AI & Vaccines Against Vector-Borne Disease in Asia
- Bridges M&C team
- 7 hours ago
- 6 min read
How AI and vaccines could help Asia anticipate and prepare for future vector-borne disease outbreaks.
Southeast Asia’s risk from vector-borne diseases is becoming harder to predict. Climate change, shifting rainfall patterns and rapid urbanisation are altering the environmental conditions that determine where and when mosquito-borne diseases such as dengue and Zika can spread.
For public health systems, this raises a larger question: how can countries prepare for outbreaks before rising case numbers make the threat obvious?
Increasingly, the answer may lie in combining two very different forms of preparedness: using artificial intelligence (AI) and predictive analytics to identify where outbreaks may emerge, while investing early in preventive tools such as vaccines.
Conventional disease surveillance, including diagnostic testing, genomic sequencing and on-the-ground risk assessment, remains the backbone of how outbreaks are detected and tracked. However, datasets are becoming larger and more complex, spanning climate records, mosquito population counts, human case reports and viral genomics. Newer approaches to infectious disease surveillance and early warning are increasingly examining how multiple sources of data can be integrated and analysed to identify emerging threats early.
The opportunity now is to determine whether these increasingly complex datasets can be used not only to track outbreaks, but to anticipate where they may emerge next.
Can AI Help Predict Where the Next Hotspots Will Emerge?
This is where AI and predictive analytics are increasingly being explored as a complement to conventional surveillance. In theory, machine learning models can be trained to identify subtle correlations across disparate datasets. Rainfall patterns, temperature anomalies, land use, population density, historical case data and even satellite imagery can be studied to flag areas at elevated risk of an emerging outbreak before conventional surveillance would detect a rise in cases.
The appeal is straightforward: if a model can reliably highlight a hotspot forming weeks or months in advance, health authorities gain a valuable window to redirect vector control resources, ramp up diagnostic testing, or issue public health advisories before transmission accelerates.
Dr Devanathan Raghunathan, CEO and co-founder of Keeping Labor Safe, explains that no single data point signals an outbreak on its own. "Rainfall, rising temperatures, mosquito counts or even a few cases of fever do not, by themselves, signal an outbreak." It is when several of these indicators move together, that a pattern becomes meaningful.
He compares it to how meteorologists forecast a hurricane, "We do not wait until a hurricane reaches the coastline before concluding that conditions are dangerous." The same logic, he argues, should apply to disease surveillance by shifting the question from where the outbreak is located to where conditions for one are developing.
The bigger obstacle in Southeast Asia, Dr Devanathan notes, usually is not a lack of data but where the data is collected. Weather, health, environmental and transport data are scattered across different agencies, recorded on different timelines and mapped to different geographic boundaries. "The challenge is therefore not simply 'better data’, but creating an interoperable regional information ecosystem" that can combine these signals fast enough to act on.
However, predictive models are only as good as the data feeding them, and in many parts of Southeast Asia, environmental and epidemiological datasets remain fragmented, inconsistently collected, or simply unavailable at the resolution needed for more granular hotspot prediction. There is also the challenge of turning a forecast into action, since identifying where an outbreak may occur is only useful if authorities are able to respond in time.
From Prediction to Prevention
Identifying a likely hotspot is only half the battle. Health systems also need the operational capacity and resources to act on that intelligence before an outbreak takes hold.
This might mean deploying additional vector control teams, increasing diagnostic capacity in a specific district, or ensuring that medical countermeasures are available where they are likely to be needed. It also requires coordination across environmental, healthcare and civil defence agencies, so that they can respond collectively to emerging risks rather than acting independently once cases begin to rise.
According to Dr Devanathan, acting on these forecasts presents its own challenges. Predictions come with uncertainty. For that reason, he argues predictive tools should not function as an “automated trigger" but as “a decision-support layer” that helps health authorities prioritise where to test, communicate and intervene earlier. As he puts it, the field is shifting "from a technical bottleneck to an execution bottleneck". The harder question is not whether useful forecasts can be generated, but whether public health systems can integrate them into existing workflows quickly enough to act.
Increasingly, experts argue that the most robust form of prevention includes tools that exist independently of the timeline of any single outbreak, most notably, vaccines. Unlike a predictive model, a vaccine cannot be developed and deployed on short notice. It requires years of research, clinical trials and regulatory review, which means the decision to invest in a vaccine programme has to be made long before an outbreak reaches crisis proportions.
Hyun Soo Kim, Chief Executive Officer of Sun Biotech Singapore, points to a structural problem that makes this even harder in practice: once an epidemic recedes, so does the data needed to prove a vaccine works. As Kim puts it, "the decline in Zika cases since the 2016 epidemic has made efficacy trials difficult to recruit." Developing a Zika vaccine presents an additional challenge: pregnant women are among those at greatest risk from the virus, yet their inclusion in vaccine trials raises important ethical considerations.
The lesson Kim draws from the COVID-19 pandemic is that "vaccine development needs government support in terms of financial backing and future purchase guarantees". Without government commitment upfront, manufacturers may be reluctant to invest in vaccine development before there is clear and immediate demand.
Vaccines Cannot Be Developed Overnight
The long lead time for vaccine development illustrates why infectious disease preparedness must begin well before an outbreak occurs. For emerging or underserved diseases that have yet to attract significant investment or global attention needed to drive vaccine development, the timeline can stretch even further without investment ahead of demonstrated need.
Sun Biotech focuses on developing and manufacturing vaccines for infectious diseases that disproportionately affect Asia but have historically attracted less R&D investment from larger pharmaceutical companies. Working with governments and local industry partners across the region, the company’s areas of focus include national immunisation priorities such as Zika and Japanese encephalitis (JE).
Kim's experience developing JE vaccines with the World Health Organization and the Bill & Melinda Gates Foundation has shaped his preference for inactivated vaccine platforms, which offer a strong safety profile and are well suited to tropical, resource-limited settings compared with faster-to-design but more challenging-to-deploy alternatives such as mRNA or DNA-plasmid vaccines.
Zika offers a useful lens through which to examine why this approach matters, and how predictive analytics and vaccine development could together strengthen preparedness for future outbreaks.
Zika: Preparing Today for Tomorrow's Threat
More than a decade after Zika emerged as a global threat linked to severe birth defects, the virus remains a public health concern.
The World Health Organization now counts evidence of current or previous local Zika transmission in 97 countries and territories worldwide, and while global case numbers have declined sharply since 2017, transmission continues at low levels across the Americas and other endemic regions, with new outbreaks still surfacing periodically, including in Asia in 2024.
Singapore illustrates how easily that low-level circulation can resurface. Thirteen Zika cases were reported in all of 2024 and seven cases were logged by mid-June 2025, prompting the National Environment Agency (NEA) and Communicable Diseases Agency (CDA) to launch wastewater and mosquito surveillance around the affected Woodlands neighbourhood, where testing revealed persistent Zika virus signals in the area.
Dr Devanathan places Zika within the wider vector-borne disease landscape, noting that the same shifts in temperature, rainfall and urbanisation also affect dengue and chikungunya. Effective predictive systems will therefore need to track multiple pathogens rather than Zika in isolation. He sees the potential for such systems to eventually provide personalised, location-specific alerts to travellers, clinicians and residents before a rise in cases becomes apparent.
The goal, he says, is not simply to “predict Zika” but to build “a forward-looking public-health system” that gives governments and individuals time to act before a threat becomes an outbreak. Vaccine development, pursued not as a reaction to crisis but as a long-term investment, offers a further layer of protection that can be implemented well before the next major outbreak emerges.
For Zika, Sun Biotech’s vaccine candidate remains at the R&D stage, alongside numerous other candidates being developed using different platforms. For Kim, this is precisely why vaccine development cannot wait for the next major outbreak. The company’s broader aim is to deliver “sustainably affordable vaccines” to vulnerable regions through partnerships with governments and international organisations, building protection before the next threat emerges.
Ultimately, the story of vector-borne disease preparedness in Southeast Asia is not one of any single solution, but of complementary systems working in concert: analytics that may help anticipate where the next threat is heading, and preventive tools such as vaccines that take years to build and must therefore be developed long before they are urgently needed.
As climate change continues to reshape the region's disease landscape, this kind of multi-layered preparedness will likely determine how well Southeast Asia is able to respond not just to Zika, but to whatever vector-borne threat emerges next.
Comments