Novarea Inc. is proud to announce that Hekma Smart Solutions, as our Representative in the GCC region, has been awarded a Research Grant on AI in Childhood Asthma from the National University of Science and Technology in Oman. The main goal of this project is to develop a Risk Prediction model in Childhood Asthma, based on Machine Learning.
Dr. Prashanth Gowda, Associate Professor and Head of the Pediatrics Department at the College of Medicine, National University of Science and Technology, is the Principal Investigator of this project and the team leader.
This achievement reflects months of dedicated work by the research team and the trust that Oman National University has placed in our scientific vision and technical expertise.
Why This Project Matters
Asthma is one of the most common chronic respiratory diseases worldwide, affecting over 339 million people. In children, the burden is especially heavy—on healthcare systems, families, and the children themselves.
Recent studies estimate the global prevalence of childhood asthma at approximately 10.2%, with most cases beginning in the first five years of life. In Oman, asthma remains a major public health concern.
Unfortunately:
• There is no cure for asthma.
• Early prediction tools are very limited and rarely usable in clinical settings.
• Children under 10 years carry the highest risk but have the fewest preventive tools available.
This is where machine learning can create a real transformation.
What Is the Goal of This Project?
We aim to design and validate a machine learning–based application that predicts the risk of developing asthma in children under 10 years by integrating clinical, family, environmental, and even wearable data.
Key features of this application:
• Early asthma risk prediction (by age 4)
• Uses longitudinal data (from birth to age 4)
• Model transparency with SHAP (interpretable for clinicians)
• Fairness assessment across sex, ancestry, and socioeconomic groups
• Ready for deployment in primary care and clinical settings
What Methods Are We Using?
Our team will use state of the art machine learning algorithms, including:
• XGBoost
• TabNet
• Support Vector Machine (SVM)
• Deep Neural Networks
Data will be sourced from validated birth cohorts and electronic health records (EHRs). For handling missing data, we will apply Multiple Imputation by Chained Equations (MICE).
Beyond accuracy (AUC), we will assess real clinical utility using Decision Curve Analysis (DCA).
Key Research Questions
-
- Can machine learning models using data up to age 4 accurately predict asthma risk between ages 4 and 8 years?
-
- Which early-life determinants (e.g., wheeze, maternal asthma, infection, antibiotics, atopy) best predict asthma risk?
-
- Which machine learning algorithm performs best in terms of accuracy, calibration, and clinical utility?
-
- Is the final model fair and interpretable across subgroups (sex, ancestry, socioeconomic status)?
Expected Outcomes and Impact
• Early identification of at-risk children – before chronic disease develops
• Reduced clinician workload – especially in busy hospitals
• Fewer unnecessary prescriptions and emergency hospitalizations
• A simple, transparent, and fair clinical tool
• A framework for ethical, interpretable AI in pediatric medicine
Acknowledgment
We extend our deepest gratitude to Oman National University for making this important project possible through their trust and support. We also thank all our research collaborators, statistical advisors, and the entire Hekma Smart Solutions team for their hard work in preparing this successful grant proposal.
Get Involved
If you are a researcher, pediatrician, investor, or AI health enthusiast and would like to collaborate with us on this journey:
Send us an email
Visit the Hekma Smart Solutions website
The full research protocol and future publications will be shared soon.
Together, we will build a future where fewer children suffer from asthma.
Hekma Smart Solutions – AI for a better, healthier life.