When the nudge runs out of road: what co-design could offer Australia’s stalling immunisation rates

In April 2026, the National Centre for Immunisation Research and Surveillance (NCIRS) released data confirming what public health observers had been quietly bracing for: childhood vaccination coverage in Australia has now fallen for a fifth consecutive yea r[1]. Coverage at 12 months of age sits at 90.5 per cent. At 24 months, 88.4 per cent. At 60 months, 92.5 per cent. Each of these figures is down meaningfully on the 2020 baseline, and each sits below the 95 per cent threshold required to maintain herd immunity against measles and other highly infectious diseases [2].

The nudge architecture we already have

Australia is, by international standards, an unusually instrumented vaccination system. We have a comprehensive national register, a tightly scheduled program, a vast and free vaccination offering and — since 2016 — a set of behavioural and financial levers known colloquially as “No Jab, No Pay” (NJNP) in relation to welfare payments.  At the state-level “No Jab, No Play” rules also govern access to subsidised childcare enrolments.

These are nudges in the strict policy sense: they restructure the choice architecture rather than mandating vaccination outright. The financial stakes are significant. By 2019, families forgoing payments under NJNP could be missing up to AUD $26,000 per child annually, with roughly half of Australian families eligible for some component of the affected payments [7].

The evaluation evidence is now reasonably mature, and it tells a story. A Lancet Regional Health before-and-after study comparing pre-NJNP and post-NJNP cohorts found that approximately 49,510 more children were fully vaccinated than would have been expected without the policy.  The largest behavioural shifts occurred amongst families with previously registered objections [8]. A 2023 scoping review across ten evaluations concluded that NJNP and No Jab No Play produced “small but gradual and significant increases” in coverage and catch-up vaccination [9].

Smaller-scale behavioural nudges show similar patterns. The Australian EPIC study — a randomised controlled trial of SMS-based nudges to pregnant women — found that three reminder messages incorporating social norming content, safety information, video testimonials and a vaccination pledge produced uptake increases of 1.6 per cent for COVID-19 and 1.7 per cent for influenza [10]. Modest in absolute terms, but at SMS-level cost, plausibly cost-effective at population scale.

The behavioural insights literature would frame this neatly. Nudges work best on what Milkman and colleagues call the intention–action gap: people who already mean to vaccinate but face friction — forgetfulness, scheduling, mild procrastination, or the small effortful cost of confirming what they already believe [11]. For this group, a well-timed text, a default appointment, a simplified booking flow, or a financial penalty that makes the default action salient will reliably move behaviour.

Where the nudge runs out

The trouble is that the population the current architecture struggles to reach is no longer the intention–action gap group. It is the group whose intentions themselves have shifted, or whose intentions sit inside a different epistemic framework altogether.

The NCIRS authors of the original NJNP evaluation were candid about this limitation. Reviewing catch-up MMR1 activity among children aged 5 to under 7, they noted that financial sanctions had “little impact on those who reject vaccination,” and that further expansion of the policy was unlikely to shift this group meaningfully [12]. Even within the broader post-NJNP cohort, the proportion of children with registered objections actually rose — from 1.1 per cent to 1.5 per cent — even as catch-up activity increased among the previously hesitant-but-not-objecting [13].

This is the predictable shape of a nudge plateau. The mechanism reaches the people whose behaviour was loosely coupled to the choice architecture. It does not reach the people whose behaviour is tightly coupled to identity, trust, lived experience, social network, or worldview. For that group, an SMS reminder reads as patronising. A financial penalty reads as coercion. A celebrity endorsement reads as part of the apparatus they already distrust. The same intervention that gently moves an “intender” can actively entrench a refuser.

The post-pandemic shift in vaccine attitudes has, if anything, sharpened this. Research from the Immunization Foundation of Australia and others suggests trust is now built less through institutional credentials and more through identity-based affinity: parents are more persuaded by someone like them — a homeschooling mother, a fellow parent at the school gate — than by a generic clinician [14]. This is not a failure of public reasoning; it is how trust has always operated under conditions of uncertainty.

What co-design could offer

If the nudge architecture has reached the limits of who it can move, the policy question becomes: what could engage the population the nudge cannot? This is exactly the terrain where codesign is built to work — and it has been substantially under-used in Australian immunisation policy to date.

Codesign, properly understood, is not a consultation exercise dressed in better language. It is a structured process in which the people most affected by a system contribute to shaping the system itself, from problem definition through to solution design and evaluation. Under the Co.Design4All framework, that means treating those who currently sit outside the immunisation system not as a target population to be persuaded, but as co-authors of an arrangement that might actually work for them and their children.

Three features distinguish what codesign with vaccine-hesitant or vaccine-opposed parents could look like — and what it would need to avoid.

First, the question on the table needs to genuinely move. A codesign process whose hidden brief is “how do we get these people to vaccinate” is not codesign — it is a focus group with extra steps. A real codesign question might be: What would a child health system look like that you would trust enough to engage with at every stage, from antenatal through to adolescence? Vaccination is part of that system, but it is not the only part. Parents who currently decline vaccination often also decline newborn vitamin K, certain antenatal screenings, or developmental checks — not in isolation, but as expressions of a broader stance toward medical institutions. The codesign frame has to be wide enough to take that stance seriously rather than carving off vaccination as the one negotiable item.

Second, the participants need to include the actual refusers, not proxies. This is where most existing community engagement on vaccination quietly compromises. It is far easier to convene “vaccine-hesitant-but-reachable” parents than parents with registered objections. The Vaccine Champions program developed in Victoria during COVID-19 illustrated the value of peer-to-peer engagement, training community, faith and industry leaders to advocate within their own networks [15]. That model is important — but it primarily moves the population in the middle, not out at the edge. A codesign process aimed at the edge has to bring in parents whose worldview the public health system has previously treated as outside the conversation. That includes parents drawn from homeschooling communities, alternative health networks, some faith-based groups, and communities with historic and well-grounded reasons to distrust medical institutions — including, but not only, Aboriginal and Torres Strait Islander families whose under-vaccination at the 60-month milestone remains below national targets [16].

Third, the process has to be willing to surface and act on findings that are uncomfortable for public health agencies. Some of what emerges will be addressable: poor experiences at specific clinics, dismissive interactions with GPs, fragmented information environments, distrust born of past harm. Some will be harder: requests for changes to the schedule, separate appointments rather than combined visits, longer informed-consent conversations, alternative information channels that do not route through government. If a codesign process surfaces these and the system is unwilling to act on any of them, the process will — correctly — be read as performative, and trust will fall further. The international evidence on community-driven vaccination strategies suggests this is not naive. Studies of equity-focused, community-led vaccine programs have shown meaningful trust and participation gains where the community is genuinely shaping the program rather than receiving it [17]. Australian research on culturally and linguistically diverse communities in Melbourne’s north reached similar conclusions: trusted sources, culturally sensitive information, and supported pathways together moved uptake in ways generic communications could not [18].

The case for sequencing, not substitution

None of this is an argument against the nudge architecture. The evidence that NJNP and SMS-based interventions move the intention–action group is solid, and that group is large. Removing those mechanisms would almost certainly lower coverage, not raise it.

The argument is for sequencing. The nudge layer is doing the work it can do, and the diminishing marginal returns are visible in the year-on-year coverage declines despite all of these policies remaining in force. The next percentage points of coverage will not come from the same toolkit. They will come from designing a child health system where people who currently sit outside it can find their way into it — which is a different problem, requiring a different method.

The National Immunisation Strategy 2025–2030 sets the policy frame for the coming period. Whether it produces a different trajectory will depend, in significant part, on whether the implementation moves beyond communication campaigns aimed at the hesitant middle and engages — genuinely, structurally — with the parents the current system has stopped reaching. Codesign is not a panacea. But it is the method best suited to the part of the problem that nudges alone cannot solve.

Article References

1. National Centre for Immunisation Research and Surveillance (NCIRS), Annual Immunisation Coverage Report 2025 — Summary, April 2026. ncirs.org.au.
2. Australian Government Department of Health, Disability and Ageing, Childhood immunisation coverage, accessed May 2026. health.gov.au.
3. National Centre for Immunisation Research and Surveillance (NCIRS), Annual Immunisation Coverage Report 2025 — Summary, April 2026. ncirs.org.au.
4. National Centre for Immunisation Research and Surveillance (NCIRS), Annual Immunisation Coverage Report 2025 — Summary, April 2026. ncirs.org.au.
5. See for example Timeliness and Equity: An Analysis of Measles Herd Immunity in a Regional Area of Australia, 2025. pmc.ncbi.nlm.nih.gov.
6. National Centre for Immunisation Research and Surveillance (NCIRS), Annual Immunisation Coverage Report 2025 — Summary, April 2026. ncirs.org.au.
7. Hull BP et al., “Impact of Australia’s No Jab, No Pay policy on vaccination uptake — a before-after study in two national birth cohorts,” Lancet Regional Health — Western Pacific, 2024. thelancet.com.
8. Hull BP et al., “Impact of Australia’s No Jab, No Pay policy on vaccination uptake — a before-after study in two national birth cohorts,” Lancet Regional Health — Western Pacific, 2024. thelancet.com.
9. Burns S et al., “The Impact of the No Jab No Play and No Jab No Pay Legislation in Australia: A Scoping Review,” International Journal of Environmental Research and Public Health, 2023. mdpi.com.
10. Marshall H et al., “Randomised controlled trials of behavioural nudges delivered through text messages to increase influenza and COVID-19 vaccine uptake among pregnant women (EPIC study) in Australia,” Vaccine, 2025. pubmed.ncbi.nlm.nih.gov.
11. Dai H et al., “Behavioural nudges increase COVID-19 vaccinations,” Nature, 2021. pmc.ncbi.nlm.nih.gov.
12. NCIRS, “No Jab No Pay raises catch-up vaccination rates,” commentary on Beard et al. analysis. ncirs.org.au.
13. Hull BP et al., “Impact of Australia’s No Jab, No Pay policy on vaccination uptake — a before-after study in two national birth cohorts,” Lancet Regional Health — Western Pacific, 2024. thelancet.com.
14. Hughes C, comments in Beyond the Comment Section: Rebuilding Trust in Vaccines Through Conversation, Applied Clinical Trials, 2026.
15. Kaufman J et al., “Vaccine Champions Training Program: Empowering Community Leaders to Advocate for COVID-19 Vaccines,” Vaccines, 2022. mdpi.com.
16. NCIRS, Annual Immunisation Coverage Report 2025 — Summary, noting Aboriginal and Torres Strait Islander coverage at 60 months remains below the 95 per cent national target.
17. See discussion in From hesitancy to confidence: Shaping the future of vaccine messaging, 2025. pmc.ncbi.nlm.nih.gov.
18. Qualitative study of CALD communities in Melbourne’s outer north, 2023. ncbi.nlm.nih.gov.