5, 10, 20, 30 and 50 responses at random from both baseline and post-concussion test data. Group mean and individual RCI scores are reported under each of these conditions. RESULTS Highly variable results were obtained when few (5 & 10) responses were included in the RCI calculation. Changes observed under these conditions were inconsistent with the athletes clinical status at the time of testing. These inconsistencies resolved as a greater number of responses were included in the RCI calculation. CONCLUSION Accurate estimates of performance on neuropsychological tests of response time require that at least 20 responses be recorded in every athlete at both baseline and post-concussion evaluations. The use of neuropsychological tests with fewer required responses may lead to erroneous clinical decisions.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call