Abstract:Background: Asthma is defined by the Global Initiative for Asthma (GINA) as a heterogeneous disease characterized by chronic airway inflammation. The pathogenesis of the disease is better understood with the comprehension of immunological pathways. These pathways differ by the type of recruited cells and released interleukin (IL). Thus, asthma can be classified into subtypes based on the underlying immune mechanism: eosinophilic asthma (EA) and non-eosinophilic asthma (NEA). Patients with EA tend to respond better to inhaled corticosteroid as compared to those with NEA. The distinction of EA is very important in the light of emergent type 2 inflammation targeted therapies. Methods: We performed a 1-year (2018) retrospective cohort analysis of the Nationwide Inpatient Database (NIS). We included all adult patients presenting with severe asthma. Patients were stratified into two groups: eosinophilic severe asthma and non-eosinophilic severe asthma. The primary outcomes measures were the prevalence of chronic steroid use, status asthmaticus, family history of asthma, food, drug and environmental allergies, presence of nasal polyps, allergic rhinitis, allergic dermatitis, need for mechanical ventilation, need for oxygen supplementation, gastroesophageal reflux disease, in-hospital mortality, and length of stay. We performed descriptive statistics. Continuous parametric variables were reported using a mean and standard deviation. Continuous nonparametric variables were reported using a median and interquartile range. To compare the characteristics of the two groups, we used the independent t-test for continuous parametric variables and the Mann-Whitney U test for continuous nonparametric variables. The Chi-square test was used to assess differences in categorical variables. Results: A total of 2,646 patients were included, out of which 882 belonged to the eosinophilic group and 1,764 were in the non-eosinophilic group. Comparing EA versus NEA, we have found that eosinophilic group was characterized by higher percentage of steroid use (18.3% vs. 9.5%, P < 0.001). This group also had higher rates of status asthmaticus and positive family history (P = 0.009 and 0.004, respectively). The presence of allergies, allergic rhinitis, nasal polyps, and allergic dermatitis was higher among patients with eosinophilia. The need for mechanical ventilation and supplemental oxygen was also higher among this group (P < 0.001 for both); however, there was no significant difference in mortality rate (P = 0.347) and the length of hospital stay was similar in both groups (P < 0.001). Conclusion: We showed herein that the eosinophilic subtype of asthma differs widely from the non-eosinophilic phenotype. Clinically, patients with eosinophilia might exhibit different symptomatology, more atopy, and concomitant comorbidities. However, this group might have better response to steroid therapy and might benefit from the new emergent T2 immune targeted therapy. The identification of EA is crucial for better disease control.

S573 Predictive Artificial Intelligence Analysis of Determinants of Health Outcomes in Hospitalized Patients with Eosinophilic Esophagitis

Harnessing Digital Pathology And Causal Learning To Improve Eosinophilic Esophagitis Dietary Treatment Assignment

Harnessing artificial intelligence to infer novel spatial biomarkers for the diagnosis of eosinophilic esophagitis

Evaluating the Impact of Social Determinants on Health Prediction in the Intensive Care Unit

Assessing socioeconomic bias in machine learning algorithms in health care: a case study of the HOUSES index

Explainable artificial intelligence model to predict acute critical illness from electronic health records

S690 Predictors of Patients Receiving No Medication for Treatment of Eosinophilic Esophagitis in the United States: Data from the TARGET-EGIDS Cohort

Social and Behavioral Determinants of Health in the Era of Artificial Intelligence with Electronic Health Records: A Scoping Review

The Determinants of Eosinophilia in Patients With Severe Asthma

A preliminary review of the utility of artificial intelligence to detect eosinophilic chronic rhinosinusitis

Development and technical validation of an artificial intelligence model for quantitative analysis of histopathologic features of eosinophilic esophagitis

EHRs Connect Research and Practice: Where Predictive Modeling, Artificial Intelligence, and Clinical Decision Support Intersect

Eosinophil may be a predictor of immune-related adverse events induced by different immune checkpoint inhibitor types: A retrospective multidisciplinary study

Predictive modeling for eosinophilic chronic rhinosinusitis: Nomogram and four machine learning approaches

Preadmission assessment of extended length of hospital stay with RFECV-ETC and hospital-specific data

Abstract 18252: Comparison of an Artificial Intelligence Heart Failure Detection Model and Clinical Prediction Models in Patients With Heart Failure With Preserved Ejection Fraction (HFPEF)

Predictors and impact of aspiration pneumonia in patients undergoing esophagogastroduodenoscopy: national inpatient sample 2016–2020

Augmenting electronic health record data with social and environmental determinant of health measures to understand regional factors associated with asthma exacerbations

Characterization and Racial Stratification of Social Determinants of Health for Individuals with Type 2 Diabetes as Recorded in Electronic Health Records: Implications for Artificial Intelligence Development.

Artificial Intelligence for Inflammatory Bowel Diseases (IBD); Accurately Predicting Adverse Outcomes Using Machine Learning

Noninvasive biomarkers identify eosinophilic esophagitis: A prospective longitudinal study in children