Research examining the neurobehavioral outcome after mild head injury has yielded inconsistent and contradictory findings. Such findings have been attributed to a variety of methodological weaknesses, such as failure to consider the preinjury status of the patient, lack of control groups, and variability in outcome time points. However, few researchers have examined the adequacy of the current diagnostic criteria. A Glasgow Coma Scale (GCS) score of 13 to 15 is one of the primary criteria in the classification of mild head injury. We propose that the use of GCS Scores 13 to 15 permits excessive heterogeneity in injury severity and contributes to variability in neurobehavioral outcome. The purpose of this study is to examine the relationship of admission GCS scores to variables indicative of injury severity. The case records of 3370 patients consecutively admitted to a Level I trauma center with nonmissile head injuries, positive loss of consciousness, and admission GCS scores of 13 to 15 were reviewed. The frequency of positive computed tomographic scan findings and the need for neurosurgical intervention within the first 24 hours were recorded. A chi 2 analysis revealed statistically significant differences between the frequency of positive computed tomographic scans and the need for neurosurgical intervention in patients with GCS scores of 13 versus 14, 14 versus 15, and 13 versus 15. These results indicate significant differences in injury severity among patients with admission GCS scores of 13 to 15. The implicit assumption of clinicopathological homogeneity among patients with such scores is challenged by these data. This study demonstrates the need for more precise research diagnostic criteria in the study of neurobehavioral outcome after mild head injury. These findings also provide compelling evidence for the re-examination of the classification of mild head injury. Serious consideration must be given to the segregation of patients with GCS scores of 15 from those with scores of 14 and 13.