External validation of the Tanta University Risk Model for predicting clinical outcomes in ICU patients with acute drug poisoning.

Abstract

Introduction: The Tanta University Risk Model has been externally validated for predicting intensive care unit admission in patients with acute poisoning, but its performance for intensive care unit outcomes has not been evaluated.

Methods: Adult patients with acute drug poisoning admitted to the intensive care unit were identified from the Medical Information Mart for Intensive Care IV database. Model performance for prolonged intensive care unit stay and 30 day mortality was evaluated using area under the receiver operating characteristic curve, sensitivity, and specificity.

Results: Among 1,329 patients, the median intensive care unit stay was 1.70 days, and the 30 day mortality was 6.5%. Stays of 2 days, 4 days, and 7 days or longer occurred in 38.8%, 15.5%, and 6.7% of patients, respectively. The model performed best for predicting stays of 7 days or longer (area under the curve = 0.837, 95% CI: 0.790-0.884) and moderately for 30 day mortality (area under the curve = 0.709, 95% CI: 0.645-0.774).

Conclusion: The Tanta University Risk Model predicts intensive care unit outcomes in acute drug poisoning, particularly prolonged stays of 7 days or longer, supporting its use as a practical clinical decision tool.