New research from Singapore is uncovering why standard allergy treatments do not always work for Asian patients, and how AI-driven data could change that.
Years ago, Ryan Lim was just another young adult constantly waking up congested
and fatigued.
Now, the 21-year-old is not just managing his allergic rhinitis (AR) better – he is helping build the solution.
Ryan is both a patient and volunteer under Project ENTenna, Asia’s first AI-trained allergic rhinitis database, in collaboration with major hospitals and research institutions, including NUH, TTSH, CGH, and A*STAR. To date, the programme has tracked over 1,600 patients, with the goal of tracking 6,000 patients from 2024 to 2026.
His story is part of a bigger national effort to reshape how AR is diagnosed, treated and understood in Singapore and in Asia.
AI and Millions of Data Points in Allergic Rhinitis
What does it mean to be an AI-trained, multimodal allergic rhinitis database? ENTenna integrates environmental data, genetics, symptom scores, treatment adherence and clinical notes into one living, evolving system, while following the patient database through their treatment journey.
“AR is a complex condition with thousands of contributing factors,” said Adjunct Associate Professor Ng Chew Lip, Senior Consultant, Department of Otolaryngology – Head & Neck Surgery (ENT), and Associate Chairman Medical Board (Care Transformation), NTFGH. He is the principal investigator of Project ENTenna.
“AI helps us make sense of all the data for us”
AI models trained on ENTenna’s data now provide clinicians with specialist-level insights, enabling personalised, preemptive care. Early results show a 45% increase in safe transitions from hospital to primary care, a 25% estimated cost avoidance, and medication adherence improvement from 40% to 65%.
Building Asia’s Allergy Database
Most allergy research is based on Western populations, where genetics, environment, and allergen exposures differ sharply.
“Asian allergies are not the same as Caucasian allergies,” Prof Ng said. “The pathways are different. Sensitisations are different.”
Singapore’s tropical climate brings exposure to unique allergens like dust mites and the seasonal haze, versus pollen and mould in more temperate countries. The ENTenna team is mapping genetic, clinical and environmental data to create a regionally relevant allergy roadmap, one they plan to share across ASEAN and South China.
Treatments for Allergic Rhinitis in Asians
Allergic rhinitis (AR) exists along a spectrum of severity. Clinicians often prescribe first-line therapies such as oral antihistamines and intranasal steroid sprays for patients with mild to moderate symptoms. Symptoms can also improve significantly with environmental control measures, such as minimising dust mite exposure or monitoring air quality. Once symptoms are stable, some patients may be able to gradually reduce or stop medication under medical supervision.
Clinicians may consider additional treatment modalities like sublingual immunotherapy (SLIT) and adjunctive surgical options for more persistent cases. By using ENTenna’s AI models to stratify patients by symptom intensity and medication adherence, clinicians can deliver more targeted and personalised treatments, ensuring each patient receives care appropriate to their condition’s complexity.
Additionally, through a partnership with A*STAR’s Singapore Immunology Network, the programme is collecting nasal swabs and blood samples to study AR at a molecular level. Using –onomics data: transcriptomics, metabolomics and genomics, the project hopes to uncover new biomolecular markers and treatment targets tailored to Asian patients.
Bringing Specialist-Level Insight to Every Clinic
ENTenna’s AI platform allows hospitals to clinicians monitor patients and stratify patients into high, moderate and low acuity groups. This system can then give the recommendation to GPs and community care partners like the polyclinics to manage milder cases, more effectively managing the patient load.
For national planners, population health maps superimpose symptom clusters against pollutant levels and patient density, helping target interventions potentially down to the neighbourhood.
On the clinician and patient level, visual dashboards show patterns like missed medication doses correlating with symptom spikes, which can advice early action and behavioural modification.
“Being part of the programme gave me control over my condition,” Ryan shared. It would also be evident if a medication was working or not working, which meant he and his doctor can switch to a different medication.
For someone who has lived with AR for nearly five years, that sense of control and understanding made a difference. “This is a good step in the right direction,” Ryan said.
The First Gen-AI Model for AR Care
At the heart of ENTenna is its most ambitious feature – a native in-house developed AI-powered Whatsapp chatbot. Built under stringent data governance and cybersecurity protocols, the system is designed to interact with patients in text-to-text and eventually voice-to-voice formats.
Currently in development and testing, the chatbot helps track weekly symptom and medication adherence. Patients like Ryan receive a combination of notifications to manage their condition:
- Weekly symptom trackers and medication adherence reporting
- Monthly standardised questionnaires
- AI-enabled behavioural nudges, leverage open source data like PSI levels
Preliminary data from ENTenna on 1,000 patients shows that such interventions have demonstrated an improvement from 40% to 65% in treatment adherence. This helps reduce unnecessary hospital visits through such proactive monitoring.
Ryan, who was part of this programme, shared: “As a patient, I appreciate how this study prioritises real patient needs like accessibility and education. The use of AI to personalise medication reminders and symptom tracking really makes a huge difference,”
Christine Wu, Assistant Director of Health Services Research and Analytics, explained. “Unlike general models like OpenAI, we work closely with clinicians to develop the chatbot.” The controlled and standardised chatbot means reduced hallucinations and ultimately patient safety.
Exciting features such as the text-to-text Q&A format of the Whatsapp chatbot are hoping to be rolled out by the end of year, after more testing and safety checks.
A Living Repository for Long-Term Change
More than a project, ENTenna is just the beginning.
As a nation-level initiative, it is in the works to be rolled out across the country beyond the lead site at Ng Teng Fong General Hospital (NTFGH) – from Tan Tock Seng Hospital (TTSH) and Changi General Hospital (CGH) – with the goal of standardising allergy care across Singapore.
But the ambition extends beyond national borders. With partnerships in South China and plans to engage other ASEAN countries, the ENTenna team hopes to build a regional allergy database that reflects the diverse genetics, environments and clinical needs of Asian populations.
At the same time, the team is preparing the platform to support other chronic conditions down the line, including asthma, tinnitus, giddiness, dementia and stroke. ENTenna is laying the digital, clinical and molecular foundation for the future of chronic disease care in Asia.
