Predicting outcome after severe head injury--a computer-assisted analysis of neurological symptoms and laboratory values.

A total number of 58 parameters (laboratory values, neurological symptoms, and vegetative parameters) were evaluated in 150 patients during the first seven days after severe head injury. The patients were divided into two groups, “survivors” and “non-survivors”. Eight easily evaluable routine parameters with the most significant differences between the two groups of patients were used for statistical evaluation of a “no survival chance score”. These highly indicative parameters are serum osmolarity and urea, blood glucose, total bilirubin, motor reaction to stimuli, body temperature, respiratory activity, and pupil reaction. A “low survival chance limit” was evaluated from each of these parameters by computer analysis. None of the patients in the series survived when three or more of these eight parameters had climbed beyond the limit. So far, the system is able to predict “no survival chances” in 50.8% of the non-survivors some six days prior to death; 80% of these predictions could be made by the fourth day after injury.