Development and validation of a predictive model for postoperative delirium in patients undergoing cardiac surgery

Figure 1. Study flowchart illustrating the construction and validation of the prediction model for postoperative delirium.

Figure 2. The predictors of the nomogram were screened using LASSO regression analysis in the training set. (A) Profiles of the coefficients for candidate variables across log(λ) values. (B) Ten-fold cross-validation was used to determine the optimal λ. The dotted vertical lines represent the λ values selected according to the minimum criterion and the 1-standard-error criterion.

Figure 3. The construction of POD prediction nomogram.

Figure 4. Discriminative performance of the prediction model assessed by area under the receiver operating characteristic (AUROC). (A) In the training set, the model yielded an AUC of 0.785 (95% CI: 0.743–0.826). (B) In the validation set, the model achieved an AUC of 0.812 (95% CI: 0.755–0.869).

Figure 5. Calibration assessment of the prediction model in the training and validation sets. (A,B) Bootstrap calibration curves for the respective datasets. The ideal line is shown as a black dashed line, the apparent line as a red dashed line, and the bias-corrected line after 500 bootstrap resamples as a blue solid line. (C,D) Flexible calibration curves for the training and validation sets. The fitted calibration curve is shown by the red solid line, the ideal reference line by the gray dashed line, and the shaded area indicates the 95% confidence interval.

Figure 6. The decision curves analysis of the nomogram in the training (A) and validation sets (B) the red line represents the net benefit of the nomogram, whereas the blue and green lines represent the strategies of treating all patients and treating no patients, respectively. In both the training and validation sets, the model showed clinical utility across the corresponding ranges of threshold probabilities.

Figure 7. The clinical impact curve of the nomogram in the training (A) and validation sets (B) the red solid line denotes the number of subjects considered high risk at each threshold probability, and the blue dashed line denotes the number of true-positive cases within the high-risk population.