Machine learning approaches for practical predicting outpatient near-future AECOPD based on nationwide electronic medical records.

In this research, we aimed to harness machine learning to predict the imminent risk of acute exacerbation in chronic obstructive pulmonary disease (AECOPD) patients. Utilizing retrospective data from electronic medical records of two Taiwanese hospitals, we identified 26 critical features. To predict 3- and 6-month AECOPD occurrences, we deployed five distinct machine learning algorithms alongside ensemble learning. The 3-month risk prediction was best realized by the XGBoost model, achieving an AUC of 0.795, whereas the XGBoost was superior for the 6-month prediction with an AUC of 0.813. We conducted an explainability analysis and found that the episode of AECOPD, mMRC score, CAT score, respiratory rate, and the use of inhaled corticosteroids were the most impactful features. Notably, our approach surpassed predictions that relied solely on CAT or mMRC scores. Accordingly, we designed an interactive prediction system that provides physicians with a practical tool to predict near-term AECOPD risk in outpatients.

Comparative efficacies of various corticosteroids for preventing postextubation stridor and reintubation: a systematic review and network meta-analysis.

Objectives: We assessed the efficacies of various corticosteroid treatments for preventing postexubation stridor and reintubation in mechanically ventilated adults with planned extubation. Methods: We searched the Pubmed, Embase, the Cochrane databases and ClinicalTrial.gov registration for articles published through September 29, 2022. Only randomized controlled trials (RCTs) that compared the clinical efficacies of systemic corticosteroids and other therapeutics for preventing postextubation stridor and reintubation were included. The primary outcome was postextubation stridor and the secondary outcome was reintubation. Results: The 11 assessed RCTs reported 4 nodes: methylprednisolone, dexamethasone, hydrocortisone, and placebo, which yielded 3 possible pairs for comparing the risks of post extubation stridor and 3 possible pairs for comparing the risks of reintubation. The risk of postextubation stridor was significantly lower in dexamethasone- and methylprednisolone-treated patients than in placebo-treated patients (dexamethasone: OR = 0.39; 95% CI = 0.22-0.70; methylprednisolone: OR = 0.22; 95% CI = 0.11-0.41). The risk of postextubation stridor was significantly lower in methylprednisolone-treated patients than in hydrocortisone-treated: OR = 0.24; 95% CI = 0.08-0.67) and dexamethasone-treated patients: OR = 0.55; 95% CI = 0.24-1.26). The risk of reintubation was significantly lower in dexamethasone- and methylprednisolone-treated patients than in placebo-treated patients: (dexamethasone: OR = 0.34; 95% CI = 0.13-0.85; methylprednisolone: OR = 0.42; 95% CI = 0.25-0.70). Cluster analysis showed that dexamethasone- and methylprednisolone-treated patients had the lowest risks of stridor and reintubation. Subgroup analyses of patients with positive cuff-leak tests showed similar results. Conclusions: Methylprednisolone and dexamethasone were the most effective agents against postextubation stridor and reintubation.

An artificial intelligence system to predict the optimal timing for mechanical ventilation weaning for intensive care unit patients: A two-stage prediction approach.

Background: For the intensivists, accurate assessment of the ideal timing for successful weaning from the mechanical ventilation (MV) in the intensive care unit (ICU) is very challenging. Purpose: Using artificial intelligence (AI) approach to build two-stage predictive models, namely, the try-weaning stage and weaning MV stage to determine the optimal timing of weaning from MV for ICU intubated patients, and implement into practice for assisting clinical decision making. Methods: AI and machine learning (ML) technologies were used to establish the predictive models in the stages. Each stage comprised 11 prediction time points with 11 prediction models. Twenty-five features were used for the first-stage models while 20 features were used for the second-stage models. The optimal models for each time point were selected for further practical implementation in a digital dashboard style. Seven machine learning algorithms including Logistic Regression (LR), Random Forest (RF), Support Vector Machines (SVM), K Nearest Neighbor (KNN), lightGBM, XGBoost, and Multilayer Perception (MLP) were used. The electronic medical records of the intubated ICU patients of Chi Mei Medical Center (CMMC) from 2016 to 2019 were included for modeling. Models with the highest area under the receiver operating characteristic curve (AUC) were regarded as optimal models and used to develop the prediction system accordingly. Results: A total of 5,873 cases were included in machine learning modeling for Stage 1 with the AUCs of optimal models ranging from 0.843 to 0.953. Further, 4,172 cases were included for Stage 2 with the AUCs of optimal models ranging from 0.889 to 0.944. A prediction system (dashboard) with the optimal models of the two stages was developed and deployed in the ICU setting. Respiratory care members expressed high recognition of the AI dashboard assisting ventilator weaning decisions. Also, the impact analysis of with- and without-AI assistance revealed that our AI models could shorten the patients' intubation time by 21 hours, besides gaining the benefit of substantial consistency between these two decision-making strategies. Conclusion: We noticed that the two-stage AI prediction models could effectively and precisely predict the optimal timing to wean intubated patients in the ICU from ventilator use. This could reduce patient discomfort, improve medical quality, and lower medical costs. This AI-assisted prediction system is beneficial for clinicians to cope with a high demand for ventilators during the COVID-19 pandemic.

Analysis of non-small cell lung cancer with miliary lung metastasis in patients harboring epidermal growth factor receptor mutations.

Miliary lung metastasis is a unique feature of lung metastasis in non-small cell lung cancer (NSCLC), indicating hematogenous dissemination. Some studies reported more frequency of epidermal growth factor receptor (EGFR) mutation and worse prognosis in these patients. Cases were identified from Chi-Mei medical center cancer registry for the period 2015-2019. Inclusion criteria were NSCLC with contra-lateral lung metastasis harboring EGFR mutation, under tyrosine kinase inhibitor (TKI) prescription. Patients with miliary or non-miliary lung metastasis were enrolled for survival analysis. 182 NSCLC patients were enrolled for assessing time to discontinuation of TKI (TD-TKI), progression-free survival (PFS) and overall survival (OS). 54 patients with miliary lung metastasis had average 13.2 months [95% confidence interval (CI) 10.7-15.6] of TD-TKI, 11.4 months (95% CI 9.3-13.6) of PFS, and 21.3 months (95% CI 16.8-25.8) of OS, which were shorter than non-miliary group with marginally statistical significance. In multivariate analysis, miliary lung metastasis had no statistical significance, and other strong prognostic indicators were found including performance status, liver metastasis, EGFR type, and generation of TKI. In NSCLC patients harboring EGRF mutation under TKI prescription, miliary lung metastasis was not a dominant indicator for outcomes evaluation.

Comparison of the risk of gastrointestinal perforation between patients with and without rheumatoid arthritis: A nationwide cohort study in Asia.

Objectives: Patients with rheumatoid arthritis (RA) may have an increased risk for gastrointestinal perforation (GIP) caused by medications or chronic inflammation. However, the risk of GIP between patients with and without RA remains unclear. Therefore, we conducted this study to clarify it. Methods: Using the Taiwan National Health Insurance Research Database, we identified patients with and without RA matched at 1:1 ratio by age, sex, and index date between 2000 and 2013 for this study. Comparison of the risk of GIP between the two cohorts was performed by following up until 2014 using Cox proportional hazard regression analyses. Results: In total, 11,666 patients with RA and an identical number of patients without RA were identified for this study. The mean age (±standard deviation) and female ratio were 55.3 (±15.2) years and 67.6% in both cohorts. Patients with RA had a trend of increased risk for GIP than patients without RA after adjusting for underlying comorbidities, medications, and monthly income [adjusted hazard ratio (AHR) 1.42; 95% confidence interval (CI) 0.99-2.04, p = 0.055]. Stratified analyses showed that the increased risk was significant in the female population (AHR 2.06; 95% CI 1.24-3.42, p = 0.005). Older age, malignancy, chronic obstructive pulmonary disease, and alcohol abuse were independent predictors of GIP; however, NSAIDs, systemic steroids, and DMARDs were not. Conclusion: RA may increase the risk of GIP, particularly in female patients. More attention should be paid in female population and those with independent predictors above for prevention of GIP.

The Impact of the Pay-for-Performance Program on the Outcome of COPD Patients in Taiwan After One Year.

Objective: To investigate the impact of a multidisciplinary intervention on the clinical outcomes of patients with COPD. Methods: This study retrospectively extracted the data of patients enrolled in the national pay-for-performance (P4P) program for COPD in four hospitals. Only COPD patients who received regular follow-up for at least one year in the P4P program between September 2018 and December 2020 were included. Results: A total of 1081 patients were included in this study. Among them, 424 (39.2%), 287 (26.5%), 179 (16.6%), and 191 (17.7%) patients were classified as COPD Groups A, B, C, and D, respectively. Dual therapy with long-acting ß2-agonist (LABA)/long-acting muscarinic antagonist (LAMA) was the most used inhaled bronchodilator at baseline (n = 477, 44.1%) patients, followed by LAMA monotherapy (n = 195, 18.0%), triple therapy with inhaled corticosteroid (ICS)/LABA/LAMA (n = 184, 17.0%), and ICS/LABA combination (n = 165, 15.3%). After one year of intervention, 374 (34.6%) and 323 (29.9%) patients had their pre- and post-bronchodilator-forced expiratory volume in one second (FEV1) increase of more than 100 mL. Both the COPD Assessment Test (CAT) and modified British Medical Research Council (mMRC) scores had a mean change of -2.2 ± 5.5 and -0.3 ± 0.9, respectively. The improvement in pulmonary function and symptom score were observed across four groups. The decreased number of exacerbations was only observed in Groups C and D, and not in Groups A and B. Conclusion: This real-world study demonstrated that the intervention in the P4P program could help improve the clinical outcome of COPD patients. It also showed us a different view on the use of dual therapy, which has a lower cost in Taiwan.

Development of an Interactive AI System for the Optimal Timing Prediction of Successful Weaning from Mechanical Ventilation for Patients in Respiratory Care Centers.

Successful weaning from prolonged mechanical ventilation (MV) is an important issue in respiratory care centers (RCCs). Delayed or premature extubation increases both the risk of adverse outcomes and healthcare costs. However, the accurate evaluation of the timing of successful weaning from MV is very challenging in RCCs. This study aims to utilize artificial intelligence algorithms to build predictive models for the successful timing of the weaning of patients from MV in RCCs and to implement a dashboard with the best model in RCC settings. A total of 670 intubated patients in the RCC in Chi Mei Medical Center were included in the study. Twenty-six feature variables were selected to build the predictive models with artificial intelligence (AI)/machine-learning (ML) algorithms. An interactive dashboard with the best model was developed and deployed. A preliminary impact analysis was then conducted. Our results showed that all seven predictive models had a high area under the receiver operating characteristic curve (AUC), which ranged from 0.792 to 0.868. The preliminary impact analysis revealed that the mean number of ventilator days required for the successful weaning of the patients was reduced by 0.5 after AI intervention. The development of an AI prediction dashboard is a promising method to assist in the prediction of the optimal timing of weaning from MV in RCC settings. However, a systematic prospective study of AI intervention is still needed.

Maintenance of low driving pressure in patients with early acute respiratory distress syndrome significantly affects outcomes.

BACKGROUND: Driving pressure (∆P) is an important factor that predicts mortality in acute respiratory distress syndrome (ARDS). We test the hypothesis that serial changes in daily ΔP rather than Day 1 ΔP would better predict outcomes of patients with ARDS. METHODS: This retrospective cohort study enrolled patients admitted to five intensive care units (ICUs) at a medical center in Taiwan between March 2009 and January 2018 who met the criteria for ARDS and received the lung-protective ventilation strategy. ∆P was recorded daily for 3 consecutive days after the diagnosis of ARDS, and its correlation with 60-day survival was analyzed. RESULTS: A total of 224 patients were enrolled in the final analysis. The overall ICU and 60-day survival rates were 52.7% and 47.3%, respectively. ∆P on Days 1, 2, and 3 was significantly lower in the survival group than in the nonsurvival group (13.8 ± 3.4 vs. 14.8 ± 3.7, p = 0.0322, 14 ± 3.2 vs. 15 ± 3.5, p = 0.0194, 13.6 ± 3.2 vs. 15.1 ± 3.4, p = 0.0014, respectively). The patients were divided into four groups according to the daily changes in ∆P, namely, the low ∆P group (Day 1 ∆P < 14 cmH2O and Day 3 ∆P < 14 cmH2O), decrement group (Day 1 ∆P ≥ 14 cmH2O and Day 3 ∆P < 14 cmH2O), high ∆P group (Day 1 ∆P ≥ 14 cmH2O and Day 3 ∆P ≥ 14 cmH2O), and increment group (Day 1 ∆P < 14 cmH2O and Day 3 ∆P ≥ 14 cmH2O). The 60-day survival significantly differed among the four groups (log-rank test, p = 0.0271). Compared with the low ΔP group, patients in the decrement group did not have lower 60-day survival (adjusted hazard ratio 0.72; 95% confidence interval [CI] 0.31-1.68; p = 0.4448), while patients in the increment group had significantly lower 60-day survival (adjusted hazard ratio 1.96; 95% CI 1.11-3.44; p = 0.0198). CONCLUSIONS: Daily ∆P remains an important predicting factor for survival in patients with ARDS. Serial changes in daily ΔP might be more informative than a single Day 1 ΔP value in predicting survival of patients with ARDS.

Correlation of Body Mass Index with Oncologic Outcomes in Colorectal Cancer Patients: A Large Population-Based Study.

It has been acknowledged that excess body weight increases the risk of colorectal cancer (CRC); however, there is little evidence on the impact of body mass index (BMI) on CRC patients' long-term oncologic results in Asian populations. We studied the influence of BMI on overall survival (OS), disease-free survival (DFS), and CRC-specific survival rates in CRC patients from the administrative claims datasets of Taiwan using the Kaplan-Meier survival curves and the log-rank test to estimate the statistical differences among BMI groups. Underweight patients (<18.50 kg/m2) presented higher mortality (56.40%) and recurrence (5.34%) rates. Besides this, they had worse OS (aHR:1.61; 95% CI: 1.53-1.70; p-value: < 0.0001) and CRC-specific survival (aHR:1.52; 95% CI: 1.43-1.62; p-value: < 0.0001) rates compared with those of normal weight patients (18.50-24.99 kg/m2). On the contrary, CRC patients belonging to the overweight (25.00-29.99 kg/m2), class I obesity (30.00-34.99 kg/m2), and class II obesity (≥35.00 kg/m2) categories had better OS, DFS, and CRC-specific survival rates in the analysis than the patients in the normal weight category. Overweight patients consistently had the lowest mortality rate after a CRC diagnosis. The associations with being underweight may reflect a reverse causation. CRC patients should maintain a long-term healthy body weight.

Incidence and management of metabolic acidosis with sodium bicarbonate in the ICU: An international observational study.

BACKGROUND: Metabolic acidosis is a major complication of critical illness. However, its current epidemiology and its treatment with sodium bicarbonate given to correct metabolic acidosis in the ICU are poorly understood. METHOD: This was an international retrospective observational study in 18 ICUs in Australia, Japan, and Taiwan. Adult patients were consecutively screened, and those with early metabolic acidosis (pH < 7.3 and a Base Excess < -4 mEq/L, within 24-h of ICU admission) were included. Screening continued until 10 patients who received and 10 patients who did not receive sodium bicarbonate in the first 24 h (early bicarbonate therapy) were included at each site. The primary outcome was ICU mortality, and the association between sodium bicarbonate and the clinical outcomes were assessed using regression analysis with generalized linear mixed model. RESULTS: We screened 9437 patients. Of these, 1292 had early metabolic acidosis (14.0%). Early sodium bicarbonate was given to 18.0% (233/1292) of these patients. Dosing, physiological, and clinical outcome data were assessed in 360 patients. The median dose of sodium bicarbonate in the first 24 h was 110 mmol, which was not correlated with bodyweight or the severity of metabolic acidosis. Patients who received early sodium bicarbonate had higher APACHE III scores, lower pH, lower base excess, lower PaCO2, and a higher lactate and received higher doses of vasopressors. After adjusting for confounders, the early administration of sodium bicarbonate was associated with an adjusted odds ratio (aOR) of 0.85 (95% CI, 0.44 to 1.62) for ICU mortality. In patients with vasopressor dependency, early sodium bicarbonate was associated with higher mean arterial pressure at 6 h and an aOR of 0.52 (95% CI, 0.22 to 1.19) for ICU mortality. CONCLUSIONS: Early metabolic acidosis is common in critically ill patients. Early sodium bicarbonate is administered by clinicians to more severely ill patients but without correction for weight or acidosis severity. Bicarbonate therapy in acidotic vasopressor-dependent patients may be beneficial and warrants further investigation.