Determining clinically meaningful cognitive impairment following traumaticbrain injury

Aims. To study the reliability and validity of ratings of neuropsychologica l impairment with test data from traumatically brain-injured patients. Methods. Neuropsychological test results from 66 traumatically brain-injure d adults and 27 orthopaedic controls were rated for neuropsychological impa irment by an experienced neuropsychologist and three undergraduate students provided with brief training in test interpretation. Ratings were based on the discrepancy between a patient's current cognitive level and their esti mated premorbid cognitive level. Decision making rules were utilised in mak ing the ratings. The raters, who were blind with respect to the patients di agnostic group membership, independently rated test results. Test results w ere rerated using the same method approximately two weeks later. Results. The ratings of novice raters were in good to very good agreement w ith the ratings of an expert. All raters evidenced very good to excellent t est-retest reliability. Higher rates of neuropsychological impairment were found in the traumatically brain-injured group than in the orthopaedic cont rol group. Discriminant function analysis suggested that raters employed in formation from all neuropsychological measures (with the exception of curre nt intellectual level), in making their overall ratings of neuropsychologic al impairment. Conclusion. These findings provide encouraging preliminary evidence in supp ort of the reliability and validity of individual case-based ratings of neu ropsychological impairment. The high false positive rate in the control sam ple may reflect the relatively low specificity of neuropsychological impair ment.