Breakthrough Pilot Study Ushers in Agentic AI for Personalized Preventive Health in Singapore

Singapore's push towards proactive healthcare has reached a pivotal milestone with the publication of a groundbreaking pilot study on agentic AI health plans. Titled "Personalised health plan development using agentic AI in Singapore’s national preventive care programme: a pilot study," the research appeared in npj Digital Medicine on March 9, 2026. Led by Han Leong Goh from Synapxe, Singapore's national health technology agency, the study demonstrates how multi-agent artificial intelligence (AI) systems can generate and refine tailored health plans, addressing the growing demand for preventive care amid an aging population.

This innovation aligns seamlessly with Healthier SG, the Ministry of Health's (MOH) flagship programme launched in 2023 to shift from reactive treatment to lifelong wellness. With over 1.3 million residents enrolled—representing more than half of eligible Singaporeans aged 40 and above—the programme emphasizes personalized Health Plans crafted by family doctors. The agentic AI pilot represents a technological leap, empowering both patients and clinicians to create actionable, dynamic plans focused on diet, exercise, and lifestyle modifications.

Understanding Healthier SG: Singapore's Blueprint for Preventive Healthcare

Healthier SG marks a paradigm shift in Singapore's healthcare landscape, prioritizing prevention over cure. Singaporeans and permanent residents aged 40 and older enroll with a preferred general practitioner (GP) to receive a Personalized Health Plan. This plan incorporates Healthier SG screenings, vaccinations, and lifestyle coaching tailored to individual risk factors like body mass index (BMI), blood pressure, and chronic disease predispositions.

Key components include:

• Annual health assessments subsidized for enrollees.
• Focus on six chronic conditions: diabetes, hypertension, dyslipidaemia, obesity, depression, and cognitive impairment.
• Integration with apps like HealthHub for tracking progress and accessing incentives via Healthy 365.

As of early 2026, enrolment continues to surge, with MOH aiming for universal coverage among eligible adults. The programme's success hinges on scalability, where AI steps in to augment GP capacity strained by workforce shortages.

Demystifying Agentic AI: From Reactive Tools to Autonomous Health Agents

Agentic AI refers to advanced systems comprising autonomous agents capable of planning, reasoning, and executing multi-step actions with minimal human oversight. Unlike traditional narrow AI, which follows predefined rules, agentic AI mimics human-like decision-making, adapting in real-time to new data.

In healthcare, examples include clinical scribes approved by Singapore's Health Sciences Authority (HSA) and predictive tools for patient triage. Synapxe, in partnership with OpenAI, is pioneering multi-agent frameworks to integrate into national platforms like HealthHub, enabling voice-driven outreach and plan refinement.

The pilot's multi-agent digital assistant exemplifies this: agents collaborate to ingest user profiles, generate initial plans, solicit feedback, and iterate—streamlining what traditionally requires multiple consultations.

The Multi-Agent Framework: How It Crafts Personalized Health Plans Step-by-Step

The study's digital assistant employs a proprietary multi-agent framework developed by Synapxe. While technical specifics remain confidential for proprietary reasons, the process unfolds as follows:

1. Input Ingestion: Collects user data—demographics, medical history, preferences, and goals—via conversational interfaces.
2. Plan Generation: Core agents synthesize evidence-based recommendations, drawing from guidelines like those from the Diabetes Prevention Program.
3. Refinement Loop: Feedback agents query users (e.g., "Is this exercise feasible?") and supervisor agents adjust plans iteratively.
4. Validation: Ensures alignment with MOH protocols, flagging risks for clinician review.

This closed-loop system produces granular plans, such as customized meal schedules or progressive exercise regimens, fostering adherence through personalization.

Pilot Study Design: Rigorous Testing with Residents and Clinicians

The pilot involved 20 residents (typical Healthier SG enrollees) and 7 clinicians from public healthcare institutions. Participants interacted with the AI assistant over simulated sessions, generating health plans focused on preventive lifestyle interventions.

Evaluation metrics assessed usability, personalization, granularity, and safety using Likert scales and sentiment analysis on feedback. Statistical significance was confirmed via p-values (e.g., personalization p=0.003), ensuring robust insights.

Diverse demographics ensured representativeness, mirroring Singapore's multi-ethnic population.

Impressive Results: High Acceptance and Positive Sentiment

Results exceeded expectations. Both cohorts rated success metrics (effectiveness, ease-of-use, trustworthiness, personalization) significantly above neutral (all p<0.05). Residents particularly praised granularity (p=0.0003) and voiced minimal concerns (p=0.941).

• Over 50% positive feedback on diet and exercise personalization.
• Clinicians noted reduced administrative burden, allowing focus on complex cases.
• Sentiment analysis: 6e-06 p-value for general features.

No adverse events reported, underscoring safety.

Implications for Singapore's Healthcare Ecosystem and Workforce

This pilot validates agentic AI's role in scaling preventive care, critical as Singapore's population ages—projected 1-in-4 over 65 by 2030. By augmenting GPs, it addresses shortages while empowering patients for self-management.

Integration with upcoming ACE-AI (rolling out 2027) for risk prediction promises end-to-end support: predict risks, generate plans, track adherence. For academics, it opens avenues in AI ethics, validation, and multimodal data fusion.Read the full study

University Research Driving AI Innovation in Health

Singapore's universities are at the forefront. Han Leong Goh, lead author and NUS Senior Adjunct Lecturer, bridges industry-academia. NUS iHealthtech and NTU's Centre for AI in Medicine (C-AIM) collaborate with Synapxe on challenges targeting chronic diseases affecting 1.8 million Singaporeans.

NTU's MSc in AI in Medicine trains the next generation, while joint initiatives like NUS-Synapxe-IMDA AI Challenge foster talent. These efforts position Singapore as a global leader in health AI.Explore higher ed jobs in AI health

Challenges, Ethical Considerations, and Safeguards

Despite promise, challenges persist: data privacy (addressed via federated learning), bias mitigation, and clinician oversight. Singapore's Model AI Governance Framework for Agentic AI, launched January 2026, provides guidelines for autonomy and accountability.

The pilot's clinician validation loop exemplifies human-in-the-loop design, ensuring AI augments rather than replaces expertise.

Future Outlook: Scaling Agentic AI Nationwide

MOH plans AI rollout under Healthier SG from 2027, potentially reaching millions. Synapxe's Tandem platform will host agentic tools, with expansions to mental health and elderly care. International collaborations could export this model.

For researchers, this paves the way for longitudinal studies on adherence and outcomes.Learn more about Healthier SG AI career advice in higher ed

Photo by Jiachen Lin on Unsplash

Stakeholder Perspectives and Broader Impact

Clinicians hailed time savings; residents appreciated empowerment. MOH's Ong Ye Kung emphasized predictive AI as healthcare's "next frontier." This aligns with Singapore's S$1B+ AI investments, fostering a vibrant ecosystem.

Explore opportunities at university jobs, higher ed jobs, or rate your professors.