Influence of Curved Video Laryngoscope Blade Sizes and Patient Heights on Video Laryngoscopic Views: A Randomized Controlled Trial.

This study aimed to compare the video laryngoscope views facilitated by curved blades 3 and 4 with an exploration of the relationship between these views and patient height. Conducted as a randomized controlled trial, this study enrolled adults scheduled for surgery under general anesthesia. Intubation procedures were recorded, and the percentage of glottic opening was measured before tube insertion. Multivariate analysis validated the impact of various factors, including blade size and patient height, on the percentage of glottic opening scores. A total of 192 patients were included. The median percentage of glottic opening scores for curved blades 3 and 4 were 100 and 83, respectively (p < 0.001). The unstandardized coefficient indicated a significant negative impact of blade 4 on the percentage of glottic opening scores (-13, p < 0.001). In the locally estimated scatterplot smoothing analysis, blade 3 exhibited a steady rise in glottic opening scores with increasing height, whereas blade 4 showed a peak followed by a decline around 185 cm. The unstandardized coefficient of height showed no significant association (0, p = 0.819). The study observed superior laryngoscopic views with blade 3 compared to blade 4. However, no significant association was found between laryngoscopic views and patient height.

Effect of sugammadex on postoperative complications in patients with severe burn who underwent surgery: a retrospective study.

This retrospective study investigated the association of sugammadex with postoperative pulmonary complication risk between 2013 and 2021 in patients with severe burn of five hospitals. Postoperative pulmonary complications included atelectasis, pulmonary edema, pulmonary effusion, pneumothorax, pneumonia, pulmonary thromboembolism, respiratory failure and acute respiratory distress. To identify whether sugammadex reduced the risk of postoperative pulmonary complication in patients with severe burn who underwent surgery, Kaplan-Meier curve were used to check the difference of incidence according to surgical cases and time-varying Cox hazard regression were used to calculate the hazard ratio. The study included 1213 patients with severe burn who underwent 2259 surgeries. Postoperative pulmonary complications were occurred in 313 (25.8%) patients. Among 2259 surgeries, sugammadex was used in 649 (28.7%) surgeries. Cumulative postoperative pulmonary complication were 268 (16.6%) cases in surgeries without sugammadex, and 45 (6.9%) cases in surgeries with sugammadex, respectively (P < 0.005). The postoperative pulmonary complications risk was reduced significantly in patients who use sugammadex than those who did not use sugammadex. (Adjusted hazard ratio, 0.61; 95% confidence interval, 0.42-0.89; P = 0.011). In conclusion, sugammadex reduced risk of postoperative pulmonary complications compared with nonuse of sugammadex in patients with severe burn who underwent surgery.

Stereotactic ablative radiotherapy for primary renal cell carcinoma.

Stereotactic ablative radiotherapy (SAbR) is an emerging non-invasive definitive treatment option for primary renal cell carcinoma (RCC), particularly when surgery is not ideal. Employing ablative doses, SAbR delivered in one to five fractions to the primary tumor has been shown to achieve high local control rates with favorable toxicity profile in multiple retrospective and prospective series, and has dispelled previous notions of RCC radio-resistance. Moreover, emerging evidence suggests possible immunomodulatory effects, leading to clinical investigations of SAbR in combination with systemic and surgical management in patients with metastatic disease. In this review, we summarize key evidence supporting SAbR delivered to the primary tumor including preclinical rationale, dose escalation studies, recent prospective trials, and outcomes from ongoing multi-institutional registries. We also discuss areas of active clinical investigation including the use of primary SAbR in combination with systemic therapies in patients with metastatic disease. The accumulated body of evidence supports SAbR as promising indication being increasingly incorporated into the multi-disciplinary management of primary RCC.

Predictive Modeling and Integrated Risk Assessment of Postoperative Mortality and Pneumonia in Traumatic Brain Injury Patients through Clustering and Machine Learning: Retrospective Study.

This study harnessed machine learning to forecast postoperative mortality (POM) and postoperative pneumonia (PPN) among surgical traumatic brain injury (TBI) patients. Our analysis centered on the following key variables: Glasgow Coma Scale (GCS), midline brain shift (MSB), and time from injury to emergency room arrival (TIE). Additionally, we introduced innovative clustered variables to enhance predictive accuracy and risk assessment. Exploring data from 617 patients spanning 2012 to 2022, we observed that 22.9% encountered postoperative mortality, while 30.0% faced postoperative pneumonia (PPN). Sensitivity for POM and PPN prediction, before incorporating clustering, was in the ranges of 0.43-0.82 (POM) and 0.54-0.76 (PPN). Following clustering, sensitivity values were 0.47-0.76 (POM) and 0.61-0.77 (PPN). Accuracy was in the ranges of 0.67-0.76 (POM) and 0.70-0.81 (PPN) prior to clustering and 0.42-0.73 (POM) and 0.55-0.73 (PPN) after clustering. Clusters characterized by low GCS, small MSB, and short TIE exhibited a 3.2-fold higher POM risk compared to clusters with high GCS, small MSB, and short TIE. In summary, leveraging clustered variables offers a novel avenue for predicting POM and PPN in TBI patients. Assessing the amalgamated impact of GCS, MSB, and TIE characteristics provides valuable insights for clinical decision making.

Postoperative Nausea and Vomiting Prediction: Machine Learning Insights from a Comprehensive Analysis of Perioperative Data.

Postoperative nausea and vomiting (PONV) are common complications after surgery. This study aimed to present the utilization of machine learning for predicting PONV and provide insights based on a large amount of data. This retrospective study included data on perioperative features of patients, such as patient characteristics and perioperative factors, from two hospitals. Logistic regression algorithms, random forest, light-gradient boosting machines, and multilayer perceptrons were used as machine learning algorithms to develop the models. The dataset of this study included 106,860 adult patients, with an overall incidence rate of 14.4% for PONV. The area under the receiver operating characteristic curve (AUROC) of the models was 0.60-0.67. In the prediction models that included only the known risk and mitigating factors of PONV, the AUROC of the models was 0.54-0.69. Some features were found to be associated with patient-controlled analgesia, with opioids being the most important feature in almost all models. In conclusion, machine learning provides valuable insights into PONV prediction, the selection of significant features for prediction, and feature engineering.

Harnessing Machine Learning for Prediction of Postoperative Pulmonary Complications: Retrospective Cohort Design.

Postoperative pulmonary complications (PPCs) are significant causes of postoperative morbidity and mortality. This study presents the utilization of machine learning for predicting PPCs and aims to identify the important features of the prediction models. This study used a retrospective cohort design and collected data from two hospitals. The dataset included perioperative variables such as patient characteristics, preexisting diseases, and intraoperative factors. Various algorithms, including logistic regression, random forest, light-gradient boosting machines, extreme-gradient boosting machines, and multilayer perceptrons, have been employed for model development and evaluation. This study enrolled 111,212 adult patients, with an overall incidence rate of 8.6% for developing PPCs. The area under the receiver-operating characteristic curve (AUROC) of the models was 0.699-0.767, and the f1 score was 0.446-0.526. In the prediction models, except for multilayer perceptron, the 10 most important features were obtained. In feature-reduced models, including 10 important features, the AUROC was 0.627-0.749, and the f1 score was 0.365-0.485. The number of packed red cells, urine, and rocuronium doses were similar in the three models. In conclusion, machine learning provides valuable insights into PPC prediction, significant features for prediction, and the feasibility of models that reduce the number of features.

Age-Related Variations in Postoperative Pain Intensity across 10 Surgical Procedures: A Retrospective Study of Five Hospitals in South Korea.

Age-related differences in pain perception have been reported in various contexts; however, their impact on postoperative pain intensity remains poorly understood, especially across different surgical procedures. Data from five hospitals were retrospectively analyzed, encompassing patients who underwent 10 distinct surgical procedures. Numeric rating scale scores were used to assess the worst postoperative pain intensity during the 24 h after surgery. The multivariate linear regression model analyzed the relationship between age and pain intensity. Subgroup analyses were performed according to sex and patient-controlled analgesia (PCA). This study included 41,187 patients. Among the surgeries studied, lumbar spine fusion (ß = -0.155, p < 0.001) consistently and significantly exhibited a decrease in worst postoperative pain with increasing age. Similar trends were observed in cholecystectomy (ß = -0.029, p < 0.001) and several other surgeries; however, the results were inconsistent across all analyses. Surgeries with higher percentages of PCA administration had lower median worst-pain scores. In conclusion, age may affect postoperative pain intensity after specific surgeries; however, a comprehensive understanding of the complex interplay between age, surgical intervention, and pain intensity is required. Pain management strategies should consider various factors, including age-related variations.

The Impact of Blood Transfusion in Developing Postoperative Delirium in Patients with Hip Fracture Surgery.

BACKGROUND: Many studies have been conducted to explore the risk factors associated with postoperative delirium (POD) in order to understand its underlying causes and develop prevention strategies, especially for hip fracture surgery. However, the relationship between blood transfusion and POD has been heatedly debated. The purpose of this study was to evaluate the risk factors of POD and the relationship between blood transfusions and the occurrence of POD in hip fracture surgery through big data analysis. METHODS: Medical data (including medication history, clinical and laboratory findings, and perioperative variables) were acquired from the clinical data warehouse (CDW) of the five hospitals of Hallym University Medical Center and were compared between patients without POD and with POD. RESULTS: The occurrence of POD was 18.7% (228 of 2398 patients). The risk factors of POD included old age (OR 4.38, 95% CI 2.77-6.91; p < 0.001), American Society of Anesthesiology physical status > 2 (OR 1.84 95% CI 1.4-2.42; p < 0.001), dementia (OR 1.99, 95% CI 1.53-2.6; p < 0.001), steroid (OR 0.53 95% CI 0.34-0.82; p < 0.001), Antihistamine (OR 1.53 95% CI 1.19-1.96; p < 0.001), and postoperative erythrocyte sedimentation rate (mm/h) (OR 0.97 95% CI 0.97-0.98; p < 0.001) in multivariate logistic regression analysis. The postoperative transfusion (OR 2.53, 95% CI 1.88-3.41; p < 0.001) had a significant effect on the incidence of POD. CONCLUSIONS: big data analytics using a CDW was a good option to identify the risk factors of POD and to prevent POD in hip fracture surgery.

Prediction of Postoperative Pulmonary Edema Risk Using Machine Learning.

Postoperative pulmonary edema (PPE) is a well-known postoperative complication. We hypothesized that a machine learning model could predict PPE risk using pre- and intraoperative data, thereby improving postoperative management. This retrospective study analyzed the medical records of patients aged > 18 years who underwent surgery between January 2011 and November 2021 at five South Korean hospitals. Data from four hospitals (n = 221,908) were used as the training dataset, whereas data from the remaining hospital (n = 34,991) were used as the test dataset. The machine learning algorithms used were extreme gradient boosting, light-gradient boosting machine, multilayer perceptron, logistic regression, and balanced random forest (BRF). The prediction abilities of the machine learning models were assessed using the area under the receiver operating characteristic curve, feature importance, and average precisions of precision-recall curve, precision, recall, f1 score, and accuracy. PPE occurred in 3584 (1.6%) and 1896 (5.4%) patients in the training and test sets, respectively. The BRF model exhibited the best performance (area under the receiver operating characteristic curve: 0.91, 95% confidence interval: 0.84-0.98). However, its precision and f1 score metrics were not good. The five major features included arterial line monitoring, American Society of Anesthesiologists physical status, urine output, age, and Foley catheter status. Machine learning models (e.g., BRF) could predict PPE risk and improve clinical decision-making, thereby enhancing postoperative management.

Effect of Pregnancy on Postoperative Nausea and Vomiting in Female Patients Who Underwent Nondelivery Surgery: Multicenter Retrospective Cohort Study.

Pregnant women usually have several risk factors of postoperative nausea and vomiting (PONV) and physiologic changes that make them susceptible to PONV development. We investigated the risk of PONV and postoperative vomiting (PV) in pregnant women in nondelivery surgery compared to nonpregnant women. This study included female adult patients who underwent nondelivery surgery at five hospitals between January 2011 and March 2021. To identify the association between pregnancy and PONV, logistic regression was used to calculate the odds ratio and 95% confidence intervals (CIs), adjusting for covariates. A total of 60,656 (nonpregnant women = 57,363 and pregnant women = 3293) complete patient outcomes and perioperative data were eligible for analysis. Although there was no significant association between pregnancy and PONV, the risk of PV in the pregnant women was 3.9-fold higher (95% confidence interval (95% CI), 3.06-4.97) than in the nonpregnant women. In addition, increased pregnancy duration increased the risk of PV (odds ratio (95% CI), 1.05 (1.01-1.09)) and preoperative nausea, and vomiting increased the risk of PONV (odds ratio (95% CI), 2.68 (1.30-5.54)) and PV (odds ratio (95% CI), 4.52 (2.36-8.69)). Pregnancy increased the risk of PV in female patients who underwent nondelivery surgery, and pregnancy duration and preoperative nausea and vomiting also were associated with PONV or PV.

Predictive model for difficult laryngoscopy using machine learning: retrospective cohort study

Abstract Background Both predictions and predictors of difficult laryngoscopy are controversial. Machine learning is an excellent alternative method for predicting difficult laryngoscopy. This study aimed to develop and validate practical predictive models for difficult laryngoscopy through machine learning. Methods Variables for the prediction of difficult laryngoscopy included age, Mallampati grade, body mass index, sternomental distance, and neck circumference. Difficult laryngoscopy was defined as grade 3 and 4 by the Cormack-Lehane classification. Pre-anesthesia and anesthesia data of 616 patients who had undergone anesthesia at a single center were included. The dataset was divided into a base training set (n = 492) and a base test set (n = 124), with equal distribution of difficult laryngoscopy. Training data sets were trained with six algorithms (multilayer perceptron, logistic regression, supportive vector machine, random forest, extreme gradient boosting, and light gradient boosting machine), and cross-validated. The model with the highest area under the receiver operating characteristic curve (AUROC) was chosen as the final model, which was validated with the test set. Results The results of cross-validation were best using the light gradient boosting machine algorithm with Mallampati score x age and sternomental distance as predictive model parameters. The predicted AUROC for the difficult laryngoscopy class was 0.71 (95% confidence interval, 0.59-0.83; p= 0.014), and the recall (sensitivity) was 0.85. Conclusion Predicting difficult laryngoscopy is possible with three parameters. Severe damage resulting from failure to predict difficult laryngoscopy with high recall is small with the reported model. The model's performance can be further enhanced by additional data training.

Effect of the preoperative physical status on postoperative nausea and vomiting risk: a matched cohort study.

BACKGROUND: The American Society of Anesthesiologists Physical Status Classification System is commonly used for preoperative assessment. Patient physical status before surgery can play an important role in postoperative nausea and vomiting. However, the relationship between the physical status classification and postoperative nausea and vomiting has not been well defined. METHODS: Adults aged ≥ 18 years who underwent procedures under anesthesia between 2015 and 2020 were included in the study. We analyzed the relationship of postoperative nausea and vomiting with physical status classification score using propensity score matching and Cox hazard regression. Differences in intraoperative use of vasopressor and inotropes and invasive monitoring were investigated according to the classification. RESULTS: A total of 163,500 patients were included in the study. After matching, classification 1 versus 2 included 43,400 patients; 1 versus ≤ 3, 13,287 patients; 2 versus ≤ 3, 23,530 patients (absolute standardized difference, 0-0.06). Patients with physical status classification ≤ 3 had a significantly lower postoperative nausea and vomiting risk than those with classification 1-2 (physical status classification 1 vs. ≤ 3, hazard ratio 0.76 [0.71-0.82], P < 0.001; 2 versus ≤ 3, hazard ratio 0.86 [0.82-0.91], P < 0.001). Intraoperative use of vasopressor or inotrope and invasive monitoring were noted more in the high physical status classification than the low physical status classification (absolute standardized difference [0.19-1.25]). CONCLUSION: There were differences in intraoperative invasive monitoring and use of vasopressor or inotrope among the classifications, and a score of 3 or higher reduced the risk of postoperative nausea and vomiting more than a score of 1-2.

The Influence of High Blood Pressure on Developing Symptomatic Lumbar Epidural Hematoma after Posterior Lumbar Spinal Fusion Surgery: Clinical Data Warehouse Analysis.

BACKGROUND: Determining the risk factors for symptomatic lumbar epidural hematoma (SLEH) is important for preventing postoperative SLEH. However, the relationship between blood pressure and SLEH is still debatable. The purpose of our study was to determine the risk factors for postoperative SLEH, to assess the influence of high blood pressure on developing SLEH after posterior lumbar spinal fusion surgery, and to evaluate the usefulness of big data analysis utilizing a clinical data warehouse (CDW). METHODS: The clinical data of patients who had undergone posterior lumbar spinal fusion surgery were acquired from the CDW of Hallym University Medical Center. The acquired clinical data were compared between patients without postoperative SLEH and with postoperative SLEH. RESULTS: Postoperative SLEH that required hematoma evacuation surgery within 72 h after posterior lumbar spinal fusion surgery occurred in 17 (1.3%) of 1313 patients. According to the multivariate logistic regression analysis, the risk factors for postoperative SLEH are platelet count difference (OR 1.28, p = 0.03), postoperative international normalized ratio (INR) difference (OR 31.4, p = 0.028), and postoperative systolic blood pressure (SBP) difference (≥10 mmHg) (OR 1.68, p = 0.048). An increase in postoperative SBP (OR 1.68, p = 0.048) had a statistically significant influence on the occurrence of postoperative SLEH. CONCLUSIONS: Big data analysis utilizing a CDW could be useful for extending our knowledge of the risk factors for postoperative SLEH and preventing postoperative SLEH after posterior lumbar spinal fusion surgery.

Preoperative Inflammatory Markers and the Risk of Postoperative Delirium in Patients Undergoing Lumbar Spinal Fusion Surgery.

We investigated the possible associations between postoperative delirium (POD) and routinely available preoperative inflammatory markers in patients undergoing lumbar spinal fusion surgery (LSFS) to explore the role of neuroinflammation and oxidative stress as risk factors for POD. We analyzed 11 years' worth of data from the Smart Clinical Data Warehouse. We evaluated whether preoperative inflammatory markers, such as the neutrophil-to-lymphocyte ratio (NLR), the monocyte-to-lymphocyte ratio (MLR), and the CRP-to-albumin ratio (CAR), affected the development of POD in patients undergoing LSFS. Of the 3081 subjects included, 187 (7.4%) developed POD. A significant increase in NLR, MLR, and CAR levels was observed in POD patients (p < 0.001). A multivariate analysis showed that the second, third, and highest quartiles of the NLR were significantly associated with the development of POD (adjusted OR (95% CI): 2.28 (1.25−4.16], 2.48 (1.3−4.73], and 2.88 (1.39−5.96], respectively). A receiver operating characteristic curve analysis showed that the discriminative ability of the NLR, MLR, and CAR for predicting POD was low, but almost acceptable (AUC (95% CI): 0.60 (0.56−0.64], 0.61 (0.57−0.65], and 0.63 (0.59−0.67], respectively, p < 0.001). Increases in preoperative inflammatory markers, particularly the NLR, were associated with the development of POD, suggesting that a proinflammatory state is a potential pathophysiological mechanism of POD.

Comparison of the Effect of Sugammadex and Pyridostigmine on Postoperative Catheter-Related Bladder Discomfort: A Retrospective Matched Cohort Analysis.

Background and Objectives: As the use of sugammadex for reversing neuromuscular blockade during general anesthesia increases, additional effects of sugammadex have been reported compared to cholinesterase inhibitors. Here, we compare the incidence of postoperative catheter-related bladder discomfort (CRBD) between sugammadex and pyridostigmine/glycopyrrolate treatments for reversing neuromuscular blockade. Materials and Methods: We retrospectively analyzed patients aged ≥ 18 years who underwent surgery under general anesthesia, received sugammadex or pyridostigmine with glycopyrrolate to reverse neuromuscular blockade, and had a urinary catheter in the post-anesthesia care unit between March 2019 and February 2021. After applying the exclusion criteria, 1179 patients were included in the final analysis. The incidence and severity of CRBD were collected from post-anesthesia recovery records. Results: The incidence was 13.7% in the sugammadex group (n = 211) and 24.7% in the pyridostigmine group (n = 968). Following propensity score matching, 211 patients each were included in the pyridostigmine and sugammadex matched group (absolute standardized difference (ASD), 0.01-0.05). Compared to the pyridostigmine group, the odds ratio for CRBD occurring in the sugammadex group was 0.568 (95% confidential interval, 0.316-1.021, p = 0.059). Conclusions: Sugammadex has a similar effect on the occurrence of postoperative CRBD compared with pyridostigmine.

Risk of Dementia According to Surgery Type: A Nationwide Cohort Study.

The relationship between dementia and surgery remains unclear. Research to elucidate the relationship between them is scarce, and conducting epidemiological research is complicated. This study aimed to investigate the incidence and risk of dementia according to the surgery type. We performed a retrospective propensity score-matched cohort study using nationwide representative cohort sample data from the Korean National Health Insurance Service in South Korea between 2003 and 2004. Incidence rates for dementia were obtained by dividing the number of patients with dementia by person-years at risk. To identify the risk of dementia according to the type of surgery, we investigated the hazard ratio by each surgery type. The incidence rates of dementia in control, musculoskeletal, and two or more surgeries groups were 9.66, 13.47, and 13.36 cases per 1000 person-years, respectively. The risk of dementia in the musculoskeletal and two or more surgeries groups was 1.44-fold higher (95% confidence interval (95% CI), 1.22-1.70) and 1.42-fold higher (95% CI, 1.17-1.72) than that in the control group, respectively. Patients who underwent musculoskeletal surgery and two or more surgeries had a higher risk of dementia; however, there was no association with the type of anesthesia administered.

The Relationship between Perioperative Blood Transfusion and Postoperative Delirium in Patients Undergoing Spinal Fusion Surgery: Clinical Data Warehouse Analysis.

Background and Objectives: For preventing postoperative delirium (POD), identifying the risk factors is important. However, the relationship between blood transfusion and POD is still controversial. The aim of this study was to identify the risk factors of POD, to evaluate the impact of blood transfusion in developing POD among people undergoing spinal fusion surgery, and to show the effectiveness of big data analytics using a clinical data warehouse (CDW). Materials and Methods: The medical data of patients who underwent spinal fusion surgery were obtained from the CDW of the five hospitals of Hallym University Medical Center. Clinical features, laboratory findings, perioperative variables, and medication history were compared between patients without POD and with POD. Results: 234 of 3967 patients (5.9%) developed POD. In multivariate logistic regression analysis, the risk factors of POD were as follows: Parkinson's disease (OR 5.54, 95% CI 2.15-14.27; p < 0.001), intensive care unit (OR 3.45 95% CI 2.42-4.91; p < 0.001), anti-psychotics drug (OR 3.35 95% CI 1.91-5.89; p < 0.001), old age (≥70 years) (OR 3.08, 95% CI 2.14-4.43; p < 0.001), depression (OR 2.8 95% CI 1.27-6.2; p < 0.001). The intraoperative transfusion (OR 1.1, 95% CI 0.91-1.34; p = 0.582), and the postoperative transfusion (OR 0.91, 95% CI 0.74-1.12; p = 0.379) had no statistically significant effect on the incidence of POD. Conclusions: There was no relationship between perioperative blood transfusion and the incidence of POD in spinal fusion surgery. Big data analytics using a CDW could be helpful for the comprehensive understanding of the risk factors of POD, and for preventing POD in spinal fusion surgery.

Interventions to Improve Prognostic Understanding in Advanced Stages of Life-Limiting Illness: A Systematic Review.

CONTEXT: Among patients with advanced life-limiting illness, an inaccurate understanding of prognosis is common and associated with negative outcomes. Recent years have seen an emergence of new interventions tested for their potential to improve prognostic understanding. However, this literature has yet to be synthesized. OBJECTIVES: To identify and characterize tested interventions, summarize intervention findings, and outline directions for future research. METHODS: Systematic searches were conducted in five databases - PubMed/MEDLINE (NLM), EMBASE (Elsevier), CINAHL (EBSCO), PsycINFO (OVID), and Cochrane Central Register of Controlled Trials (Wiley) - to identify interventions evaluated within a randomized controlled design for their impact on prognostic understanding. RESULTS: Of the 2354 initial articles, 12 were selected for final inclusion, representing nine unique interventions. Intervention types included decision aids accompanying medical consults, palliative care consultations, communication training for patients and physicians, and targeted discussions regarding prognosis and treatment decision making. Common components of interventions included provision of prognostic information, assistance with end-of-life care planning, and provision of decisional and emotional support during discussions. Most interventions were associated with some indication of improvement in prognostic understanding. However, even after intervention, inaccurate prognostic understanding was common, with 31-95 percent of patients in intervention groups exhibiting inaccurate perceptions of their prognosis. CONCLUSION: Prognostic understanding interventions hold the potential to improve patient understanding and thus informed decision making, but limitations exist. Future research should examine why many patients receiving intervention may continue to maintain inaccurate perceptions, and identify which intervention components can best enhance informed, value-consistent decision making.

Predictive model for difficult laryngoscopy using machine learning: retrospective cohort study.

BACKGROUND: Both predictions and predictors of difficult laryngoscopy are controversial. Machine learning is an excellent alternative method for predicting difficult laryngoscopy. This study aimed to develop and validate practical predictive models for difficult laryngoscopy through machine learning. METHODS: Variables for the prediction of difficult laryngoscopy included age, Mallampati grade, body mass index, sternomental distance, and neck circumference. Difficult laryngoscopy was defined as grade 3 and 4 by the Cormack-Lehane classification. Pre-anesthesia and anesthesia data of 616 patients who had undergone anesthesia at a single center were included. The dataset was divided into a base training set (n = 492) and a base test set (n = 124), with equal distribution of difficult laryngoscopy. Training data sets were trained with six algorithms (multilayer perceptron, logistic regression, supportive vector machine, random forest, extreme gradient boosting, and light gradient boosting machine), and cross-validated. The model with the highest area under the receiver operating characteristic curve (AUROC) was chosen as the final model, which was validated with the test set. RESULTS: The results of cross-validation were best using the light gradient boosting machine algorithm with Mallampati score x age and sternomental distance as predictive model parameters. The predicted AUROC for the difficult laryngoscopy class was 0.71 (95% confidence interval, 0.59-0.83; p =  0.014), and the recall (sensitivity) was 0.85. CONCLUSION: Predicting difficult laryngoscopy is possible with three parameters. Severe damage resulting from failure to predict difficult laryngoscopy with high recall is small with the reported model. The model's performance can be further enhanced by additional data training.