Abstract: Objective: To explore the predictors of deterioration in patients after craniotomy and to construct a risk prediction model. Methods: A total of 1,576 eligible patients were selected at three tertiary grade A hospitals in Hunan province from January 2018 to March 2020. The research samples were randomly divided into the training set (n=1,106) and the verification set (n=470) according to the ratio of 7:3, and the BP neural network was used to construct the prediction model. The predictive effect of the model was verified. Results: In the BP neural network variable importance score, intracranial hematoma on CT within 24 h after craniotomy and SpO2 contributed more to the model classification. The sensitivity of the prediction model was 77.1%, the specificity was 91.7%, the accuracy was 86.8%, the positive predictive value was 82.3%, and the negative predictive value was 88.9%. Conclusion: The prediction model based on the BP neural network algorithm has good predictive effects, which might provide a scientific and objective reference for clinical staff to predict the risk of patients' deterioration after craniotomy.

Key words: craniotomy; deterioration; risk prediction; neurosurgery; BP neural network