Speaking in Seremban on 27 August, the Health Minister gave two numbers and one answer. The numbers: more than 61,000 dengue cases to epidemiological week 33, up 60.8% on the same period last year, and 58 deaths. The answer, to whether the dengue vaccine will enter the National Immunisation Programme: not yet.

The shape matters as much as the total. At week 13, at the end of March, regional surveillance recorded 15,931 cases and 14 deaths for Malaysia. Twenty weeks later, the Minister gave more than 61,000 and 58. The two figures come from different channels and may not be strictly comparable—but the direction is unmistakable, and ours is going one way.

His words were careful. One vaccine has been approved and registered and is being used in private facilities—conditionally. The Drug Control Authority gave conditional approval in February 2024—MSIDC dates the original approval to 9 February and DPAM to 8 February; the one-day discrepancy is unexplained. At its 416th meeting on 8 January 2026, the DCA did two things: it renewed the conditional registration for a further two years, and separately lifted a different condition, the Act2Care registry requirement, effective from 9 February 2026 in both cases. A cost-benefit analysis is under way. To make it part of the NIP is not easy. Give us some time.

He is not wrong, and I want to be fair to him before I am difficult about it. But there is a sentence buried in that exchange which deserves more attention than it will get: a vaccine that Malaysians can buy is one that Malaysians cannot be given. The registration decision has been made. What has not been made is a decision about who gets it.

I should declare an interest before going further, because it is not a small one. From January 2014 to March 2016 I was Senior Director for Dengue Medical Affairs, Asia-Pacific, at Sanofi Pasteur, the vaccines division of Sanofi, through the period in which Dengvaxia—the first dengue vaccine—was being brought to registration. I left in March 2016.

I have no current relationship with any manufacturer of a dengue vaccine, and nothing to gain from what Malaysia decides. I should note that in October 2021, through my own consultancy, I did a short piece of scientific input work for Merck, which was also developing a dengue vaccine. It was brief and it ended there. But I am not a neutral observer of this field. I have sat on the manufacturer's side of exactly the conversation the Ministry is now having, for the product that came before this one.

So I want to set out what is actually being weighed, and what physicians might reasonably want to know before accepting the answer.

What the vaccine is

The vaccine is TAK-003, sold under the trade name Qdenga, made by Takeda. It is a live attenuated tetravalent vaccine built on a dengue-2 backbone carrying the surface proteins of the other three serotypes. Two doses, three months apart, from four years of age.

The important thing about it is what it is not. Its predecessor, Dengvaxia, could only be given to people with laboratory-confirmed prior dengue infection, because vaccinating the never-infected raised their risk of severe disease. That requirement is what limits it as a mass public health tool: screening every child before vaccination is impractical in most of the places where dengue is worst.

Qdenga can be given regardless of serostatus, with no pre-vaccination screening. That is the whole reason it is interesting. In September 2023 WHO SAGE recommended it for children aged 6 to 16 in settings with high dengue transmission, and it has since been prequalified by WHO from six years of age. Brazil began a national immunisation programme with it in 2024.

What it does, and what it does not

The pivotal TIDES trial followed 20,099 children and adolescents aged 4 to 16 at 26 centres across eight endemic countries in Latin America and Asia. Malaysia was not among them. Takeda chose a multi-country design deliberately, to capture variation in circulating serotypes and prior flavivirus exposure. That was sound trial design. It also means these efficacy figures were measured in populations selected for their heterogeneity, and ours was not among them. Efficacy against virologically confirmed dengue was 80.2% in the first year after the second dose, settling to 61.2% against disease and 84.1% against dengue requiring hospitalisation at four and a half years. Those are real numbers and worth having. An 84% reduction in hospitalised dengue is not nothing—a vaccine doing that much work, in a country recording 61,000 cases in eight months, is worth taking seriously on its own terms.

But efficacy varies by serotype and by prior infection, and this is where a Malaysian reader should pay attention. WHO's position paper, as summarised in DPAM's position paper on dengue vaccination. Updated twice since first publication, most recently in July 2026, it is the source for several of the Malaysian figures below. It records that more DENV-3 infections were reported among baseline seronegative children in the vaccine group, and an observed increase in severe dengue and dengue haemorrhagic fever among seronegative recipients—all of it linked to DENV-3. Neither difference reached statistical significance. But an increased risk of hospitalised or severe DENV-3 disease in the never-infected cannot be ruled out.

What "approved" means, and where it has not been

Forty-one countries have authorised Qdenga, and some 18.6 million doses have been distributed. It is worth knowing where it has not.

Takeda withdrew its application to the United States Food and Drug Administration in July 2023. In Singapore, the Health Minister told Parliament that Qdenga is not approved for use there, its application having been withdrawn "following a review by the Health Sciences Authority on its submitted scientific data".

I want to be careful with that. A withdrawn application is not a refusal, and companies may withdraw applications for commercial as well as scientific reasons. Takeda's account of the American decision is that the FDA sought data not captured within the trial protocol—a protocol the agency had previously reviewed and accepted—and treated as missing what the protocol never required. That is a reasonable position and worth stating. But two regulators of standing have reviewed applications for the product without licensing it, while ours licensed it conditionally. That is worth knowing before deciding what a cost-benefit analysis should assume.

The serotype question

In June the Health Minister said that DENV-3 is now dominant in the Malaysian infection cycle, and attributed the surge to two things: the expected four-to-five-year cycle, and this shift in circulating serotype. The Ministry has since repeated the attribution. Set that against the vaccine's recognised area of uncertainty in seronegative recipients exposed to DENV-3, and the two questions now overlap.

I want to be careful, because this is where a column like this overstates. I am not the first to notice it, and the people who noticed first were more measured than I would have been.

Dengue Prevention Advocacy Malaysia, or DPAM, is a coalition led by several Malaysian paediatric, infectious-disease and public-health professional bodies, with the MMA among those in support. Its co-chairman, Professor Dr Zamberi Sekawi, was precise about what was actually at stake:

He told The Star: the concern is not that DENV-3 is more infectious, but that a major serotype shift increases the number of people who are susceptible, particularly where a population has been more exposed to other serotypes. That raises case numbers, and may increase the risk of severe disease in those reinfected with a different serotype. His conclusion: the rise of DENV-3 should be taken seriously, but not with panic. Professor Dr Sharifa Ezat Wan Puteh of UKM made the same point from the other end—those never exposed to DENV-3 are at risk from it even if they have had dengue before.

A systematic review published in April adds the uncomfortable part: across twenty Malaysian studies, DENV-3 was the serotype linked to higher disease severity, and mixed-serotype infection significantly raised the risk of severe outcomes.

So the serotype where the vaccine's evidence in the never-infected is thinnest is the serotype now circulating, and the one our own literature associates with severe disease. That is not an argument against the vaccine. It is an argument for knowing what the cost-effectiveness model assumed about efficacy—pooled trial figures, or figures adjusted for the serotype we actually have. That choice could materially change the answer.

How much this matters depends on how many children in the vaccine target age group have not yet met dengue—and we know roughly, though not as precisely as the DENV-3 argument requires. A secondary analysis of Malaysian baseline data from two Dengvaxia phase 3 trial sites in Kuala Lumpur and Penang, enrolling children between 2011 and 2013, found overall dengue seropositivity of 32--34% in the 2-to-8 age band and 58--67% in the 9-to-16 band. Among children aged 2--8, roughly two-thirds were seronegative; among those aged 9--16, roughly one-third to two-fifths were. This demonstrates that a substantial seronegative population exists within and around the vaccine target ages. The data come from a Sanofi Pasteur paper, which is worth noting alongside my own disclosure above. What neither this nor any other published Malaysian dataset gives us is a serotype-specific breakdown: how many of those seronegative children are DENV-3 naive specifically. That is the more precise figure the serotype argument requires, and it is not in the literature.

On safety, the record so far is reassuring and I should say so plainly: no evidence of disease enhancement in vaccine recipients, and no important safety risks identified to date.

What a cost-benefit analysis actually asks

The phrase is doing a lot of work in the Minister's answer, and most readers will take it to mean "we are checking whether it is affordable." It means something more specific, and the specifics are where the decision is really made.

1. What does a case cost us now? Not just the hospital bed. The lost work, the lost school, the fogging, the enforcement, the ward that fills in a bad year. A vaccine looks expensive against a drug budget and cheap against an epidemic.

2. What efficacy figure went in? Pooled trial data, or something adjusted for our serotype distribution and our seroprevalence? The difference is not academic; it could materially alter the result.

3. Who is vaccinated, and when? WHO's recommendation is 6 to 16 in high-transmission areas. A national programme, a state programme in Selangor and Kuala Lumpur where the cases are, and a school-based catch-up are three quite different propositions with three quite different prices.

4. What threshold is being applied? Every country that does this work has an implicit ceiling on what it will pay per healthy year of life gained. Malaysian academics reference the WHO convention of one to three times GDP per capita—and a published analysis of the earlier vaccine found an incremental cost-effectiveness ratio of USD 122,000 per quality-adjusted life year, described by one of them as exuberant. But the threshold the Ministry applies is not published. It should be. A decision made against an unstated threshold cannot meaningfully be scrutinised.

5. And what price was negotiated? The list price is not the price. Two doses are reported at RM400 to RM500 in Malaysian private clinics. National procurement is not private procurement—and the Malaysian model that produced the RM4 billion figure assumed USD 25 a dose, less than half the private rate. So somebody has already costed this at a negotiated price. Whether the Ministry used the same figure, nobody outside it can say.

The modelling exists. It is just not the government's.

My complaint, drafting this, was that nobody had done the sums. That was wrong too. Dr Amirah Azzeri of Universiti Sains Islam Malaysia presented a model calibrated to local incidence, finding that routine vaccination from age seven with catch-up cohorts could prevent 34 to 42% of symptomatic cases over twenty years, at a net societal saving of up to USD 1.06 billion over thirty years.

Takeda has its own Malaysian cost-effectiveness analysis, presented at ISPOR Europe in 2024, calibrated to confirmed age-specific incidence from 2014 to 2023. And a decade ago, in the Dengvaxia era, a value-based pricing assessment using dynamic transmission modelling reached the conclusion that has stayed with me since: dengue vaccination is a potentially good investment if the purchaser can negotiate a price at or below the cost-effectiveness threshold price. I had no part in that paper and did not commission it. But I was working on that vaccine in this region while the data behind it was being generated, and I have not seen the point put better.

So we have modelling from a Malaysian university, modelling from the manufacturer, modelling from the previous manufacturer's era, and a professional position paper. What we do not have is the Ministry's own analysis, its assumptions, or its threshold.

How our neighbour does it

Singapore faced the same decision with the earlier vaccine, and answered it in public. When its Expert Committee on Immunisation advised against a national Dengvaxia programme, the Ministry published the reasoning in full: the vaccine works best in those already infected; Singapore's seroprevalence is about 10% below eighteen and 20 to 40% between eighteen and forty; below the 50% threshold WHO sets, it would be less effective there. So they would not roll it out.

Their regulator did the same. HSA published its assessment—efficacy by serotype, the age restriction and the reasons for it, and the composition of the expert panels consulted. Whether or not one agrees with the conclusion, one can see how it was reached.

That is my complaint, put positively. It is not that Malaysia has decided wrongly. It is that Malaysia has not shown its working, and a country an hour away demonstrates that showing it is perfectly possible.

And who decides?

A 2017 country brief on Malaysian immunisation financing, commissioned by ThinkWell, confirms that the National Committee on Immunization Practices does routinely request evidence from MaHTAS—the technical machinery exists and is used. But the same brief is explicit that budget considerations and political will are the key determinants of whether a new vaccine is introduced. What the public cannot see is how those considerations were balanced against the evidence in this case.

The NCIP itself is described the same way—it prioritises budget considerations and political support when considering adoption. That is a committee doing what committees are supposed to do, weighing cost against benefit. What is missing is not the process. It is any public account of where that weighing lands, and why.

One line in the brief is worth sitting with. Written in 2017: a dengue vaccine is of interest for the Ministry of Health, and Malaysia is hosting dengue vaccine clinical trials. That is eight years before this week's answer. Whatever "give us some time" means, the clock did not start with Qdenga's conditional registration in 2024. It has been running for the better part of a decade.

One disclosure the brief itself makes, and I should repeat it: it was produced with funding from Merck Sharp & Dohme—the same company for whom I did short consultancy work in 2021, on the same disease. Neither of us is a neutral narrator here. Read the brief, and this paragraph, accordingly.

Meanwhile, it is on sale

Here is the part that should trouble us.

The vaccine is registered and available in private facilities. So a Malaysian family that can find eight hundred to a thousand ringgit can vaccinate two children this month. A family that cannot, waits for the analysis.

Dengue does not distribute itself evenly. It concentrates in dense housing, in areas with unreliable water supply where containers are stored, in construction accommodation, in the places where fogging is announced and not always done. The people at highest risk are, with dreary predictability, the people least able to buy their way out.

A registered vaccine sitting outside the national programme therefore does something specific: for now it leaves a potentially useful public-health intervention functioning as a consumer product, during an epidemic. That may be unavoidable while the analysis runs. It should at least be named.

It is worth noting that the Malaysian profession has not been silent on this. Dengue Prevention Advocacy Malaysia has published a position paper on dengue vaccination—a coalition of Malaysian paediatric, infectious disease and public health bodies. It sets out the registration conditions imposed on Takeda—participation in the DEN-401 effectiveness study in Southeast Asia, a risk management plan with a Malaysia-specific annex, and a registry of every recipient. It also discloses its own funding: an unconditional educational grant from Takeda Malaysia. That disclosure is to its credit, and I note it in the same spirit as my own.

And the rest of it

None of this displaces what actually reduces dengue, and the Minister was right to list it: the operations room activated, personnel deployed to hotspots in Kuala Lumpur and Selangor, Wolbachia expanded, and the reminder that this is not the Ministry's job alone but requires local authorities, agencies and residents to stop breeding the mosquito.

But vector control has been our primary strategy for fifty years, and it is where more than nine tenths of the dengue vector control budget goes—largely on fogging, which the Auditor-General found no robust evidence for. This year we passed a full year's caseload by July. It is not working well enough on its own, and saying so is not defeatism.

Wolbachia, which deserves more credit than it gets

The Wolbachia Malaysia project is the most interesting thing our vector control has done in fifty years, and most Malaysians have never heard of it. Wolbachia is a bacterium found naturally in many insects but not in Aedes aegypti mosquitoes. When introduced into Aedes aegypti, it reduces their ability to transmit dengue and spreads through local populations over time, suppressing transmission without chemical intervention.

The Institute for Medical Research established the wAlbB strain in six sites in greater Kuala Lumpur from 201931446-0), including the Mentari Court and Section 7 flats; releases now run at dengue hotspots in the Klang Valley and Penang. A later analysis across 20 release sites against 76 controls in high-rise residential areas found an average reduction in dengue of 62.4%, confidence intervals 50 to 71%.

Effectiveness has proved genuinely heterogeneous—the published Malaysian estimates vary widely, and the Auditor-General was less impressed, concluding in 2022 that the project had yet to demonstrate comprehensive results—while also finding no robust scientific evidence that fogging, our default response for half a century, works at all.

That last point is the unglamorous truth of dengue control. No vaccine at 61% efficacy substitutes for not having standing water. The iDengue hotspot map exists, it is public, it works, and it is under-used. If a locality near you is on it, that is more immediately actionable than anything in this article.

And here is the circle

A cost-effectiveness analysis of Wolbachia is under way at the Institute for Health Systems Research. But the Deputy Minister was explicit about its limits: it compares Wolbachia only against existing dengue control measures, and cannot compare it against vaccination—because a national immunisation programme using dengue vaccines has not been implemented here. He gave March 2026 as the date that analysis was expected. It is now the end of August, and I can find nothing published.

But implementation is not a prerequisite for modelling. That is what modelling is for. Thailand published a dynamic transmission model of combined vector control and dengue vaccination. Indonesia published one on vaccination integrated with Wolbachia and health education. The question Malaysia says it cannot answer has been answered twice next door.

Read that alongside the Minister's answer this week. The vaccine is not adopted, pending a cost-benefit analysis. The Wolbachia analysis cannot weigh vaccination, because vaccination has not been adopted. So the two most promising things we have are each being assessed in isolation from the other, and the country will end up with two numbers and no comparison.

That is nobody's fault. It is what happens when decisions are sequenced rather than compared. But somebody should say so, because the interesting question was never whether the vaccine is worth it. It was what the best use of the next ringgit spent on dengue would be—and nothing currently under way is designed to answer that.

For scale: Wolbachia runs at roughly RM2.4 million a year, with expansion to ten new localities costed at RM4 million annually. A national vaccination programme would be an order of magnitude more. Both numbers are knowable. Neither is being set against the other.

Who is dying, and who would be vaccinated

There is a mismatch here that nobody has put plainly, and it is not an objection so much as a scheduling problem.

Severe dengue in Malaysia concentrates in adults with chronic illness. An emergency physician described it to the New Straits Times in August: the patients affected most are often older adults living with obesity and diabetes, and when they come in late, management becomes much more difficult. The supporting evidence is consistent—hospitalised adults with pre-existing chronic conditions face roughly twice the odds of severe dengue, a pattern documented across endemic settings in this region.

Now look at who the vaccine is for. WHO recommends it for children aged 6 to 16 in high-transmission settings. The pivotal trial enrolled 4 to 16 year olds. The Malaysian model that produced the RM4 billion figure runs routine vaccination from age seven with catch-up cohorts to thirteen. For adults, the evidence is immunobridging—antibody and T-cell responses in adults compared against the trial populations. That is a legitimate regulatory route, and it is how the label extends past sixteen. It is not efficacy data.

The adults in whom severe outcomes are concentrated are therefore the population for whom direct vaccine-efficacy evidence is thinnest and for whom there is no WHO programmatic recommendation.

That is the mismatch. Here is what follows from it.

Which is an argument for starting, not for waiting

That is not an argument for an adult vaccination programme; the efficacy evidence is not there. It is an argument for recognising that today's mortality problem and a childhood vaccination strategy operate on different timescales.

Vaccinate at seven and the cohort carries protection forward. Suppressing transmission in children protects adults indirectly, and the Malaysian model accounts for it. That is the correct foundation and it is what WHO recommends. A childhood programme builds protection forward; it cannot directly solve the risk carried by older adults today.

There was, however, one way to learn more from adults already receiving the vaccine. A condition of registration required every Malaysian recipient to be enrolled in a national registry, Act2Care, for periodic surveillance. The Drug Control Authority removed it on 8 January 2026 and the registry closed on 9 February 2026—the same day, as it happens, that the vaccine's renewed two-year registration took effect, at the same 416th meeting described above, whose minutes are not public. There may be good reasons; enrolment conditions are often lifted once early safety data reassure, and DEN-401 continues. But the vaccine is still on private sale, adults are still being vaccinated outside any programme, while Act2Care is no longer systematically recording those recipients. We had a mechanism for systematically following Malaysian recipients, including adults vaccinated outside a national programme, and it was closed while real-world use continued. Whether those data could have answered meaningful effectiveness questions depends on what the registry collected; the decision to stop collecting them deserves explanation.

That belongs in the same conversation as the price.

In short

A dengue vaccine exists, is registered here, and is being sold here. It is imperfect: 61% against disease, 84% against hospitalisation, and an important area of uncertainty precisely where our current serotype sits. It is not a solution. It is a tool, with a price and a set of trade-offs.

The Ministry is right to analyse before committing. But a cost-benefit analysis is a set of assumptions wearing a number, and the assumptions are not published. What efficacy was assumed. What a case is costed at. What threshold applies. What price was actually offered.

Publish the assumptions behind those choices and the decision becomes arguable, which is what a public health decision ought to be. Withhold them and we are asked to accept a conclusion on trust, during an epidemic, while the vaccine is available to anyone who can pay for it.

Give us some time, the Minister said. Fair enough. But tell us what you are counting.