Predicting Sudden Sensorineural Hearing Loss Recovery with Patient-Personalized Seigel's Criteria Using Machine Learning.

BACKGROUND: Accurate prognostic prediction is crucial for managing Idiopathic Sudden Sensorineural Hearing Loss (ISSHL). Previous studies developing ISSHL prognosis models often overlooked individual variability in hearing damage by relying on fixed frequency domains. This study aims to develop models predicting ISSHL prognosis one month after treatment, focusing on patient-specific hearing impairments. METHODS: Patient-Personalized Seigel's Criteria (PPSC) were developed considering patient-specific hearing impairment related to ISSHL criteria. We performed a statistical test to assess the shift in the recovery assessment when applying PPSC. The utilized dataset of 581 patients comprised demographic information, health records, laboratory testing, onset and treatment, and hearing levels. To reduce the model's reliance on hearing level features, we used only the averages of hearing levels of the impaired frequencies. Then, model development, evaluation, and interpretation proceeded. RESULTS: The chi-square test (p-value: 0.106) indicated that the shift in recovery assessment is not statistically significant. The soft-voting ensemble model was most effective, achieving an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.864 (95% CI: 0.801-0.927), with model interpretation based on the SHapley Additive exPlanations value. CONCLUSIONS: With PPSC, providing a hearing assessment comparable to traditional Seigel's criteria, the developed models successfully predicted ISSHL recovery one month post-treatment by considering patient-specific impairments.

A machine learning-based approach for predicting renal function recovery in general ward patients with acute kidney injury.

BACKGROUND: Acute kidney injury (AKI) is a significant challenge in healthcare. While there are considerable researches dedicated to AKI patients, a crucial factor in their renal function recovery, is often overlooked. Thus, our study aims to address this issue through the development of a machine learning model to predict restoration of kidney function in patients with AKI. METHODS: Our study encompassed data from 350,345 cases, derived from three hospitals. AKI was classified in accordance with the Kidney Disease: Improving Global Outcomes. Criteria for recovery were established as either a 33% decrease in serum creatinine levels at AKI onset, which was initially employed for the diagnosis of AKI. We employed various machine learning models, selecting 43 pertinent features for analysis. RESULTS: Our analysis contained 7,041 and 2,929 patients' data from internal cohort and external cohort respectively. The Categorical Boosting Model demonstrated significant predictive accuracy, as evidenced by an internal area under the receiver operating characteristic (AUROC) of 0.7860, and an external AUROC score of 0.7316, thereby confirming its robustness in predictive performance. SHapley Additive exPlanations (SHAP) values were employed to explain key factors impacting recovery of renal function in AKI patients. CONCLUSION: This study presented a machine learning approach for predicting renal function recovery in patients with AKI. The model performance was assessed across distinct hospital settings, which revealed its efficacy. Although the model exhibited favorable outcomes, the necessity for further enhancements and the incorporation of more diverse datasets is imperative for its application in real- world.

Hearing Recovery Prediction for Patients with Chronic Otitis Media Who Underwent Canal-Wall-Down Mastoidectomy.

Background: Chronic otitis media affects approximately 2% of the global population, causing significant hearing loss and diminishing the quality of life. However, there is a lack of studies focusing on outcome prediction for otitis media patients undergoing canal-wall-down mastoidectomy. Methods: This study proposes a recovery prediction model for chronic otitis media patients undergoing canal-wall-down mastoidectomy, utilizing data from 298 patients treated at Korea University Ansan Hospital between March 2007 and August 2020. Various machine learning techniques, including logistic regression, decision tree, random forest, support vector machine (SVM), extreme gradient boosting (XGBoost), and light gradient boosting machine (light GBM), were employed. Results: The light GBM model achieved a predictive value (PPV) of 0.6945, the decision tree algorithm showed a sensitivity of 0.7574 and an F1 score of 0.6751, and the light GBM algorithm demonstrated the highest AUC-ROC values of 0.7749 for each model. XGBoost had the most efficient PR-AUC curve, with a value of 0.7196. Conclusions: This study presents the first predictive model for chronic otitis media patients undergoing canal-wall-down mastoidectomy. The findings underscore the potential of machine learning techniques in predicting hearing recovery outcomes in this population, offering valuable insights for personalized treatment strategies and improving patient care.

Prediction Model of Acute Respiratory Failure in Patients with Acute Pesticide Poisoning by Intentional Ingestion: Prediction of Respiratory Failure in Pesticide Intoxication (PREP) Scores in Cohort Study.

Acute respiratory failure is the primary cause of mortality in patients with acute pesticide poisoning. The aim of the present study was to develop a new and efficient score system for predicting acute respiratory failure in patients with acute pesticide poisoning. This study was a retrospective observational cohort study comprised of 679 patients with acute pesticide poisoning by intentional poisoning. We divided this population into a ratio of 3:1; training set (n = 509) and test set (n = 170) for model development and validation. Multivariable logistic regression models were used in developing a score-based prediction model. The Prediction of Respiratory failure in Pesticide intoxication (PREP) scoring system included a summation of the integer scores of the following five variables; age, pesticide category, amount of ingestion, Glasgow Coma Scale, and arterial pH. The PREP scoring system developed accurately predicted respiratory failure (AUC 0.911 [0.849-0.974], positive predictive value 0.773, accuracy 0.873 in test set). We came up with four risk categories (A, B, C and D) using PREP scores 20, 40 and 60 as the cut-off for mechanical ventilation requirement risk. The PREP scoring system developed in the present study could predict respiratory failure in patients with pesticide poisoning, which can be easily implemented in clinical situations. Further prospective studies are needed to validate the PREP scoring system.

Arrhythmia and Heart Rate Variability during Long Interdialytic Periods in Patients on Maintenance Hemodialysis: Prospective Observational Cohort Study.

Sudden cardiac death among hemodialysis patients is related to the hemodialysis schedule. Mortality is highest within 12 h before and after the first hemodialysis sessions of a week. We investigated the association of arrhythmia occurrence and heart rate variability (HRV) using an electrocardiogram (ECG) monitoring patch during the long interdialytic interval in hemodialysis patients. This was a prospective observational study with 55 participants on maintenance hemodialysis for at least six months. A patch-type ECG monitoring device was applied to record arrhythmia events and HRV during 72 h of a long interdialytic period. Forty-nine participants with sufficient ECG data out of 55 participants were suitable for the analysis. The incidence of supraventricular tachycardia and ventricular tachycardia did not significantly change over time. The square root of the mean squared differences of successive NN intervals (RMSSD), the proportion of adjacent NN intervals differing by >50 ms (pNN50), and high-frequency (HF) increased during the long interdialytic interval. The gap in RMSSD, pNN50, HF, and the low-frequency/high-frequency (LF/HF) ratio between patients with and without significant arrhythmias increased significantly over time during the long interdialytic interval. The daily changes in RMSSD, pNN50, HF, and the LF/HF ratio were more prominent in patients without significant arrhythmias than in those with significant arrhythmias. The electrolyte fluctuation between post-hemodialysis and subsequent pre-hemodialysis was not considered in this study. The study results suggest that the decreased autonomic response during interdialytic periods in dialysis patients is associated with poor cardiac arrhythmia events.

Prediction of In-Hospital Cardiac Arrest Using Shallow and Deep Learning.

Sudden cardiac arrest can leave serious brain damage or lead to death, so it is very important to predict before a cardiac arrest occurs. However, early warning score systems including the National Early Warning Score, are associated with low sensitivity and false positives. We applied shallow and deep learning to predict cardiac arrest to overcome these limitations. We evaluated the performance of the Synthetic Minority Oversampling Technique Ratio. We evaluated the performance using a Decision Tree, a Random Forest, Logistic Regression, Long Short-Term Memory model, Gated Recurrent Unit model, and LSTM-GRU hybrid models. Our proposed Logistic Regression demonstrated a higher positive predictive value and sensitivity than traditional early warning systems.

Impact of Acid-Base Status on Mortality in Patients with Acute Pesticide Poisoning.

We investigated clinical impacts of various acid-base approaches (physiologic, base excess (BE)-based, and physicochemical) on mortality in patients with acute pesticide intoxication and mutual intercorrelated effects using principal component analysis (PCA). This retrospective study included patients admitted from January 2015 to December 2019 because of pesticide intoxication. We compared parameters assessing the acid-base status between two groups, survivors and non-survivors. Associations between parameters and 30-days mortality were investigated. A total of 797 patients were analyzed. In non-survivors, pH, bicarbonate concentration (HCO3-), total concentration of carbon dioxide (tCO2), BE, and effective strong ion difference (SIDe) were lower and apparent strong ion difference (SIDa), strong ion gap (SIG), total concentration of weak acids, and corrected anion gap (corAG) were higher than in survivors. In the multivariable logistic analysis, BE, corAG, SIDa, and SIDe were associated with mortality. PCA identified four principal components related to mortality. SIDe, HCO3-, tCO2, BE, SIG, and corAG were loaded to principal component 1 (PC1), referred as total buffer bases to receive and handle generated acids. PC1 was an important factor in predicting mortality irrespective of the pesticide category. PC3, loaded mainly with pCO2, suggested respiratory components of the acid-base system. PC3 was associated with 30-days mortality, especially in organophosphate or carbamate poisoning. Our study showed that acid-base abnormalities were associated with mortality in patients with acute pesticide poisoning. We reduced these variables into four PCs, resembling the physicochemical approach, revealed that PCs representing total buffer bases and respiratory components played an important role in acute pesticide poisoning.

New Model for Predicting the Presence of Coronary Artery Calcification.

Coronary artery calcification (CAC) is a feature of coronary atherosclerosis and a well-known risk factor for cardiovascular disease (CVD). As the absence of CAC is associated with a lower incidence rate of CVD, measurement of a CAC score is helpful for risk stratification when the risk decision is uncertain. This was a retrospective study with an aim to build a model to predict the presence of CAC (i.e., CAC score = 0 or not) and evaluate the discrimination and calibration power of the model. Our data set was divided into two set (80% for training set and 20% for test set). Ten-fold cross-validation was applied with ten times of interaction in each fold. We built prediction models using logistic regression (LRM), classification and regression tree (CART), conditional inference tree (CIT), and random forest (RF). A total of 3,302 patients from two cohorts (Soonchunhyang University Cheonan Hospital and Kangbuk Samsung Health Study) were enrolled. These patients' ages were between 40 and 75 years. All models showed acceptable accuracies (LRM, 70.71%; CART, 71.32%; CIT, 71.32%; and RF, 71.02%). The decision tree model using CART and CIT showed a reasonable accuracy without complexity. It could be implemented in real-world practice.