PHOENIX (Nov. 12, 2026) – Since 2021, The Allergists' Foundation, the philanthropic arm of the American College of Allergy, Asthma and Immunology, has funded innovative projects led by practicing allergists that close health equity gaps and provide richer, fuller lives for patients.
Grants are awarded in the following two categories:
- Planning/exploratory projects (up to $10,000 for 12 months) to address issues that have not been fully explored, for which data needs to be gathered.
- Implementation grants (up to $25,000 for 24 months) for projects in which preliminary data has been gathered and initial planning has taken place. These grants help take the work to the "next level."
This year, three grantees will present their work at the ACAAI National Scientific Meeting in Phoenix. The awards are made possible thanks to generous donations from College members and support from GSK.
"Each year, the Allergists' Foundation receives a strong response to our call for proposals under the Community Grant Program," said allergist Todd Mahr, MD, ACAAI Executive Medical Director and chair of the Foundation's Grants Allocation Committee. "While that makes the job of selecting projects for funding challenging, it's a challenge we happily accept. We are impressed with the quality and creativity of the proposals we receive, and we look forward to reviewing more project ideas with each new grant cycle."
The abstracts for the grants being presented at the 2026 ACAAI Annual Scientific Meeting are below.
R322
PREPARING FOR THE UNEXPECTED: TACKLING ASTHMA AND ANAPHYLAXIS EMERGENCIES IN URBAN SCHOOLS
D. Pillay*1, M. Pansare2, M. El-Bohy2, D. Seth2, R. Revan2, 1. Flower Mound, TX; 2. Detroit, MI.
Introduction: The incidence of food allergies and asthma is rising in US school age children, as are classroom emergencies related to these conditions. Urban schools experience barriers to providing appropriate interventions such as inadequate training, staffing, and funding due to various social determinants of health. Our project focused on identifying and filling gaps in knowledge regarding asthma/anaphylaxis care and assessing changes in knowledge and comfort level after an educational intervention.
Methods: This was a cross-sectional study assessing the impact of hands-on education of non-clinical school staff and school nurses regarding asthma/anaphylaxis management in urban schools. Responses to a brief questionnaire collected pre- and post-intervention were used for analysis.
Results: There were 72 respondents in the non-clinical staff group and 168 in the school nurses' group. Following the intervention, non-clinical staff showed a percentage increase in comfort recognizing anaphylaxis of 28% (p=0.016), using an epinephrine auto-injector (EAI) of 45% (p=0.0001), recognizing asthma emergencies of 45% (p<0.0001), and using an albuterol inhaler correctly of 48% (p<0.0001). Correct identification of epinephrine as first-line anaphylaxis treatment also rose, from 57% to 89% (p=0.007) post-intervention. Among school nurses, comfort recognizing anaphylaxis increased by 11% (p=0.048), recognizing asthma emergencies by 13% (p=0.018), and correct EAI use by 12% (p=0.008).
Conclusion: There was a significant increase in comfort/knowledge after our intervention. Hands-on education by allergists is an effective tool for improving comfort/knowledge of clinical and non-clinical school staff regarding asthma/anaphylaxis management in the school setting.
R275
ASSESSING THE ROLE OF FOOD INSECURITY IN PEDIATRIC FOOD ALLERGY AT AN ACADEMIC ALLERGY CENTER
R. Moreno*1, J. Witonsky1, S. Cooper2, L. Sanchez1, 1. San Francisco, CA; 2. San Francisco, CA.
Introduction: Health disparities in food allergy (FA) have been well-documented, with disparate outcomes for patients who are low-income or a racial/ethnic minority. These groups are underdiagnosed and experience lower rates of oral food challenge (OFC) completion. Efforts to address barriers to FA care have been proposed, including food insecurity (FI) screening, but its effects on FA care have not been well described.
Methods: Retrospective review of pediatric FA patients ages 0-18 years presenting to a two-site academic pediatric allergy center from January 2018 - December 2025 (N=2869). Data included demographics and responses to the Hunger Vital Sign, a validated FI screening tool. OFC completion was determined by ICD codes. Bivariate analyses were conducted, using Chi-square or Fisher's exact test for categorical variables.
Results: The racial composition of patients differed by OFC completion status (p < 0.001), with patients who identified as Black (7.9% v. 2.6%) and Other (22.0% v. 12.9%) being over-represented in the OFC non-completion group compared to the completion group. Latine patients were over-represented in the non-completion group (21.4% v. 15.9%; p = 0.019). Publicly insured patients were over-represented in the non-completion group (27.8% v. 13.3%; p < 0.001). Completion status did not differ by FI status (p = 0.784) or language (English v. Non-English; p = 0.217).
Conclusion: While FI or language were not associated with differences in OFC completion rate, race/ethnicity and insurance were. Targeted interventions are needed to address disparities in OFC completion.
AI Use Statement
During the preparation of this work the author(s) used UCSF Chat GPT Enterprise in order to revise Python code for the analyses. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.
R147
DEVELOPMENT OF A MACHINE LEARNING ALGORITHM TO PREDICT 30-DAY EMERGENCY DEPARTMENT READMISSIONS FOR ASTHMA EXACERBATIONS
R. Lee, W. Hsu, J. Bernstein, P. Ryan, J. Bernstein*, Cincinnati, OH.
Introduction: Asthma affects over 4.5 million US children, with 1 in 12 requiring emergency department (ED) visits. This study sought to develop machine learning algorithms (MLAs) capable of predicting 30-day pediatric asthma ED readmissions
Methods: A retrospective analysis using de-identified data from a quaternary pediatric center was conducted for asthma-related ED visits in children 6-21 years old between 1/1/2010–12/31/2022. Information collected at admission included patient demographics, medications, systemic glucocorticoid use, asthma comorbidities, diagnostic data and environmental determinants (temperature, humidity, barometric pressure, pollen and mold counts, air quality index). R, Python and SAS were used to develop MLA models and conduct statistical analyses with data split into 80% for training and 20% for testing.
Results: Eight MLAs were developed using 31,779 individual ED encounters for 15,579 asthmatics to analyze a comprehensive model utilizing all variables versus a simple model using only patient demographics, medications, and ED admission season. The logistic mixed-effects (LG Mix) algorithm yielded the highest areas under the curve (AUC) for these comprehensive (0.897) and simple (0.897) training models. With test data, the support vector machine (SVM) algorithm generated the highest AUC of 0.741 for comprehensive and 0.745 for simple models. No statistically significant differences between the LG Mix and SVM algorithms were identified for the comprehensive (p = 0.796) and simple models (p= 0.398).
Conclusion: Using MLAs identified a simple model for identifying a pediatric asthma population at high-risk for 30-day ED readmission. These findings have important healthcare cost ramifications as more aggressive outpatient management could prevent overutilization of expensive hospital resources.
Simple and comprehensive test models similarly predict 30-day ED asthma readmission risk.
(Table here can be found in online abstracts)
Area under the curve (AUC) for comprehensive and simple test models similarly predict 30-day emergency department readmission rate. Empiric p-values > 0.05 indicate the observed AUC difference is not larger than expected under the null of no difference. LR = logistic regression, EN = elastic net, LASSO = least absolute shrinkage and selection operator, LG MIX = logistic mixed-effects, NN = neural network, XGB = XGBoost, SVM = support vector machine, KNN = K-nearest neighbor
