Comparison of logistic regression and machine learning methods for predicting postoperative delirium in elderly patients: A retrospective study

Song, Yuxiang; Yang, Xiaodong; Luo, Yungen; Ouyang, Chun‐lei; Yu, Yanxia; Ma, Yulong; Li, Hao; Lou, Jingsheng; Liu, Yan‐hong; Chen, Yi‐qiang; Cao, Jiangbei; Mi, Weidong

doi:10.1111/cns.13991

Cited by 16 publications

(9 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Development of postoperative complications: Most studies focused on the prediction of postoperative acute complications 26–28,42,44–49,59,99,110,112,113,118,123 such as pain and opioid use, 53–57 postoperative atrial fibrillation (new-onset atrial fibrillation), 82 postoperative risk of stroke or myocardial infarction, 50,71,77 and delirium or cognitive decline. 65–70 Other models focused on the risk of developing pneumonia or respiratory failure, 83,85,125 acute kidney injury, 43,52,58,60–63,120–122 liver failure 117 or development of sepsis or surgical site infection. 74–76,87,99,131 Suhre et al .…”

Section: Resultsmentioning

confidence: 99%

“…Elective noncardiac surgery Mortality 30 days and/or 1 yr [26][27][28][29][30][31][32][33][34][35][36][37][38] In surgical patients with perioperative SarS-CoV-2 39 Morbidity Multiple postoperative complications 26,27,29,[40][41][42][43][44][45][46][47][48][49][50][51] acute and chronic pain 52-57 acute kidney failure 52,58-63 aSa score prediction 64 Delirium and cognitive decline [65][66][67][68][69][70] Cerebral/myocardial infarction 71 Difficult intubation prediciton 72 Ileus 73 Infection risk [74][75][76] Myocardial injury 77 Nausea and vomiting 78 Obstructive apnoea screening 79 Perioperative transfusion 80,81 Postoperative atrial fibrillation 82 respiratory failure and depression Liver failure 117 Major bleeding 118,119 Kidney failure…”

Section: Surgery Outcomes and Eventsmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools

Arina,

Kaczorek,

Hofmaenner

et al. 2023

Anesthesiology

View full text Add to dashboard Cite

Background The utilization of artificial intelligence and machine learning as diagnostic and predictive tools in perioperative medicine holds great promise. Indeed, many studies have been performed in recent years to explore the potential. The purpose of this systematic review is to assess the current state of machine learning in perioperative medicine, its utility in prediction of complications and prognostication, and limitations related to bias and validation. Methods A multidisciplinary team of clinicians and engineers conducted a systematic review using the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) protocol. Multiple databases were searched, including Scopus, Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Cochrane Library, PubMed, Medline, Embase, and Web of Science. The systematic review focused on study design, type of machine learning model used, validation techniques applied, and reported model performance on prediction of complications and prognostication. This review further classified outcomes and machine learning applications using an ad hoc classification system. The Prediction model Risk Of Bias Assessment Tool (PROBAST) was used to assess risk of bias and applicability of the studies. Results A total of 103 studies were identified. The models reported in the literature were primarily based on single-center validations (75%), with only 13% being externally validated across multiple centers. Most of the mortality models demonstrated a limited ability to discriminate and classify effectively. The PROBAST assessment indicated a high risk of systematic errors in predicted outcomes and artificial intelligence or machine learning applications. Conclusions The findings indicate that the development of this field is still in its early stages. This systematic review indicates that application of machine learning in perioperative medicine is still at an early stage. While many studies suggest potential utility, several key challenges must be first overcome before their introduction into clinical practice. Editor’s Perspective What We Already Know about This Topic What This Article Tells Us That Is New

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Surgery Outcomes and Eventsmentioning

confidence: 99%

Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools

Arina,

Kaczorek,

Hofmaenner

et al. 2023

Anesthesiology

View full text Add to dashboard Cite

show abstract

“…More common classification techniques used, such as rough set theory (RST) [26], DT [27], RF [28], and MLP [29,30], have become an important research trend at present. Moreover, Bayes network (BN) learning [31], logistic regression (LR) [32], naïve Bayesian (NB) [33], and support vector machine (SVM) [34] classifiers are always emerging techniques helpful for industry application fields; thus, they were also selected and emphasized in this study for the sake of comparison. (4) The rule extraction method: The extracted decision rule set has attracted the attention of many researchers and established an effective decision rule knowledge base to mine hidden information, such as the famous C4.5 of the decision tree and the LEM2 algorithms of the rough set theory, and it has been applied in different fields [35] with outstanding performance.…”

Section: Continuous Research Motivation and Research Originalitymentioning

confidence: 99%

“…We have also learned from the past that it has advantageous performance in different fields, such as vault settlement in urban tunnels [74]. (2) LR [32,75] is mainly used to discuss the relationship between the dependent and the independent variables, which is suitable for the situation of the binary category data of the dependent variable. For example, the two results (with and without insurance) that are suitable for predicting LTCI are addressed in a method of reaction variable statistical model analysis on learning from the literature, and its past classification performance is also very outstanding.…”

Section: Other Well-known Classifier Techniques and Related Applicati...mentioning

confidence: 99%

A Study to Identify Long-Term Care Insurance Using Advanced Intelligent RST Hybrid Models with Two-Stage Performance Evaluation

Chen

Hung

Lin³

2023

Mathematics

View full text Add to dashboard Cite

With the motivation of long-term care 2.0 plans, forecasting models to identify potential customers of long-term care insurance (LTCI) are an important and interesting issue. From the limited literature, most past researchers emphasize traditional statistics techniques to address this issue; however, these are lacking in some areas. For example, intelligent hybrid models for LTCI are lacking, performance measurement of components for hybrid models is lacking, and research results for interpretative capacities are lacking, resulting in a black box scenario and difficulty in making decisions, and the gap between identifying potential customers and constructing hybrid models is unbridged. To solve the shortcomings mentioned above, this study proposes some advanced intelligent single and hybrid models; the study object is LTCI customers. The proposed hybrid models were used on the experimental dataset collected from real insurance data and possess the following advantages: (1) The feature selection technique was used to simplify variables for the purpose of improving model performance. (2) The performance of hybrid models was evaluated against some machine learning methods, including rough set theory, decision trees, multilayer perceptron, support vector machine, genetic algorithm, random forest, logistic regression, and naive Bayes, and sensitivity analysis was performed in terms of accuracy, coverage, rules number, and standard deviation. (3) We used the C4.5 algorithm of decision trees and the LEM2 algorithm of rough sets to extract and provide valuably comprehensible decisional rules as decision-making references for the interested parties for their varied benefits. (4) We used post hoc testing to verify the significant difference in groups. Conclusively, this study effectively identifies potential customers for their key attributes and creates a decision rule set of knowledge for use as a reference when solving practical problems by forming a structured solution. This study is a new trial in the LTCI application field and realizes novel creative application values. Such a hybrid model is rarely seen in identifying LTCI potential customers; thus, the study has sufficient application contribution and managerial benefits to attract much concern from the interested parties.

show abstract

“…characteristics of older patients in ICU settings(45)(46)(47). As far as we know, this is the first study on the risk prediction of delirium in critically ill patients aged 65 years and older.…”

mentioning

confidence: 99%

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

Tang,

Ma,

2024

Front. Med.

View full text Add to dashboard Cite

Background and objectiveDelirium 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.MethodsThis 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.ResultsNine 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.ConclusionML 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.

show abstract

Comparison of logistic regression and machine learning methods for predicting postoperative delirium in elderly patients: A retrospective study

Cited by 16 publications

References 39 publications

Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools

Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools

A Study to Identify Long-Term Care Insurance Using Advanced Intelligent RST Hybrid Models with Two-Stage Performance Evaluation

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

Contact Info

Product

Resources

About