This paper re-examines the theoretical concept of severe brain injury
focusing on the duration of coma as a precise indicator of the clinica
l profile. A retrospective hospital chart study of 361 traumatic brain
-injured patients was undertaken to determine the homogeneity of the s
ubsample of the severely brain-injured (defined as 2 or more days of c
oma) with respect to the probability of four types of impairment: atax
ia, contractures, paralysis and speech impairment. The current concept
of severity assumes homogeneity among the 'severely brain-injured'. H
owever, our results indicate significant differences in impairment wit
hin this population. The authors feel strongly that future studies mus
t describe coma duration in finer gradations, and test for homogeneity
within samples before inferences are made. Improvements in life-susta
ining technologies have resulted in longer coma durations. The need to
use coma days as an indicator of impairment rather than a broad categ
ory of severity is emphasized.