Prediction models for postoperative delirium in elderly patients with machine-learning algorithms and SHapley Additive exPlanations

Development of a Disease Model for Predicting Postoperative Delirium Using Combined Blood Biomarkers

OBJECTIVE

Postoperative delirium, a common neurocognitive complication after surgery and anesthesia, requires early detection for potential intervention. Herein, we constructed a multidimensional postoperative delirium risk-prediction model incorporating multiple demographic parameters and blood biomarkers to enhance prediction accuracy.

METHODS

We included 555 patients undergoing radical surgery for colorectal cancer. Demographic characteristics and lipid profiles were collected preoperatively, and perioperative anesthesia and surgical conditions were recorded; blood biomarkers were measured before and after surgery. The 3D-CAM scale was used to assess postoperative delirium occurrence within 3 days after surgery. Patients were divided into the postoperative delirium (N = 100) and non-postoperative delirium (N = 455) groups. Based on machine learning, linear and nine non-linear models were developed and compared to select the optimal model. Shapley value-interpretation methods and mediation analysis were used to assess feature importance and interaction.

RESULTS

The median age of the participants was 65 years (interquartile range: 56-71 years; 57.8% male). Among the 10 machine-learning models, the random forest model performed the best (validation cohort, area under the receiver operating characteristic curve of 0.795 [0.704-0.885]). Lipid profile (total cholesterol, triglycerides, and trimethylamine-N-oxide) levels were identified as key postoperative delirium predictors. Mediation analysis further confirmed mediating effects among total cholesterol, trimethylamine-N-oxide, and postoperative delirium; a nomogram model was developed as a web-based tool for external validation and use by other clinicians.

INTERPRETATION

Blood biomarkers are crucial in predicting postoperative delirium and aid anesthesiologists in identifying its risks in a timely manner. This model facilitates personalized perioperative management and reduces the occurrence of postoperative delirium.

TRIAL REGISTRATION

ChiCTR2300075723.

Machine Learning Multimodal Model for Delirium Risk Stratification

Importance

Automating the identification of risk for developing hospital delirium with models that use machine learning (ML) could facilitate more rapid prevention, identification, and treatment of delirium. However, there are very few reports on the performance of ML models for delirium risk stratification in live clinical practice.

Objective

To report on development, operationalization, and validation of a multimodal ML model for delirium risk stratification in live clinical practice and its associations with workflow and clinical outcomes.

Design, Setting, and Participants

This quality improvement study developed an ML model supported by automated electronic medical records to stratify the risk of non-intensive care unit delirium in live clinical practice using the Confusion Assessment Method as the diagnostic reference standard, with an iterative model update method. Data from patients aged at least 60 years admitted to non-intensive care units at Mount Sinai Hospital between January 2016 and January 2020 were used to train and test the ML model presented. The model was validated in live clinical practice from March 2023 to March 2024. Analysis of the model's associations with workflow and clinical outcomes was conducted retrospectively in 2024, comparing hospitalized patients prior to deployment of any model version (pre-ML cohort) and during model clinical deployment (post-ML cohort).

Main Outcomes and Measures

Outcomes of interest were area under the receiver operating characteristic curve, monthly delirium detection rates, median length of hospital stay, and daily doses of opiate, benzodiazepine, and antipsychotic medications administered.

Results

The overall sample included 32 284 inpatient admissions (mean [SD] age, 73.56 (9.67) years, 15 157 [46.9%] women). A total of 25 261 inpatient admissions of older patients with both medical and surgical primary diagnoses represented the combined model testing and training cohort (median age, 73.37 [66.42-81.36] years) and live clinical deployment validation cohort (median [IQR] age, 72.11 [62.26-78.97] years), while 7023 inpatient admissions of older patients with both medical and surgical primary diagnoses represented the combined pre-ML (median [IQR] age, 74.00 [68.00-81.00] years) and post-ML (median [IQR] age, 75.33 [68.34-82.91] years) cohorts. The model presented is a fusion of electronic medical record patient data features and clinical note features processed by natural language processing. The results of model validation in live clinical practice included an area under the curve of 0.94 (95% CI, 0.93-0.95). Median (IQR) monthly delirium detection rates of inpatients assessed for delirium with the Confusion Assessment Method increased from 4.42% (95% CI, 3.70%-5.14%) in the pre-ML cohort to 17.17% (95% CI, 15.54%-18.80%) in the post-ML cohort (P < .001). Post-ML vs pre-ML cohorts received lower daily doses of benzodiazepines (median [IQR] 0.93 [0.42-2.28] diazepam dose equivalents vs 1.60 [0.66-4.27] diazepam dose equivalents; P < .001) and olanzapine (median [IQR], 1.09 [0.38-2.46] mg vs 2.50 [1.17-6.65] mg; P < .001).

Conclusions and Relevance

This quality improvement study demonstrates the feasibility of a novel multimodal ML model to automate delirium risk stratification in live clinical practice. The model demonstrated acceptable performance in live clinical practice and may facilitate resource allocation to enhance delirium identification and care.

Correlation analysis of a novel artificial intelligence optical microscope-assisted semen assessment system with IVF outcomes

PURPOSE

On the day of fresh oocyte retrieval in in vitro fertilization (IVF) cycles, a novel portable artificial intelligence optical microscope (AIOM) was employed to assist in the assessment of semen parameters. This study analyzed the correlation between sperm kinetic and morphological parameters with short-term IVF outcomes. Additionally, it explored whether these parameters could serve as predictive indicators for rescue intracytoplasmic sperm injection (R-ICSI) in IVF patients.

METHODS

A retrospective analysis was conducted on patients undergoing short-term IVF at the West China Second Hospital of Sichuan University between May 2021 and May 2024. Based on fertilization outcomes, the short-term IVF patients were categorized into a successful fertilization group (group A, n = 281) and a group requiring R-ICSI after failed fertilization (group B, n = 49). AIOM was utilized to analyze semen parameters including pH, sperm concentration, sperm motility parameters, sperm movement trajectory parameters, and sperm morphological parameters. The study further investigated the correlation between these short-term IVF fertilization-related laboratory indicators and IVF outcomes.

RESULTS

No statistically significant difference was observed in semen pH between the two groups. However, there were significant differences in sperm concentration and the majority of motility parameters. Specifically, compared to group A, patients in group B exhibited lower sperm concentration (p = 0.01), motility (p = 0.01), local motility (p = 0.01), progressive motility (PR) (p = 0.00), total motility (p = 0.01), and amplitude of lateral head displacement (ALH) (p < 0.01), along with higher immotility (p = 0.00). No statistically significant differences were found between the two groups in other sperm motility, velocity, or trajectory parameters. Additionally, sperm morphological parameters were also associated with short-term IVF fertilization outcomes. Compared to group A, group B had higher sperm head length mean (p < 0.01), head perimeter mean (p < 0.01), and head area mean (p = 0.01), as well as lower tail length mean (p = 0.01). Multivariate regression analysis of fertilization outcomes indicated that higher immotility (p = 0.01) and head length mean (p < 0.01), along with lower tail length mean (p = 0.04), were independent risk factors affecting successful short-term IVF fertilization. Notably, head length mean showed a significant negative correlation with polyspermy rate (p < 0.01), whereas tail length mean was significantly positively correlated with polyspermy rate (p < 0.01).

CONCLUSION

Optimization of semen parameters with AIOM at the time of fertilization is significantly associated with short-term IVF fertilization outcomes. Abnormal semen parameters at fertilization-specifically, higher immotility and head length mean, along with lower tail length mean-can be considered risk factors for fertilization failure and may serve as predictive indicators for potential R-ICSI.

Enhanced machine learning predictive modeling for delirium in elderly ICU patients with COPD and respiratory failure: A retrospective study based on MIMIC-IV

BACKGROUND AND OBJECTIVE

Elderly patients with Chronic obstructive pulmonary disease (COPD) and respiratory failure admitted to the intensive care unit (ICU) have a poor prognosis, and the occurrence of delirium further worsens outcomes and increases hospitalization costs. This study aimed to develop a predictive model for delirium in this patient population and identify associated risk factors.

METHODS

Data for the machine learning model were obtained from the MIMIC-IV database. Feature variable screening was conducted using Lasso regression and the best subset method. Four models-K-nearest neighbor, random forest, logistic regression, and extreme gradient boosting (XGBoost)-were trained and optimized to predict delirium risk. The stability of the model is evaluated using ten-fold cross validation and the effectiveness of the model on the validation set is evaluated using accuracy, F1 score, precision and recall. The SHapley Additive exPlanations (SHAP) method was used to explain the importance of each variable in the model.

RESULTS

A total of 1,155 patients admitted to the intensive care unit between 2008 and 2019 were included in the study, with a delirium incidence of 12.9% (149/1,155). Among the four ML models evaluated, the XGBoost model demonstrated the best discriminative ability. In the validation set, it achieved an AUC of 0.932, indicating superior performance with high accuracy, precision, recall, and F1 scores of 0.891, 0.839, 0.795, and 0.810, respectively. Key features identified through SHAP analysis included the Glasgow Coma Scale (GCS) verbal score, length of hospital stay, mean SpO₂ on the first day of ICU admission, Modification of Diet in Renal Disease (MDRD) equation score, mean diastolic blood pressure, GCS motor score, gender, and duration of noninvasive ventilation. These findings provide valuable insights for individualized risk management.

CONCLUSIONS

The developed prediction model effectively predicts the occurrence of delirium in elderly COPD patients with respiratory failure in the ICU. This model can assist clinical decision-making, potentially improving patient outcomes and reducing healthcare costs.

Analysis of influencing factors and clinical application of a predictive model for emergence agitation from general anesthesia after abdominal surgery

OBJECTIVES

To identify factors influencing emergence agitation (EA) in abdominal surgery patients and develop a predictive model for early clinical intervention.

METHODS

We retrospectively analyzed data from 794 patients who underwent abdominal surgery between June 2022 and June 2024. Independent risk factors for EA were identified using multivariate logistic regression, which informed the construction of a nomogram model. The dataset was split into a training set (67%) and a validation set (33%), with an additional 119 patients serving as an external validation set. Data analysis was performed using SPSS 26.0 and R 4.3.3, and model performance was assessed using Receiver Operating Characteristic (ROC) and calibration curves.

RESULTS

Multivariate analysis revealed nine independent risk factors for EA: age, ASA classification, type of surgery, duration of surgery, intraoperative fluid volume, use of analgesic pumps, catheter usage, postoperative pain, and smoking history. The model's area under the curve (AUC) was 0.787 in the training set, 0.623 in the validation set, and 0.666 in the external validation set, indicating good predictive performance. Calibration curves demonstrated a strong agreement between predicted and observed outcomes, confirming the model's accuracy and consistency.

CONCLUSION

The developed nomogram integrates multiple risk factors to predict EA risk in abdominal surgery patients. It demonstrates high stability and applicability across different datasets, facilitating early identification of high-risk patients and supporting individualized postoperative management.

Development and validation of a prediction model for myelosuppression in lung cancer patients after platinum-based doublet chemotherapy: a multifactorial analysis approach

OBJECTIVE

To develop an individualized prediction model for myelosuppression risk in lung cancer patients undergoing platinum-based doublet chemotherapy and validate its predictive efficacy.

METHODS

A retrospective analysis was conducted on the clinical data of 584 lung cancer patients who received platinum-based doublet chemotherapy at The Affiliated Hospital of Qingdao University between January 2016 and December 2020. Patients were randomly assigned to a training cohort (n=391) and a validation cohort (n=193). Myelosuppression occurred in 280 (71.6%) patients in the training cohort and 132 (68.4%) in the validation cohort. Univariate analysis and LASSO regression were used to identify independent risk factors for myelosuppression. Prediction models were developed using Support Vector Machine (SVM), Random Forest, Extreme Gradient Boosting (XGBoost), and Adaptive Boosting (Adaboost). Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and Decision Curve Analysis (DCA). The SHAP algorithm was employed to evaluate feature importance, and a nomogram was developed for individual risk prediction.

RESULTS

LASSO regression identified 10 independent risk factors for myelosuppression: age, body mass index (BMI), white blood cell count, neutrophil count, platelet count, total protein, gender, treatment regimen, targeted therapy, and first chemotherapy cycle. In the training cohort, the XGBoost model exhibited the best performance, with an area under the curve (AUC) of 0.855 (95% CI: 0.813-0.897), while the AUC in the validation cohort was 0.793. SHAP analysis identified white blood cell count, platelet count, neutrophil count, BMI, and age as the most influential predictors. The SHAP analysis based on the XGBoost model demonstrated substantial value.

CONCLUSION

This study successfully developed an individualized prediction model for myelosuppression risk in lung cancer patients following platinum-based doublet chemotherapy, with the XGBoost model achieving high predictive accuracy and clinical utility. The model provides a valuable tool for guiding precision medicine.

Machine Learning for Prediction of Postoperative Delirium in Adult Patients: A Systematic Review and Meta-analysis

PURPOSE

This meta-analysis aimed to evaluate the performance of machine learning (ML) models in predicting postoperative delirium (POD) and to provide guidance for clinical application.

METHODS

PubMed, Embase, Cochrane Library, and Web of Science databases were searched from inception to April 29, 2024. Studies reported ML models for predicting POD in adult patients were included. Data extraction and risk of bias assessment were performed using the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis - AI (TRIPOD-AI) and Prediction model Risk Of Bias ASsessment Tool (PROBAST) tools. Meta-analysis with the area under the curve (AUC) was performed using MedCalc software.

FINDINGS

A total of 23 studies were included after screening. Age (n = 20, 86.95%) and Random Forest (RF) (n = 24, 17.27%) were the most frequently used feature and ML algorithm, respectively. The meta-analysis showed an overall AUC of 0.792. The ensemble models (AUC = 0.805) showed better predictive performance than single models (AUC = 0.782). Additionally, considerable variations in AUC were found among different ML algorithms, with AdaBoost (AB) demonstrating good performance with AUC of 0.870. Notably, the generalizability of these models was uncertain due to limitations in external validation and bias assessment.

IMPLICATIONS

The performance of ensemble models were higher than single models, and the AB algorithms demonstrated better performance, compared with other algorithms. However, further research was needed to enhance the generalizability and transparency of ML models.

Development and validation of a risk prediction model for severe postoperative complications in elderly patients with hip fracture

OBJECTIVE

This study aimed to investigate risk factors associated with severe postoperative complications following hip fracture surgery in elderly patients and to develop a nomogram-based risk prediction model for these complications.

METHODS

A total of 627 elderly patients with hip fractures treated at Yongchuan Hospital of Chongqing Medical University from January 2015 to April 2024 were collected. 439 patients were assigned to the training cohort for model development, and 188 to the validation cohort for model assessment. The training cohort was stratified based on the presence or absence of severe complications. We employed LASSO regression, as well as univariate and multivariate logistic regression analyses, to identify significant factors. A nomogram was constructed based on the outcomes of the multivariate regression. The model's discriminative ability was assessed using the area under the receiver operating characteristic curve (AUC), while calibration plots and decision curve analysis (DCA) evaluated its calibration and stability. Internal validation was performed using the validation cohort.

RESULTS

Out of the 627 patients, 118 (18.82%) experienced severe postoperative complications. Both LASSO regression and multivariate logistic analysis identified the modified 5-item frailty index (mFI-5) and the preoperative C-reactive protein to albumin ratio (CAR) as significant predictors of severe complications. The nomogram model, derived from the multivariate analysis, exhibited strong discriminative ability, with an AUC of 0.963 (95% CI: 0.946-0.980) for the training cohort and 0.963 (95% CI: 0.938-0.988) for the validation cohort. Calibration plots demonstrated excellent agreement between the nomogram's predictions and actual outcomes. Decision curve analysis (DCA) indicated that the model provided clinical utility across all patient scenarios. These findings were consistent in the validation cohort.

CONCLUSIONS

Both the mFI-5 and CAR are predictive factors for severe postoperative complications in elderly patients undergoing hip fracture surgery.

Preoperative serum ferritin as a biomarker for predicting delirium among elderly patients receiving non-cardiac surgery: a retrospective cohort study

Iron metabolism disorder has been identified as a contributor to the pathogenesis and progression of multiple cognitive dysfunction-related diseases, including postoperative delirium. However, the association between preoperative iron reserves and postoperative delirium risk remains elusive. This retrospective cohort study aimed to explore the impact of preoperative serum ferritin levels on the risk of postoperative delirium in elderly patients undergoing non-neurosurgical and non-cardiac procedures. Conducted at the Chinese PLA General Hospital between January 2014 and December 2021, the study finally included 12,841 patients aged 65 years and above. Preoperative serum ferritin levels were assessed within 30 days before surgery, and postoperative delirium occurrence within the first seven days after surgery was determined through medical chart review. The analyses revealed that both low and high levels of serum ferritin were associated with an increased risk of postoperative delirium. Patients in the lowest quintile of serum ferritin exhibited an 81% increased risk, while those in the highest quintile faced a 91% increased risk compared to those in the second quintile. Furthermore, mediation analyses indicated that the direct effect of preoperative serum ferritin on postoperative delirium contradicted its indirect effect mediated by hemoglobin levels. These findings suggest that maintaining serum ferritin within moderate range preoperatively could be beneficial for managing postoperative delirium risk among elderly patients.

Interpretable machine learning model for early prediction of delirium in elderly patients following intensive care unit admission: a derivation and validation study

Background and objective

Delirium is the most common neuropsychological complication among older adults admitted to the intensive care unit (ICU) and is often associated with a poor prognosis. This study aimed to construct and validate an interpretable machine learning (ML) for early delirium prediction in older ICU patients.

Methods

This was a retrospective observational cohort study and patient data were extracted from the Medical Information Mart for Intensive Care-IV database. Feature variables associated with delirium, including predisposing factors, disease-related factors, and iatrogenic and environmental factors, were selected using least absolute shrinkage and selection operator regression, and prediction models were built using logistic regression, decision trees, support vector machines, extreme gradient boosting (XGBoost), k-nearest neighbors and naive Bayes methods. Multiple metrics were used for evaluation of performance of the models, including the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, recall, F1 score, calibration plot, and decision curve analysis. SHapley Additive exPlanations (SHAP) were used to improve the interpretability of the final model.

Results

Nine thousand seven hundred forty-eight adults aged 65 years or older were included for analysis. Twenty-six features were selected to construct ML prediction models. Among the models compared, the XGBoost model demonstrated the best performance including the highest AUC (0.836), accuracy (0.765), sensitivity (0.713), recall (0.713), and F1 score (0.725) in the training set. It also exhibited excellent discrimination with AUC of 0.810, good calibration, and had the highest net benefit in the validation cohort. The SHAP summary analysis showed that Glasgow Coma Scale, mechanical ventilation, and sedation were the top three risk features for outcome prediction. The SHAP dependency plot and SHAP force analysis interpreted the model at both the factor level and individual level, respectively.

Conclusion

ML is a reliable tool for predicting the risk of critical delirium in elderly patients. By combining XGBoost and SHAP, it can provide clear explanations for personalized risk prediction and more intuitive understanding of the effect of key features in the model. The establishment of such a model would facilitate the early risk assessment and prompt intervention for delirium.

Risk stratification and predictive modeling of postoperative delirium in chronic subdural hematoma

Background- Postoperative delirium is a common complication associated with the elderly, causing increased morbidity and prolonged hospital stay. However, its risk factors in chronic subdural hematoma patients have not been well studied. Methods- A total of 202 consecutive patients with chronic subdural hematoma at Peking University Third Hospital between January 2018 and January 2023 were enrolled. Various clinical indicators were analyzed to identify independent risk factors for postoperative delirium using univariate and multivariate regression analyses. Delirium risk prediction models were developed as a nomogram and a Markov chain. Results- Out of the 202 patients (age, 71 (IQR, 18); female-to-male ratio, 1:2.7) studied, 63 (31.2%) experienced postoperative delirium. Univariate analysis identified age (p < 0.001), gender (p = 0.014), restraint belt use (p < 0.001), electrolyte imbalance (p < 0.001), visual analog scale score (p < 0.001), hematoma thickness (p < 0.001), midline shift (p < 0.001), hematoma side (p = 0.013), hematoma location (p = 0.018), and urinal catheterization (p = 0.028) as significant factors. Multivariate regression analysis confirmed the significance of restraint belt use (B = 7.657, p < 0.001), electrolyte imbalance (B = -3.993, p = 0.001), visual analog scale score (B = 2.331, p = 0.016), and midline shift (B = 0.335, p = 0.007). Hematoma thickness and age had no significant impact. Conclusion- Increased midline shift and visual analog scale scores, alongside restraint belt use and electrolyte imbalance elevate delirium risk in chronic subdural hematoma surgery. Our prediction models may offer reference value in this context.