Healthcare was one of three priority areas (alongside agriculture and smart cities) targeted for early delivery of National AI Use Cases in Malaysia's National AI Roadmap 2021–2025. The roadmap framed AI in health around supply-chain efficiency, prevention, and diagnostic imaging, naming four use cases. It is worth reading the original framing carefully, because the gap between these descriptions and what a ministry can actually procure is the story of this page.
Project 6 — Autonomous Vaccine Distribution & Management
Conceived during COVID-19, this use case aimed to boost the operational efficiency of vaccine distribution — targeting a 20% reduction in supply spend and a 40% reduction in logistics cost — by overlaying an AI layer on existing workflows so that dose scheduling could be administered autonomously at scale.
Project 7 — Personalized Proactive Healthcare
A national proactive-healthcare strategy for cardiovascular disease, intended to shift care upstream and reduce overall healthcare cost. Cardiovascular disease has been Malaysia's leading cause of death for decades, so the target was well chosen; the delivery mechanism was never specified beyond the concept.
Project 8 — Autonomous A-eye System
An autonomous system to help prevent blindness using AI-powered retinal image analysis. The clinical context matters here: diabetic retinopathy is the leading cause of visual loss among working-age Malaysians, and the MOH already runs manual retinal-photography screening for diabetic patients through Klinik Kesihatan. The roadmap's bet was that AI grading could be layered onto that existing fundus-camera workflow.
Project 9 — AI-Nasoalveolar (AI-Na) System
An AI-assisted predictive model for 3D-printed presurgical nasoalveolar moulds, predicting changes in cleft separation after treatment — the most modest and most clinically specific of the four.
What actually happened, 2021–2025
Healthcare is the sector where the roadmap's framing collided most directly with events. Project 6 was written during the pandemic and effectively delivered by the pandemic — but through a different vehicle than the roadmap imagined. And the period's real achievement is one the roadmap barely mentions: getting a paper-based health system onto digital records at all.
Vaccine logistics: solved operationally, not as a roadmap project
Malaysia's COVID-19 vaccination programme ran through MySejahtera, owned by the Government and managed by the Ministry of Health, coordinated by the JKJAV special committee. By March 2022, around 98.7% of the adult population had received two doses[1] — from a near-standing start thirteen months earlier, with single-day throughput exceeding half a million doses at the August 2021 peak. As a logistics achievement it stands comparison with anything in the region.
But look at how it was done: appointment scheduling, call-centre outreach, manual supply coordination, PPV mega-centres, and a great deal of human improvisation. The “autonomous AI-scheduling layer” of Project 6 played no documented role, and the project's 20%/40% cost-reduction targets have never been publicly measured or reported against. The honest verdict is that the problem Project 6 existed to solve was solved — conventionally, under pressure, and faster than any AI procurement cycle could have moved.
The real delivery: EMR digitisation and a national health record
What the Ministry of Health actually built over the roadmap period is a multi-year digitisation programme the roadmap did not foreground. By end-2025, 412 government health clinics ran the Cloud-Based Clinic Management System (CCMS), and the MOH is consolidating data toward a “One Individual, One Record” national health record — integrating 30 million vaccination records, 7 million prescription records and, in early 2026, 90,000 medical-imaging records into MySejahtera.[2] Hospital digitisation lags well behind: only 23 hospitals run the Total Hospital Information System, with full coverage not targeted until end-2029.[2]
This sequencing is the single most important fact for judging the roadmap's healthcare chapter. Diagnostic AI — the A-eye vision of Project 8, and every imaging use case after it — consumes digital records and standardised images. In 2021 the roadmap assumed that backbone; in 2026 the MOH is still laying it. The roadmap did not fail to deliver AI so much as it skipped the decade of plumbing that AI presupposes.
Project-by-project verdicts
Scored against what the roadmap actually wrote, using the public record as at July 2026. Where no public delivery trail exists, we say so rather than infer failure — but the absence of any published measurement against the roadmap's own numeric targets is itself a finding.
The underlying mission — vaccinate the country efficiently — was accomplished at remarkable speed, but by conventional logistics under emergency coordination. No public evidence that the AI layer was built, or that the −20% supply / −40% logistics targets were ever measured. A useful lesson in what crisis-mode government can do without waiting for an AI programme.
No AI-driven national cardiovascular strategy emerged under this name. The prevention agenda continued through conventional channels — NCD screening at Klinik Kesihatan, the National Diabetes Registry, health-promotion campaigns — and risk-stratification tools remain confined to research settings and private wellness apps. The premise, that upstream care cuts long-run cost, resurfaced instead in the health-financing reform debate (see the cost section below), which is arguably where it always belonged.
The strongest concept of the four. Diabetic retinopathy is the leading cause of visual loss among working-age Malaysians, manual fundus-photography screening already runs in government clinics, and deep-learning DR graders have performed at or above human graders in nationwide programmes elsewhere in the region — so the clinical case is proven in principle. What Malaysia lacks is the deployment substrate: standardised digital imaging flowing from clinics into an interoperable record, which Fig. 2 shows arriving only now. Meanwhile MaHTAS assessments of AI-assisted chest X-ray reading (detection-rate gains of ~15.5% among trainees) show the MOH's evaluative machinery is warm to imaging AI. Expect this one to ship eventually — just not as an autonomous system, and not credited to the roadmap.
The most modest use case, and plausibly the one that came closest to its own definition — as university-hospital clinical research in presurgical cleft care. But there is no public trail of national-scale deployment or outcome reporting, which for a “National AI Use Case” is the relevant bar. Verdict: a legitimate research project mislabelled as national infrastructure.
Roadmap promised
- Autonomous AI vaccine-distribution layer (−20% supply, −40% logistics cost)
- National proactive cardiovascular-care strategy
- Autonomous A-eye retinal screening to prevent blindness
- AI-Na predictive cleft-surgery moulding
What 2026 shows
- Vaccine rollout succeeded via MySejahtera — as logistics, not an autonomous AI programme
- Effort went into EMR/CCMS digitisation & a national health record
- AI diagnostics still pending the imaging-data backbone now being built
- Named Projects 7–9 have no clear public delivery trail at national scale
Where it stands now (rev 2026-07)
The Ministry of Health is pursuing its digital agenda through CCMS for clinics, the Dental Information System, and THIS for hospitals, with primary-care digitisation targeted for 2027 and hospitals by 2029.[2] Interoperability work — the Malaysia Patient Summary under the Malaysia Digital Health Certification Network — is the precondition for any national diagnostic-AI deployment, and worth watching more closely than any AI announcement.
National AI coordination, meanwhile, has shifted from MOSTI toward the Ministry of Digital and the National AI Office (established December 2024), with the September 2024 National Guidelines on AI Governance and Ethics setting the framework clinical AI will have to satisfy on privacy, safety and accountability. In practice, a second gatekeeper matters just as much: MaHTAS, the MOH's health technology assessment unit, which evaluates clinical AI for safety, efficacy and cost-effectiveness before public adoption — a role nobody formally designed but which now shapes what reaches government hospitals. For the current adoption picture across public and private care, see the live sector page: AI in healthcare.
The cost problem the roadmap never priced in
Project 7 promised that proactive, AI-assisted care would reduce overall healthcare cost. Five years later, the cost curve went the other way — and it is the one part of this story that reaches every insured household directly.
Whatever else the roadmap got wrong, its economic instinct was sound: Malaysian healthcare costs needed bending downward, and prevention was the lever. What happened instead is that medical cost inflation accelerated — roughly 12.6% in 2023, about 15% in 2024 and 2025, and projected around 16% in 2026, well above global and regional averages.[5][6] Bank Negara has named advancing medical technology as one of the drivers, alongside the rising burden of the very non-communicable diseases Project 7 was meant to get ahead of. There is a certain irony in AI-era medicine showing up first on the cost side of the ledger: new diagnostics and treatment systems arrive in private hospitals as premium equipment, billed at premium rates, paid through insurance claims.
The consequences landed on policyholders. Insurance claims costs outran premiums for years, and the repricing letters that followed prompted Bank Negara's December 2024 interim measures: premium increases staggered over at least three years, keeping annual adjustments under 10% for most policyholders — but only until the end of 2026.[5] Beyond that sit the structural reforms: a transition from fee-for-service billing to diagnosis-related group (DRG) pricing at private hospitals, and a government-designed base medical and health insurance/takaful (MHIT) plan piloting in the second half of 2026 ahead of a 2027 rollout, with repriced policyholders able to switch in without fresh underwriting.[6]
Put plainly: the roadmap imagined AI cutting healthcare costs by 2025. The reality of 2026 is Malaysians re-reading their medical cards to work out whether they can still afford modern care — while the caps protecting them from full repricing expire at year-end.
If this section describes your renewal letter — checking whether an existing plan's annual limits, room-and-board rates and exclusions still match what treatment actually costs in 2026, and how the new base MHIT plan compares as a fallback, is hard to do from policy wording alone.
medicard.my maintains plain-language comparisons of Malaysian medical cards and health insurance plans, including how the BNM repricing measures affect existing policyholders and what to review before switching or topping up cover. General information only — not financial advice.
Why this matters — and what to watch
Healthcare shows the roadmap's recurring gap between named flagship projects and the unglamorous infrastructure that actually determines outcomes. The pandemic proved Malaysia could mobilise a national health-logistics effort at speed; the post-pandemic question is whether that urgency carries into the slow work of records, interoperability and data governance — the foundation every clinical-AI use case in the roadmap quietly assumed. It also shows something subtler: the roadmap's diagnoses (imaging AI, upstream prevention, logistics optimisation) have aged far better than its prescriptions (named autonomous systems on five-year timelines).
For 2026–2030, watch four things: whether the “One Individual, One Record” integration actually completes on schedule; whether hospital digitisation accelerates past its slow 2029 trajectory; whether AI diagnostics (retinal, imaging) move from research pilots into MaHTAS-accredited national screening once the data backbone exists; and whether the DRG and base-MHIT reforms manage to bend the cost curve that Project 7 could not. See also the 11 use cases for how healthcare compares with other sectors.
References & further reading
- “COVID-19 vaccination in Malaysia” (adult two-dose coverage; JKJAV; MySejahtera; daily-dose peak), Wikipedia, accessed 2026. en.wikipedia.org
- “MOH to Integrate 90,000 Medical Imaging Records into MySejahtera” (CCMS 412 clinics; THIS 23 hospitals; digitisation targets), CodeBlue, Feb 2026. codeblue.galencentre.org
- The National AI Office (NAIO), launched 12 Dec 2024 — MyDIGITAL. mydigital.gov.my
- “Fostering Digital Health Innovation in Malaysia” (MySejahtera role; sandboxes), ICMHI 2025 proceedings, ACM. dl.acm.org
- “Bank Negara Caps Medical Insurance Premium Hikes At 10% For Most Policyholders” (interim measures to end-2026; 15% medical inflation; DRG mandate), CodeBlue, Dec 2024. codeblue.galencentre.org
- “New govt medical insurance plan to offer coverage up to age 85” (Base MHIT White Paper; 2026 pilot, 2027 rollout; ~16% projected 2026 medical inflation), The Edge Malaysia, Jan 2026. theedgemalaysia.com
- MaHTAS health technology assessment on AI-assisted chest X-ray analysis (trainee detection-rate improvement), Ministry of Health Malaysia, 2022; summarised in “Malaysia AI Assisted Healthcare Technology”, International Trade Administration, 2025. trade.gov