OFFICIAL INFORMATION ACT REQUESTS

Ministry for Primary Industries

REQUEST: AUGUST 20, 2026 Request No. OIA26-0675
RESPONSE: SEPTEMBER 8, 2026 Request No. OIA26-0675
NEW REQUEST: SEPTEMBER 17, 2026

Department of Conservation

REQUEST: AUGUST 18, 2026 Request No. OIAD-9589
RESPONSE: SEPTEMBER 14, 2026 Request No. OIAD-9589
NEW REQUEST: SEPTEMBER 17, 2026 OIAD-9771

What is happening?

DOC has begun a programme to vaccinate approximately 300 core breeding birds from some of New Zealand's most threatened native species against H5N1 avian influenza. These include kākāpō, takahē, tūturuatu/shore plover, kakī/black stilt and kākāriki karaka/orange-fronted parakeet. This is a significant intervention because these are small, threatened populations and some of the birds are extremely important to the survival of their species.

Why are we asking questions?

DOC publicly described its 2024–2025 vaccination trial as showing that vaccination was 'safe and effective' and would help protect these birds from H5N1. But the trial did not expose vaccinated birds to H5N1. It measured antibody responses as a proxy for expected protection, and the trial report itself stated that overall efficacy would not be known until natural exposure occurred.

The vaccine being used now is not the vaccine that was trialled.

The threatened-bird trial used an H5N3 vaccine, Poulvac Flufend (A009733). The approximately 300 birds in the 2026 programme are instead to receive a different H5N2 vaccine (A012218). The earlier threatened-species trial therefore cannot, by itself, establish the safety or protective efficacy of the different product now being administered to these birds.

‘Highly pathogenic’ does not mean every infected species will suffer high mortality.

Pathogenicity concerns the capacity of a virus to cause disease. Different bird species can experience very different disease severity and mortality. The important conservation question is therefore not simply whether H5N1 is classified as HPAI, but what evidence shows that each of these threatened species faces a sufficiently serious risk of disease, death or population decline to justify intervention.

What triggered vaccination?

DOC has described the detection of H5N1 in migratory seabirds in Australia as a ‘trigger to start vaccinating some of our most critically endangered’ New Zealand birds. But neither MPI's One Health document nor the related Cabinet paper identifies a published threshold for moving from preparedness into vaccination. The unanswered question is: what risk-benefit assessment showed that this trigger justified vaccinating these particular populations?

Vaccination itself is not risk-free.

These birds must be captured, restrained and injected, potentially repeatedly if boosters are required. Some free-living birds may then be difficult to recapture and monitor. The risk calculation therefore has two sides: the risk posed by H5N1 to each species versus the known and uncertain risks associated with intervention.

There are important unanswered questions about the original trial.

The trial principally demonstrated an antibody response to an H5N3 poultry vaccine; it did not demonstrate protection following H5N1 challenge. The published material does not provide complete underlying numbers showing how many birds of each species were vaccinated, subsequently tested and followed over time. Importantly, H5N1-specific HI testing was undertaken only in takahē, while chicken-derived HI thresholds were used to infer expected protection. The critical scientific question is therefore how robustly this single immunological proxy predicts protection from disease and death in each of the threatened species.

What are we asking DOC and MPI to disclose?

The OIA requests seek the underlying trial data; deaths, adverse events and longer-term outcomes; species-specific evidence of H5N1 risk; the scientific basis for extrapolating HI antibody responses to protection; the evidence supporting subsequent use of a different H5N2 vaccine; the expected booster regime; and records showing who decided to proceed, according to what criteria, and how the programme was funded. MPI's own documentation for the H5N2 product states: ‘Full efficacy and potency data is pending’ and ‘Duration of immunity is not established.’

How extensively did the authorities deliberate on the risks, benefits and uncertainties?

This is a novel intervention involving core breeding animals from highly threatened species. The responsible authorities should be able to demonstrate that they adequately considered the risks and expected benefits, and that the evidence relied upon was sufficiently robust. In particular, where expected protection rests substantially upon HI antibody responses rather than direct H5N1 challenge evidence, the validity and limitations of that proxy should be explicit. The evidence supporting both the expected benefit and the risks of intervention should be transparent, species-specific and available for independent scrutiny before further intervention occurs.

September Update following the Rejection of the August Official Information Act Requests

The first two requests, sent to MPI OIA26-0675 and DOC OIAD-9589 in August 2026, were rejected (MPI rejection; DOC rejection). The responses pointed to information already publicly available and refused the remaining material principally on the basis that locating and providing it would require substantial collation or research.

Scientifically, and from a policy and governance perspective, significant decisions are usually distilled into a relatively narrow set of records. The underlying scientific literature may be extensive, but the evidence relied upon for a particular decision is ordinarily evaluated, summarised or communicated through identifiable assessments, advice, briefings or decision documents.

This should be particularly true here. The programme concerns a defined and unprecedented intervention in core breeding populations of highly threatened species, involving approximately 300 birds from five species, a specific disease threat and a specific vaccine. The decision was made over a relatively short period and would reasonably be expected to involve a relatively small group of senior scientific, veterinary and programme personnel.

If MPI and DOC determined that the risk from H5N1 justified intervening in these core breeding populations, somebody made that recommendation or decision. If it was scientifically informed, somebody considered the relevant evidence. There should therefore be a reasonably short and identifiable chain connecting the scientific evidence, its assessment by responsible scientific personnel, the advice provided to senior decision-makers, and the recommendation or decision to proceed.

In a democracy, consequential decisions by public agencies should leave an identifiable record showing the evidence considered, the reasoning applied, the advice given and who ultimately authorised the decision.

The published trial cannot, on its own, provide a sufficiently transparent scientific basis for the intervention. It involved very small cohorts, did not undertake viral challenge, used antibody responses as a proxy for protection, and expressly acknowledges that overall efficacy will not be known until natural challenge occurs.

Of particular scientific concern, its principal results are reported as percentages without consistently providing the underlying numbers tested at each time point, despite deaths, difficulties with recapture, incomplete vaccination and adverse events associated with capture and handling. If these percentages and antibody responses formed a material part of the justification for progressing to vaccination of approximately 300 core breeding birds, the absence of transparent denominators, individual outcomes and accounting for loss to follow-up is deeply concerning.

Without these data, attrition and potential attrition bias cannot be independently assessed, nor can the strength and durability of the observed immune response be adequately evaluated. An antibody response used as a proxy for protection cannot itself establish protection against H5N1 disease or mortality.

These are not frivolous or arbitrary requests they arise from a genuine concern that consequential scientific decisions affecting highly threatened species should be demonstrably evidence-based, transparent and accountable. Trust in science depends upon the scientific process being open to scrutiny. Where scientific evidence is relied upon to justify consequential public policy, confidence cannot rest simply upon an agency's declaration that the evidence is trustworthy. The underlying methods, data, assumptions, uncertainties and inferential steps should be sufficiently transparent for the reasoning to be independently examined.

This is particularly important because some quite extraordinary uncertainties prevail: the vaccination programme appears to depend upon a chain of scientific inferences, with a single immunological surrogate providing the principal bridge between the experimental trial and expected protection.

The Inferential Problem

The 2024–2025 trial did not expose vaccinated birds to H5N1 and therefore did not directly test whether vaccination prevented infection, clinical disease, mortality or viral shedding. Instead, expected protection was inferred principally from haemagglutination-inhibition (HI) antibody titres. DOC expressly states that overall efficacy will not be known until natural challenge occurs and defines ‘efficacy’ for the purposes of the trial as the detection of antibodies following vaccination. Antibody persistence is itself described as only an approximate measure of duration of protection.

The inferential problem becomes more significant when the species-specific evidence is examined. H5N1-specific HI testing was undertaken only in takahē. Even in takahē, the measured endpoint was an immunological surrogate rather than demonstrated protection following H5N1 challenge, while the HI thresholds used to interpret those titres as indicating expected protection were derived from studies in chickens. The evidential bridge therefore runs from a chicken-derived relationship between HI titre and protection, to H5N1-specific HI titres measured in takahē, and from that surrogate towards an expectation of protection against H5N1. The trial did not demonstrate that a given HI titre has the same relationship with morbidity or mortality in takahē as it does in chickens. For the other four threatened species, equivalent H5N1-specific HI testing was not reported, requiring a further inferential step if the takahē findings are used to support expectations of protection across species.

These issues are of concern because demonstrating that vaccination induces antibodies is not equivalent to demonstrating that those antibodies reliably predict clinically meaningful protection in a particular species. The scientific strength of a surrogate endpoint depends upon how well the relationship between the surrogate and the outcome of interest has been established in the relevant biological context. Where the outcomes that matter are survival, reduced disease and protection of critically endangered breeding populations, the question is therefore not simply whether HI titres increased, but what evidence validates those titres as predictors of those outcomes in each species concerned.

A further speculative or inferential step using these poultry vaccines arise because the experimental trial used the H5N3 vaccine Poulvac Flufend i AI H5N3 RG, whereas the 2026 programme is proceeding with a different H5N2 vaccine. The reasoning must therefore bridge not only from HI response to expected protection, and potentially from takahē to other species, but also from the H5N3 trial formulation to the H5N2 vaccine now being administered, and ultimately to protection against contemporary H5N1 clade 2.3.4.4b. Each step may be scientifically defensible, but each requires evidence.

These scientific leaps are accompanied by other important uncertainties, including the effects of capture and vaccination stress on breeding populations, prior exposure to H5 variants, and the ultimately unknown protective effect in each species. The brief trial report cannot plausibly represent the entirety of the scientific reasoning behind such a consequential intervention, in which an immunological response measured in a small experimental trial becomes an expectation of protection sufficient to support intervention in core breeding populations.

INTERNATIONAL GUIDANCE

International veterinary guidance similarly recognises these limitations. The World Organisation for Animal Health (WOAH) and the WOAH/FAO Network of Expertise on Animal Influenza (OFFLU) recognise vaccination as a potentially valuable complementary measure for controlling highly pathogenic avian influenza (HPAI), including in particular circumstances involving wild birds. However, WOAH treats vaccination as one component of a wider risk-based control strategy, rather than as a conclusion flowing simply from the detection of antibodies.

WOAH's guidance on emergency vaccination of wild birds describes a decision process encompassing outbreak epidemiology, target populations, the vaccine and vaccination strategy, diagnostic capacity, regulatory considerations and available resources. It also acknowledges that vaccination of free-ranging wildlife remains limited and places vaccination within a broader process of risk assessment and coordinated response planning.

Of particular relevance to the question of vaccine matching, OFFLU's Avian Influenza Matching (AIM) programme exists because antigenic variation among circulating HPAI viruses can reduce vaccine effectiveness. OFFLU notes that this is particularly important for killed, inactivated adjuvanted vaccines that depend substantially upon humoral immune responses. The OFFLU-AIM programme therefore compares the antigenic characteristics of circulating viruses with vaccine antigens to assist governments in selecting and updating appropriate vaccines. Its July 2024 assessment reports substantial variation in antigenic distance between contemporary H5 viruses and different vaccine antigens.

The fact that a vaccine and circulating virus both possess an H5 haemagglutinin therefore does not, by itself, establish equivalent protective efficacy. Relevant considerations include antigenic matching, vaccine formulation and potency, dose and vaccination regime, the immune response of the target species, and the relationship between measured antibody titres and clinically meaningful protection.

PSGR NZ therefore considers it reasonable to expect identifiable scientific records showing how DOC addressed these evidential bridges before progressing from a small experimental H5N3 trial to vaccination of approximately 300 core breeding birds using a different H5N2 vaccine. There should likewise be an identifiable record of the scientific evidence and advice informing MPI's role in the programme and the pathway by which the decision to proceed was reached.

Policy-relevant science depends upon more than an assertion that a decision was evidence-based. The evidence, reasoning and uncertainties must be capable of scrutiny. If the scientific basis for this intervention is sound and defensible, the principal records demonstrating that basis should be capable of public release.

H5N1 VACCINATION OF ENDANGERED NATIVE BIRDS

September Update following the Rejection of the August Official Information Act Requests

HAVE YOU CHECKED OUT OUR SUBSTACK:

SCIENCE, STEWARDSHIP & SCALABILITY?

Information

Providing scientific & medical information & analysis in the service of the public's right to be independently informed on issues relating to human & environmental health.