A prediction model for identifying alcohol withdrawal seizures

A retrospective review of alcohol withdrawal seizures was performed at a private chemical-dependence treatment facility to help identify patients who were at high risk for having a seizure. Patients were identified by two means: controlled substance records were reviewed to determine patients having received intramuscular phenobarbital, and patient charts were reviewed for all patients with a discharge diagnosis of a seizure disorder. Two thousand and one patient records were reviewed; alcohol withdrawal seizure patients were identified. Twenty-eight randomly selected nonseizure patient records served as controls. The statistical test consisted of a discriminant function analysis. The data yielded a statistically significant predictive model for alcohol withdrawal seizures based on six interdependent patient variables which will be helpful in treating future patients undergoing alcohol withdrawal.