Turkish Neurosurgery
Internal validation of two models developed for determining prognosis of patients with isolated traumatic brain injury
Ahmet Tolgay Akıncı1, Osman Şimşek1, Kocatürk Murat2
1Trakya University, Neurosurgery Department, Edirne,
2Çiğli Regional Education Hospital, Neurosurgery Department, İzmir,
DOI: 10.5137/1019-5149.JTN.30145-20.6
Aim:To evaluate the efficiency of two models for prognostication of patients with isolated traumatic brain injury (TBI).Material and Methods:The models were developed using patient data in 10 years; internal validation was performed using the data of patients who were admitted within the following 5 years. We reviewed the records of 204 patients with TBI accepted into the Neurosurgery Department and Intensive Care Units. Models were applied to estimate prognosis. Calculations were statistically compared with the actual clinical outcomes of patients using discriminant analysis.Results:For Model 1, the correct classification rate was 87.9%, the specificity was 66.7%, the sensitivity was 94.2%, the positive predictive value was 68.8% and the negative predictive value was 93.6%. For Model 2, the correct classification rate was 90.2%, specificity was 57.6%, sensitivity was 96.5%, positive predictive value was 76% and negative predictive value was 92.2%.Conclusion:Both models had decent correct classification rates and may be efficient tools for estimating the prognosis of unfavourable outcomes in patients with isolated TBI. These models are good candidates for wide use following validation in national and international multicentric studies.
Corresponding author : Ahmet Tolgay Akıncı