Abstract
Since the first COVID‐19 case, the Aotearoa New Zealand response has been adjusted to match the evolving virus, transitioning from elimination to protection. Using whole‐of‐population data, this study aimed to examine variables associated with COVID‐19 vaccination among Māori and quantify the gap in COVID‐19 immunisation coverage between Māori and non‐Māori. A retrospective study was conducted using microdata from Stats NZ. Logistic regression was used to estimate the odds of being unvaccinated or only partially vaccinated as a function of the independent variables, adjusting for the effects of all other variables in the adjusted model. Primary Health Organisation (PHO) enrolees (adjusted odds ratio [aOR] = 0.10, 95% confidence interval [CI]: 0.10–0.11) and individuals with a household income exceeding NZ$100,001 per annum (aOR = 0.33, 95% CI: 0.30–0.36) had lower odds of being unvaccinated (more likely to be vaccinated). Disability (aOR = 1.52, 95% CI: 1.48–1.56), household crowding (aOR = 1.20, 95% CI: 1.17–1.24), poor housing quality (aOR = 1.10, 95% CI: 1.07–1.14), three or more generation family type (aOR = 1.17, 95% CI: 1.13–1.21) and the presence of a child aged under five in the household (aOR = 1.51, 95% CI: 1.47–1.55) were associated with being unvaccinated. We found sizeable gaps in COVID‐19 vaccination rates between Māori and non‐Māori.
Keywords: Aotearoa New Zealand, COVID‐19 vaccination, Health equity , Māori and Non‐Māori , severe acute respiratory syndrome coronavirus 2, vaccine uptake
1. Introduction
Coronavirus disease (COVID‐19) is an infectious disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) (World Health Organization 2023). Since the confirmation of the first COVID‐19 case in Aotearoa New Zealand on February 28, 2020, as of February 2024, there have been over 2,580,365 cases, with 37,502 hospitalisations, 1072 intensive care unit admissions and 5479 deaths (Te Whatu Ora Health New Zealand 2024b). Severe COVID‐19 outcomes have been associated with older age, smoking, obesity, comorbidities, immunocompromising conditions and pregnancy (Barron et al. 2020; Williamson et al. 2020; Reddy et al. 2021). The risk of COVID‐19 mortality was found to increase with age, relative socioeconomic deprivation, disability status and comorbidity (Williamson et al. 2020). In Aotearoa New Zealand, age‐adjusted COVID‐19 mortality was found to be higher among Māori and Pacific peoples compared to the general population (Jefferies et al. 2020; Sonder et al. 2022; Ministry of Health 2024b). Māori have higher rates of hospitalisation and death following a positive test compared to non‐Māori (Satherley et al. 2023). At the outset of the pandemic, Aotearoa New Zealand pursued a strategy focused on suppressing the community spread of SARS‐CoV‐2 (Jefferies et al. 2020). The approach to COVID‐19 was adjusted in an attempt to respond to the evolving virus, transitioning from elimination to minimisation and prioritising protection (Ministry of Health 2024b).
A vaccination (once available) is one of the tools that can be used to reduce the impact of any new or emerging infectious disease pandemic (Rodrigues and Plotkin 2020). This is particularly important in countries including Aotearoa New Zealand, which do not have a limitless healthcare system capacity in the event of a pandemic (Health Quality and Safety Commission 2015; Goodyear‐Smith and Ashton 2019; Hitchon et al. 2024). The national medical regulatory agency granted provisional approval for the use of the Pfizer‐BioNTech COVID‐19 vaccine (Comirnaty) on February 3, 2021. An equity‐driven vaccination prioritisation strategy was developed. This was facilitated by widespread and rapid vaccine delivery amid uncertain and evolving conditions (Satherley et al. 2023). The vaccine rollout commenced ith priority groups (those most at risk of contracting COVID‐19 and experiencing poor health outcomes from infection, e.g. border and managed isolation and quarantine workers, their household contacts and families) on February 20, 2021 (Ministry of Health 2024a).
From January 2022, all individuals aged 5 and older in Aotearoa New Zealand became eligible for vaccination. Additionally, in February 2023, vaccination eligibility was extended to certain infants as young as 6 months (Ministry of Health 2024a). Several additional vaccines have been approved and administered to a limited segment of the vaccinated population. These include AstraZeneca (Vaxzevria, approved in July 2021), Janssen (July 2021), Novavax (Nuvaxovid, March 2022) and Bivalent mRNA‐CVs (December 2022) (Ministry of Health 2024a). According to the Aotearoa Immunisation Register, as of February 2024, more than 12,958,300 COVID‐19 vaccinations, including 1,032,211 boosters, have been administered in Aotearoa New Zealand (Te Whatu Ora Health New Zealand 2024a). Enabling infrastructure and systems such as innovative service delivery models, data sharing, effective communication, targeted campaigns, a diverse and deep workforce, technology tools, transparency and secured supply contributed to high vaccine uptake (Ministry of Health 2021).
The COVID‐19 pandemic has highlighted and exacerbated health inequities between Māori and other New Zealanders. For example, a recent study using the June 2020 ‘Administrative Population Census (APC) resident population estimates' found that an excess of 2661 Māori tested positive, 1333 were hospitalised, and 162 died, compared to what would be expected if Māori experienced these outcomes at the same rate as non‐Māori (Satherley et al. 2023).
Minimising the impact and inequity of the epidemic requires the accurate quantification of vaccine uptake and timing. We previously analysed the daily case and immunisation data provided by the Ministry of Health and found that vaccine uptake recording was incomplete (particularly for Māori) and that, despite vaccination reducing case rates, some people were still experiencing severe COVID‐19 outcomes. There was, therefore, a need to further investigate the factors associated with vaccine uptake, in particular, to support efforts to optimise the COVID‐19 vaccination programme for Māori, who are more susceptible to severe COVID‐19 outcomes (Jefferies et al. 2020).
Building on the integrated data from Stats NZ microdata, we aimed to examine independent variables associated with COVID‐19 vaccination among Māori and quantify the gap in COVID‐19 immunisation coverage between Māori and non‐Māori.
2. Materials and Methods
2.1. Study Design and Population
A retrospective population study was conducted using microdata from the Stats NZ Integrated Data Infrastructure (IDI). The IDI is a database of linked, deidentified individual and household‐level administrative data. For this study, data were sourced from several collections in the IDI, including the 2018 Census, Ministry of Health and core data derived by Stats NZ, including the APC, full death dates and address notifications (Figure 1).
FIGURE 1.
Summary of data joining steps for variables of interest in the IDI.
We used two separate denominator sources to present population counts of vaccination status and regression models of vaccination status, respectively. The 2018 Census Usually Resident Population was used for the regression models, as most independent variables of interest were only available in the 2018 Census. However, for COVID‐19 vaccination counts stratified by ethnicity and age, we used Stats NZ's June 30, 2020, APC (Stats NZ 2022) because it provides a more up‐to‐date count of the Māori and non‐Māori populations, during (rather than two years before) the COVID‐19 pandemic. The APC provides a census‐like count of the Aotearoa New Zealand resident population (specified on June 30, 2020, for the present study), based on the presence of administrative records for individuals in the IDI (e.g. tax data, migration, birth and death records) that indicate the residence of an individual in Aotearoa New Zealand. The information on ethnicity in the APC is obtained using a source‐ranked approach of other data sources in the IDI (Stats NZ 2024). This involves assessing the potential of administrative data to provide ethnicity information comparable to the current census. This enables the codes for ethnicity to be taken from the most reliable source of information available for each individual, rather than from a single source.
2.2. Outcomes
2.2.1. COVID‐19 Vaccination Status
Most of the Aotearoa New Zealand population received the Pfizer‐BioNTech (Comirnaty) vaccine (Ministry of Health 2024a). In collaboration with mainstream services and multiple stakeholders, the Ministry of Health delivered COVID‐19 vaccinations on a large scale, targeting communities with traditionally lower access to health services. The COVID‐19 vaccine was administered by regulated (e.g. authorised vaccinators, pharmacists and intern pharmacists) and unregulated (e.g. pharmacy technicians, kaiāwhina, kaiārahi, Kaimanaaki, healthcare assistants and practitioners) healthcare professionals who had completed the appropriate COVID‐19 vaccinator education (Ministry of Health 2024a).
We used codes provided directly by the Ministry of Health to determine vaccination status. This included data on whether an individual was fully vaccinated (1) or partially vaccinated (2) up to May 2023, the latest date of available records in the IDI. Individuals who were part of a given population denominator, but did not have a vaccination status recorded in the Ministry of Health data, were assumed to be unvaccinated (0). Qualifying as fully vaccinated vs. partially vaccinated is dependent on factors such as age, time since the last primary dose and/or booster, and whether an individual is at an increased risk of severe COVID‐19 infection (Ministry of Health 2024a). This is taken into consideration within the vaccination status records as provided by the Ministry of Health. Fully vaccinated is defined as having completed the primary course of vaccination. For the majority of people, this meant receiving two doses of the vaccine. For some groups, more than two doses are required to meet the fully vaccinated definition (e.g. immunocompromised infants requiring three doses). Since March 7, 2024, with the release of the new XBB COVID‐19 vaccine, the definition of the primary course has changed, with only one initial dose being required as the primary course of vaccination. Partially vaccinated is defined as having received at least one dose but not having completed the primary course of vaccination. For the majority of people in this category, this essentially means that someone has received one dose of the old vaccine. Unvaccinated means zero doses have been received by the person (Ministry of Health 2024a; Te Whatu Ora Health New Zealand 2025). Full and partial vaccination can significantly decrease the risk of COVID‐19 hospitalisation and mortality, alleviating the strain on hospitals and healthcare systems and ensuring their capacity to remain accessible during times of critical demand (COVID‐19 Group 2023; Mbinta et al. 2024).
2.3. Covariates
We developed the Directed Acyclic Graph (DAG) using the Evidence Synthesis for Constructing Directed Acyclic Graphs (ESC‐DAGs) protocol (Ferguson et al. 2020) and ggdag (R package for visualising and analysing DAGs) (Barrett 2025). This process enabled us to construct an evidence‐based conceptual model, providing a visual representation of covariates in our multivariable analysis of COVID‐19 vaccine uptake among Māori. Additional information can be found in the Supporting Information (Figure S1 and Table S1).
2.3.1. Household Independent Variables
2.3.1.1. Household Composition
Household composition was obtained through the 2018 Census variable, which records the following categories: one‐family household, two‐family household, three or more family household, other multiperson household and one‐person household.
2.3.1.1.1. Extended Family Type
These were obtained from the 2018 Census and coded as 0, less than three generations living together, or 1, three or more generations together (e.g. a couple with children and grandchildren or grandparents and grandchildren).
2.3.1.1.2. Child Under Five
The presence of a child under five was based on the age of the youngest child in a family/household in the 2018 Census. A code of 0 indicates no child under five was present, and a code of 1 indicates a child under five was present. Note that because COVID‐19 outcomes occurred during 2020–2023, children would be under the age of 10 in a household when the COVID‐19 outcome occurred. As such, the variable captures a 0–5‐year‐old cohort at the time of the 2018 Census,P but will miss children born after the 2018 Census and may, therefore, underestimate the effect of the presence of a child under five.
2.3.1.1.3. Household Crowding
Household crowding was coded from the 2018 Census variable used by Stats NZ, based on the Canadian National Occupancy Standard (Stats NZ 2012). Those not needing additional bedrooms were coded 0 (not crowded), and those needing one or more additional bedrooms were coded 1 (crowded).
2.3.1.1.4. Housing Quality
Housing quality was derived from the dwelling dampness and mould variables in the 2018 Census. A code of 1 represents a dwelling that was always damp or that always had mould over A4 in size, and a code of 0 represents the presence of dampness or mould sometimes or never.
2.3.1.1.5. Household Income
Jensen equivalised household income was obtained from the 2018 Census and coded into the following categories: loss of or no income, NZ$1–NZ$15,000, NZ$15,001–NZ$30,000, NZ$30,001–NZ$50,000, NZ$50,001–NZ$70,000, NZ$70,001–NZ$100,000 and NZ$100,001+.
2.3.2. Individual Independent Variables
2.3.2.1. Age and Sex
Age in years, birth month and year and sex were obtained from the 2018 Census. Age in years at the 2018 Census was used for the regression models. For analyses presenting vaccination status by age band, age was recoded to reflect age on June 30, 2022. This meant age bands were more relevant to vaccination availability (both the timing of the vaccine rollout and the minimum age criteria for general vaccinations, i.e. 5+ years). For the regression models, age was divided by 10, with each unit increase therefore corresponding to an additional 10 years (rather than 1 year) of age. This made the effect size for age easier to demonstrate in the results.
2.3.2.2. Ethnicity
Māori ethnicity was coded for those recorded as Māori only or Māori and at least one other ethnic group in the 2018 Census for the logistic regression models of vaccination status. For overall population counts of vaccination status, Māori ethnicity was obtained from the APC table, and it was backfilled using ethnicity in the IDI personal detail table where missing.
2.3.2.3. Disability Status
Disability status was coded as 0 (not disabled) or 1 (disabled) using the 2018 Census. This is based on Stats NZ's definition of disability, that is, having a lot of difficulty or being unable to do one or more of six activities (seeing, hearing, walking, or climbing steps, remembering, or concentrating, washing all over or dressing or communicating) (Stats NZ 2014).
2.3.2.4. Primary Health Organisation Enrolment
Primary Health Organisation (PHO) enrolment was derived from the National Enrolment Service (NES) records as an indicator variable (not enrolled = 0 and enrolled = 1) based on enrolment status in the month prior (July 2021) to the start of the vaccination rollout for the general population.
2.3.2.5. Residential Mobility
This was derived as a count of the number of unique property IDs recorded for each individual during 2018–2019 (before the pandemic) in the address notification table (thus the number of implied address changes). This data table contains a list of residential addresses reported for an individual obtained through their interaction with various government agencies, derived by Stats NZ. Individuals without a record in the address notification table during this period were coded as 0 (with the assumption that they had not changed their address during this time). Residential mobility was then categorised as no moves or one move (coded 0; low mobility), two or three moves (coded 1; medium mobility) or 4 or more moves (coded 2; high mobility) during the 2‐year period.
2.4. Statistical Analysis
We tabulated counts of vaccination status and calculated vaccination rates by age groups, separately for Māori and non‐Māori. We then computed ‘excess’ vaccination outcomes for Māori compared to non‐Māori (e.g. the number of additional Māori who would have been fully vaccinated against COVID‐19 had the Māori vaccination rate been the same as the non‐Māori rate).
Logistic regression was used to assess vaccination status among Māori (being partially vaccinated vs. fully vaccinated, not vaccinated vs. fully vaccinated and not vaccinated vs. partially vaccinated). We estimated unadjusted odds ratios (ORs) by running separate logistic regression models for each independent variable without covariate adjustment. In the adjusted models, the change in odds of being unvaccinated or only partially vaccinated associated with a one‐unit increase in each independent variable (e.g. each additional 10 years of age, or for binary coded variables, for the change from one level of the variable to the other, such as living in a crowded, relative to an uncrowded household) while adjusting for the effects of all other independent variables. Age was also modelled with a nonlinear specification. These estimates are ORs, which indicate how many times greater (for ORs greater than 1) or lower (for ORs below 1) the odds of an outcome (COVID‐19 vaccination status) are with each unit increase in each independent variable. The 95% confidence intervals (CIs) corresponding to each OR provide a general indication of uncertainty around the effect (wider CIs reflect less precision). Collinearity was assessed using Stata's _rmcoll command, which revealed no collinearity among the included independent variables. The model predominantly comprised categorical variables and transformed continuous terms, thereby minimising the risk of influential outliers. The analysis was specifically designed to estimate associations rather than optimise predictive performance; therefore, no stepwise variable selection procedures were used. Because the data consisted of people within households experiencing similar circumstances (e.g. people within households all experience the same housing quality and crowding), and COVID‐19 tends to spread within households, we used generalised estimating equations with robust standard errors for all regression models, with household as the cluster variable to account for potential nonindependence in model residuals (Hanley et al. 2003). Analyses were performed in Stata 18.0 using the xtgee command.
2.5. Ethics and Permissions
An ethics approval (number: 30627) was obtained from the Human Ethics Committee of Te Herenga Waka–Victoria University of Wellington, Aotearoa New Zealand.
3. Results
3.1. COVID‐19 Vaccination Counts, Rates and Excess Outcomes Stratified by Age and Ethnicity
The analysis comprised of three comparison groups: unvaccinated versus fully vaccinated (N = 425,208), partially vaccinated versus fully vaccinated (N = 360,495), and unvaccinated versus partially vaccinated (N = 94,260). In all regression samples, the majority of individuals were enrolled in a PHO, had no disabilities, lived in one‐family households and reported satisfactory housing conditions. Females slightly outnumbered males, and most households did not have children under five. Higher household income, particularly in income levels 4 and 5, and lower residential mobility were more prevalent among those who were fully vaccinated. For instance, 72.6% of fully vaccinated individuals were in the top two income bands, compared to 45.7% of the partially vaccinated and 30.5% of the unvaccinated (Table S2).
Tables 1–3 show differences in vaccination between Māori and non‐Māori across age groups. Across all groups, females slightly outnumber males, with the majority enrolled in a PHO. Most participants had no disability, did not live with extended family and resided in one‐family households. Higher income levels, better housing quality and the absence of crowding are more common among those who are fully vaccinated. Māori in the 60–64, 65–69, and 70–74 age bands were most likely to be fully vaccinated. The full vaccination rate was 839 per 1000 in adults 50 years and older. Overall, 78,880 fewer Māori across all age bands were fully vaccinated than would be expected if vaccination rates were the same as for non‐Māori.
TABLE 1.
Counts and rates of fully vaccinated individuals against COVID‐19 by age group for Māori and non‐Māori as of May 2023.
| Age group (June 2022) | Fully vaccinated | June 2020 population (APC) | Fully vaccinated per 1000 (in age group) | Fully vaccinated, assuming non‐Māori rate (count) | Māori non‐Māori rate difference (excess) |
|---|---|---|---|---|---|
| Māori | |||||
| 0–4 | 72 | 50,502 | 1.43 | 404 | −332 |
| 5–9 | 11,619 | 83,688 | 138.84 | 28,709 | −17,090 |
| 10–14 | 43,254 | 88,752 | 487.36 | 58,676 | −15,422 |
| 15–19 | 65,082 | 77,502 | 839.75 | 71,480 | −6398 |
| 20–24 | 58,587 | 71,316 | 821.51 | 64,759 | −6172 |
| 25–29 | 53,202 | 67,353 | 789.90 | 59,308 | −6106 |
| 30–34 | 48,762 | 62,343 | 782.16 | 54,787 | −6025 |
| 35–39 | 39,204 | 49,305 | 795.13 | 43,854 | −4650 |
| 40–44 | 36,615 | 44,925 | 815.03 | 40,245 | −3630 |
| 45–49 | 37,212 | 44,427 | 837.60 | 40,112 | −2900 |
| 50–54 | 37,485 | 44,409 | 844.09 | 40,053 | −2568 |
| 55–59 | 34,035 | 40,341 | 843.68 | 36,348 | −2313 |
| 60–64 | 30,126 | 35,202 | 855.80 | 31,900 | −1774 |
| 65–69 | 21,384 | 25,008 | 855.09 | 22,663 | −1279 |
| 70–74 | 14,895 | 17,409 | 855.59 | 15,830 | −935 |
| 75–79 | 8328 | 9876 | 843.26 | 8961 | −633 |
| 80+ | 7650 | 9843 | 777.20 | 8305 | −655 |
| Total | 547,512 | 822,201 | 626,392 | −78,880 | |
| Non‐Māori | |||||
| 0–4 | 999 | 124,959 | 7.99 | — | — |
| 5–9 | 78,354 | 228,408 | 343.04 | — | — |
| 10–14 | 161,529 | 244,326 | 661.12 | — | — |
| 15–19 | 219,483 | 237,975 | 922.29 | — | — |
| 20–24 | 228,711 | 251,868 | 908.06 | — | — |
| 25–29 | 256,533 | 291,330 | 880.56 | — | — |
| 30–34 | 282,129 | 321,039 | 878.80 | — | — |
| 35–39 | 260,637 | 293,037 | 889.43 | — | — |
| 40–44 | 239,898 | 267,798 | 895.82 | — | — |
| 45–49 | 240,276 | 266,124 | 902.87 | — | — |
| 50–54 | 257,034 | 284,988 | 901.91 | — | — |
| 55–59 | 248,208 | 275,472 | 901.03 | — | — |
| 60–64 | 241,701 | 266,721 | 906.19 | — | — |
| 65–69 | 211,113 | 232,962 | 906.21 | — | — |
| 70–74 | 187,338 | 206,022 | 909.31 | — | — |
| 75–79 | 141,588 | 156,051 | 907.32 | — | — |
| 80+ | 191,697 | 227,184 | 843.80 | — | — |
| Total | 3,447,228 | 4,176,264 | — | — | — |
Note: Denominators are June 2020 APC population estimates, and age groups are as of June 2022. The COVID‐19 vaccines were initially available to specific groups based on a prioritisation schedule. However, since January 2022, everyone aged 5 years and older has been eligible for a primary vaccination course. In February 2023, vaccination was extended to high‐risk infants aged 6 months to 4 years.
Abbreviation: APC = Administrative Population Census.
TABLE 3.
Counts and rates of unvaccinated individuals against COVID‐19 by age group for Māori and non‐Māori as of May 2023.
| Age group (June 2022) | Unvaccinated | June 2020 population (APC) | Unvaccinated per 1000 (in age group) | Unvaccinated assuming non‐Māori rate (count) | Māori non‐Māori rate difference (excess) |
|---|---|---|---|---|---|
| Māori | |||||
| 0–4 | 50,298 | 50,502 | 995.96 | 49,531 | 767 |
| 5–9 | 57,633 | 83,688 | 688.67 | 35,484 | 22,149 |
| 10–14 | 32,937 | 88,752 | 371.11 | 18,091 | 14,846 |
| 15–19 | 9807 | 77,502 | 126.54 | 5359 | 4448 |
| 20–24 | 10,410 | 71,316 | 145.97 | 5920 | 4490 |
| 25–29 | 11,940 | 67,353 | 177.27 | 7384 | 4556 |
| 30–34 | 11,628 | 62,343 | 186.52 | 6963 | 4665 |
| 35–39 | 8727 | 49,305 | 177.00 | 5009 | 3718 |
| 40–44 | 7305 | 44,925 | 162.60 | 4339 | 2966 |
| 45–49 | 6438 | 44,427 | 144.91 | 4013 | 2425 |
| 50–54 | 6264 | 44,409 | 141.05 | 4081 | 2183 |
| 55–59 | 5847 | 40,341 | 144.94 | 3775 | 2072 |
| 60–64 | 4758 | 35,202 | 135.16 | 3139 | 1619 |
| 65–69 | 3432 | 25,008 | 137.24 | 2247 | 1185 |
| 70–74 | 2421 | 17,409 | 139.07 | 1518 | 903 |
| 75–79 | 1461 | 9876 | 147.93 | 874 | 587 |
| 80+ | 2085 | 9843 | 211.83 | 1446 | 639 |
| Total | 233,391 | 822,201 | 159,172 | 74,219 | |
| Non‐Māori | |||||
| 0–4 | 122,556 | 124,959 | 980.77 | — | — |
| 5–9 | 96,846 | 228,408 | 424.00 | — | — |
| 10–14 | 49,803 | 244,326 | 203.84 | — | — |
| 15–19 | 16,455 | 237,975 | 69.15 | — | — |
| 20–24 | 20,907 | 251,868 | 83.01 | — | — |
| 25–29 | 31,938 | 291,330 | 109.63 | — | — |
| 30–34 | 35,856 | 321,039 | 111.69 | — | — |
| 35–39 | 29,772 | 293,037 | 101.60 | — | — |
| 40–44 | 25,866 | 267,798 | 96.59 | — | — |
| 45–49 | 24,039 | 266,124 | 90.33 | — | — |
| 50–54 | 26,187 | 284,988 | 91.89 | — | — |
| 55–59 | 25,779 | 275,472 | 93.58 | — | — |
| 60–64 | 23,781 | 266,721 | 89.16 | — | — |
| 65–69 | 20,928 | 232,962 | 89.83 | — | — |
| 70–74 | 17,961 | 206,022 | 87.18 | — | — |
| 75–79 | 13,815 | 156,051 | 88.53 | — | — |
| 80+ | 33,381 | 227,184 | 146.93 | — | — |
| Total | 615,870 | 4,176,264 | — | — | |
Note: Denominators are June 2020 APC population estimates, and age groups are as of June 2022. The COVID‐19 vaccines were initially available to specific groups based on a prioritisation schedule. However, since January 2022, everyone aged 5 years and older has been eligible for a primary vaccination course. In February 2023, vaccination was extended to high‐risk infants aged 6 months to 4 years.
Abbreviation: APC = Administrative Population Census.
Māori had higher rates of partial vaccination compared to non‐Māori, except for the 0–4 and 5–9 age bands, suggesting Māori children were less likely to have started the vaccination process (Table 2). The COVID‐19 vaccines were initially available to specific groups based on a prioritisation schedule. However, since January 2022, everyone aged 5 years and older has been eligible for a primary vaccination course. In February 2023, vaccination was extended to high‐risk infants aged 6 months to 4 years (Ministry of Health 2024a). There was an excess of 74,219 Māori people who did not receive any COVID‐19 vaccination dose (first dose, second dose or booster) across age bands, compared to the expected count if vaccination rates were the same as for non‐Māori (Table 3).
TABLE 2.
Counts and rates of partially vaccinated individuals against COVID‐19 by age group for Māori and non‐Māori as of May 2023.
| Age group (June 2022) | Partially vaccinated | June 2020 population (APC) | Partially vaccinated per 1000 (in age group) | Partially vaccinated, assuming non‐Māori rate (count) | Māori non‐Māori rate difference (excess) |
|---|---|---|---|---|---|
| Māori | |||||
| 0–4 | 135 | 50,502 | 2.67 | 567 | −432 |
| 5–9 | 14,439 | 83,688 | 172.53 | 19,496 | −5057 |
| 10–14 | 12,567 | 88,752 | 141.60 | 11,985 | 582 |
| 15–19 | 2610 | 77,502 | 33.68 | 663 | 1947 |
| 20–24 | 2319 | 71,316 | 32.52 | 638 | 1681 |
| 25–29 | 2214 | 67,353 | 32.87 | 660 | 1554 |
| 30–34 | 1950 | 62,343 | 31.28 | 592 | 1358 |
| 35–39 | 1377 | 49,305 | 27.93 | 442 | 935 |
| 40–44 | 1011 | 44,925 | 22.50 | 341 | 670 |
| 45–49 | 777 | 44,427 | 17.49 | 303 | 474 |
| 50–54 | 660 | 44,409 | 14.86 | 276 | 384 |
| 55–59 | 456 | 40,341 | 11.30 | 217 | 239 |
| 60–64 | 318 | 35,202 | 9.03 | 164 | 154 |
| 65–69 | 189 | 25,008 | 7.56 | 99 | 90 |
| 70–74 | 93 | 17,409 | 5.34 | 61 | 32 |
| 75–79 | 87 | 9876 | 8.81 | 41 | 46 |
| 80+ | 105 | 9843 | 10.67 | 91 | 14 |
| Total | 41,307 | 822,201 | 36,638 | 4604 | |
| Non‐Māori | |||||
| 0–4 | 1404 | 124,959 | 11.24 | — | — |
| 5–9 | 53,211 | 228,408 | 232.96 | — | — |
| 10–14 | 32,994 | 244,326 | 135.04 | — | — |
| 15–19 | 2037 | 237,975 | 8.56 | — | — |
| 20–24 | 2253 | 251,868 | 8.95 | — | — |
| 25–29 | 2856 | 291,330 | 9.80 | — | — |
| 30–34 | 3051 | 321,039 | 9.50 | — | — |
| 35–39 | 2628 | 293,037 | 8.97 | — | — |
| 40–44 | 2034 | 267,798 | 7.60 | — | — |
| 45–49 | 1815 | 266,124 | 6.82 | — | — |
| 50–54 | 1770 | 284,988 | 6.21 | — | — |
| 55–59 | 1482 | 275,472 | 5.38 | — | — |
| 60–64 | 1239 | 266,721 | 4.65 | — | — |
| 65–69 | 924 | 232,962 | 3.97 | — | — |
| 70–74 | 723 | 206,022 | 3.51 | — | — |
| 75–79 | 648 | 156,051 | 4.15 | — | — |
| 80+ | 2106 | 227,184 | 9.27 | — | — |
| Total | 113,175 | 4,176,264 | — | — | — |
Note: Denominators are June 2020 APC population estimates, and age groups are as of June 2022. The COVID‐19 vaccines were initially available to specific groups based on a prioritisation schedule. However, since January 2022, everyone aged 5 years and older has been eligible for a primary vaccination course. In February 2023, vaccination was extended to high‐risk infants aged 6 months to 4 years.
Abbreviation: APC = Administrative Population Census.
3.2. Independent Variables Associated With COVID‐19 Vaccination Status
The results of logistic regression models of COVID‐19 vaccination status (fully vaccinated, partially vaccinated and unvaccinated) among Māori are displayed in Tables 4–6.
TABLE 4.
Logistic regression models of vaccination status as of May 2023 among Māori (full vaccination vs. no vaccination).
| Characteristics | OR (95% CI)a | aOR (95% CI)b |
|---|---|---|
| Sex | ||
| Male (ref) | — | — |
| Female | 1.191 (1.179–1.204) | 0.996 (0.980–1.00) |
| Age | ||
| Age (per 10 years) | 0.758 (0.755–0.761) | 0.865 (0.861–0.869) |
| Primary Health Organisation enrolment | ||
| No (ref) | — | — |
| Yes | 0.149 (0.146–0.152) | 0.102 (0.100–0.105) |
| Disability status | ||
| Nondisabled (ref) | — | — |
| Disabled | 1.521 (1.483–1.560) | 1.522 (1.481–1.564) |
| Residential mobility (2018–2019) | ||
| 0–1 moves (ref) | — | — |
| 2–3 moves | 0.827 (0.816–0.839) | 0.907 (0.888–0.926) |
| 4+ moves | 0.808 (0.783–0.834) | 0.939 (0.897–0.983) |
| Child under five | ||
| No (ref) | — | — |
| Yes | 3.292 (3.244–3.340) | 1.510 (1.472–1.548) |
| Extended family | ||
| <3 generations (ref) | — | — |
| ≥3 generations | 1.398 (1.370–1.426) | 1.170 (1.127–1.214) |
| Household overcrowding | ||
| No (ref) | — | — |
| Yes | 1.772 (1.740–1.803) | 1.200 (1.165–1.237) |
| Housing quality | ||
| Never or sometimes damp or mould (ref) | — | — |
| Always | 1.429 (1.399–1.461) | 1.102 (1.069–1.136) |
| Household income | ||
| No income (ref) | — | — |
| $1–$15,000 | 0.952 (0.886–1.024) | 0.954 (0.864–1.054) |
| $15,001–$30,000 | 0.952 (0.889–1.019) | 0.962 (0.876–1.056) |
| $30,001–$50,000 | 0.741 (0.692–0.793) | 0.756 (0.688–0.830) |
| $50,001–$100,000 | 0.589 (0.551–0.630) | 0.541 (0.493–0.593) |
| $100,001+ | 0.371 (0.347–0.397) | 0.332 (0.303–0.364) |
| Household composition | ||
| One family (ref) | — | — |
| Two or more | 1.264 (1.235–1.293) | 1.115 (1.068–1.164) |
| Multiperson | 0.824 (0.800–0.849) | 0.835 (0.798–0.874) |
| One person | 0.906 (0.884–0.928) | 0.969 (0.934–1.006) |
Note: Unvaccinated vs. fully vaccinated (N = 425,208).
OR = unadjusted odds ratio (unadjusted logistic regression model—each predictor run separately).
aOR = adjusted odds ratio (logistic regressions were estimated using generalised estimating equations with household as the clustering variable and robust standard errors).
TABLE 6.
Logistic regression models of vaccination status as of May 2023 among Māori (partially vaccinated vs. fully vaccinated).
| Characteristics | OR (95% CI)a | aOR (95% CI)b |
|---|---|---|
| Sex | ||
| Male (ref) | — | — |
| Female | 1.243 (1.213–1.274) | 1.126 (1.089–1.165) |
| Age | ||
| Age 10 | 0.328 (0.320–0.336) | 0.397 (0.385–0.409) |
| Primary Health Organisation enrolment | ||
| No (ref) | — | — |
| Yes | 0.942 (0.893–0.993) | 0.879 (0.823–0.940) |
| Disability status | ||
| Nondisabled (ref) | — | — |
| Disabled | 0.848 (0.792–0.908) | 1.245 (1.155–1.342) |
| Residential mobility | ||
| 0–1 moves (ref) | — | — |
| 2–3 moves | 1.003 (0.973–1.034) | 0.932 (0.895–0.970) |
| 4+ moves | 1.045 (0.983–1.111) | 1.047 (0.964–1.136) |
| Child under five | ||
| No (ref) | — | — |
| Yes | 4.752 (4.627–4.880) | 1.767 (1.698–1.838) |
| Extended family | ||
| <3 generations (ref) | — | — |
| ≥3 generations | 1.356 (1.307–1.406) | 1.075 (1.009–1.145) |
| Household overcrowding | ||
| No (ref) | — | — |
| Yes | 1.643 (1.591–1.697) | 1.151 (1.097–1.207) |
| Housing quality | ||
| Never or sometimes damp or mould (ref) | — | — |
| Always | 1.266 (1.214–1.319) | 1.058 (1.003–1.117) |
| Household income | ||
| No income (ref) | — | — |
| $1–$15,000 | 0.959 (0.825–1.115) | 0.978 (0.792–1.209) |
| $15,001–$30,000 | 1.004 (0.872–1.158) | 0.889 (0.727–1.087) |
| $30,001–$50,000 | 0.793 (0.688–0.913) | 0.761 (0.623–0.929) |
| $50,001–$100,000 | 0.751 (0.653–0.862) | 0.619 (0.508–0.753) |
| $100,001+ | 0.578 (0.503–0.664) | 0.468 (0.385–0.570) |
| Household composition | ||
| One family (ref) | — | — |
| Two or more | 1.108 (1.063–1.155) | 1.081 (1.007–1.161) |
| Multiperson | 0.280 (0.254–0.309) | 0.631 (0.548–0.726) |
| One person | 0.255 (0.234–0.278) | 2.135 (1.861–2.450) |
Note: Partially vs. fully vaccinated (N = 360,495).
OR = unadjusted odds ratio (unadjusted logistic regression model—each predictor run separately).
aOR = adjusted odds ratio (logistic regressions were estimated using generalised estimating equations with household as the clustering variable and robust standard errors).
For those not receiving any COVID‐19 vaccine dose (compared to those fully vaccinated), the following factors were associated with greater odds of not receiving any vaccine dose: disability status (aOR = 1.52, 95% CI: 1.48–1.56), the presence of a child under five in the household (aOR = 1.51, 95% CI: 1.47–1.55), living with three or more generations of extended family (aOR = 1.17, 95% CI: 1.13–1.21), household crowding (aOR = 1.20, 95% CI: 1.17–1.24), poorer housing quality (aOR = 1.10, 95% CI: 1.07–1.14) and a household composition of two or more families (aOR = 1.12, 95% CI: 1.07–1.16). Older people (aOR = 0.87, 95% CI: 0.86–0.87), recent PHO enrolees (aOR = 0.10, 95% CI: 0.10–0.11), higher household income bands (aOR range = 0.33–0.76) and multiperson households (aOR = 0.84, 95% CI: 0.80–0.87) had lower odds of not receiving any COVID‐19 vaccine. There was no difference in full vaccination uptake between females and males, and different levels of residential mobility.
There were fewer statistically significant factors associated with being unvaccinated compared to being partially vaccinated (Table 5). These factors include an additional 10 years of age (aOR = 1.44, 95% CI: 1.42–1.46), recent PHO enrolment status (aOR = 0.22, 95% CI: 0.21–0.23), residential mobility (aOR = 1.06, 95% CI: 1.02–1.11) and having a child under five (aOR = 1.08, 95% CI: 1.04–1.13). The household and individual independent variables of being partially vs. fully vaccinated were broadly consistent with those of being unvaccinated vs. fully vaccinated (Table 6). However, PHO enrolment was not a significant factor associated with being partially vs. fully vaccinated. However, being female (aOR = 1.13, 95% CI: 1.09–1.17) and having higher residential mobility (aOR = 1.05, 95% CI: 1.00–1.14) were associated with greater odds of being partially vaccinated. Those living with three or more (but not two) generations were more likely to be partially vaccinated (aOR = 1.08, 95% CI: 1.01–1.15). Living in a one‐person (vs. single family) household was associated with 2.14 times greater odds (95% CI: 1.86–2.45) of being partially vs. fully vaccinated, while living in multiperson households (vs. single family) was associated with 37% lower odds (aOR = 0.63, 95% CI: 0.55–0.73) of being partially (vs. fully) vaccinated. Additional findings, including unadjusted measures of association and logistic regression models of vaccination status among non‐Māori as of May 2023, are presented in Tables 4–6 and S3. Logistic regression models that included nonlinear (polynomial) specifications of age yielded results comparable to those from the primary linear models. Detailed results can be found in the Supporting Information (Table S4 and Figure S2).
TABLE 5.
Logistic regression models of vaccination status as of May 2023 among Māori (partial vaccination vs. no vaccination).
| Characteristics | OR (95% CI)a | aOR (95% CI)b |
|---|---|---|
| Sex | ||
| Male (ref) | — | — |
| Female | 0.975 (0.955–0.997) | 0.976 (0.943–1.011) |
| Age | ||
| Age 10 | 1.307 (1.297–1.316) | 1.436 (1.416–1.457) |
| Primary Health Organisation enrolment | ||
| No (ref) | — | — |
| Yes | 0.157 (0.150–0.164) | 0.219 (0.205–0.234) |
| Disability status | ||
| Nondisabled (ref) | — | — |
| Disabled | 1.848 (1.730–1.975) | 0.947 (0.876–1.023) |
| Residential mobility | ||
| 0–1 (ref) | — | — |
| 2–3 | 0.843 (0.820–0.866) | 1.062 (1.018–1.108) |
| 4+ | 0.849 (0.804–0.897) | 1.053 (0.965–1.149) |
| Child under five | ||
| No (ref) | — | — |
| Yes | 0.660 (0.642–0.680) | 1.082 (1.036–1.130) |
| Extended family | ||
| <3 generations (ref) | — | — |
| ≥3 generations | 1.012 (0.973–1.052) | 0.975 (0.910–1.045) |
| Household overcrowding | ||
| No (ref) | — | — |
| Yes | 1.060 (1.024–1.097) | 1.024 (0.972–1.078) |
| Housing quality | ||
| Never or sometimes damp or mould (ref) | — | — |
| Always | 1.140 (1.091–1.191) | 1.050 (0.991–1.113) |
| Household income | ||
| No income (ref) | — | — |
| $1–$15,000 | 1.016 (0.869–1.187) | 1.002 (0.802–1.253) |
| $15,001–$30,000 | 0.946 (0.817–1.096) | 1.065 (0.862–1.316) |
| $30,001–$50,000 | 0.930 (0.803–1.076) | 1.017 (0.823–1.255) |
| $50,001–$100,000 | 0.765 (0.663–0.883) | 0.951 (0.773–1.170) |
| $100,001+ | 0.625 (0.542–0.721) | 0.749 (0.609–0.922) |
| Household composition | ||
| One family (ref) | — | — |
| Two or more | 1.134 (1.085–1.1850) | 1.122 (1.038–1.212) |
| Multiperson | 3.202 (2.900–3.536) | 1.453 (1.255–1.682) |
| One person | 3.705 (3.391–4.049) | 0.927 (0.808–1.064) |
Note: Unvaccinated vs. partially vaccinated (N = 94,260).
OR = unadjusted odds ratio (unadjusted logistic regression model—each predictor run separately).
aOR = adjusted odds ratio (logistic regressions were estimated using generalised estimating equations with household as the clustering variable and robust standard errors).
4. Discussion
In this retrospective population study, we assessed COVID‐19 vaccine uptake among the Aotearoa New Zealand population from December 2020 to May 2023 and the independent variables associated with vaccination among the Māori population. We found lower vaccination rates among Māori. Based on the June 2020 APC, we estimated that 78,880 fewer Māori were fully vaccinated and 74,219 more Māori were unvaccinated than would have been the case had the Māori rates of vaccination and nonvaccination been the same as for the non‐Māori population. Findings from the unadjusted and adjusted models were directionally consistent.
We found that PHO enrolment was a crucial factor associated with whether Māori received at least one COVID‐19 vaccination. Several household factors, including overcrowding, housing quality, household income, household composition and living with extended family, were associated with vaccination status; for example, those living in lower‐quality and crowded housing (along with lower household incomes) had greater odds of being unvaccinated. Recent PHO enrolment significantly predicted COVID‐19 vaccination status. Among Māori, the likelihood of not being vaccinated (compared to being fully or partially vaccinated) was lower among individuals who had enrolled with a PHO in comparison to those who did not have recent enrolment. These findings indicate the utmost importance of ensuring that vaccination programmes are effective at reaching individuals who have not had recent contact or affiliation with a healthcare service provider. An examination of administrative data in 2019 showed that 6% of the population was not registered with a PHO, but this was higher among Māori at 9%, and overall enrolment rates declined from 95% in 2015 to 93% in 2018 but slightly increased to 94% in 2019 (Irurzun‐Lopez et al. 2021). Previous research has shown that the level of immunisation coverage depends on the interaction between individuals and healthcare providers, as well as the knowledge of healthcare professionals. Factors such as promoting a positive attitude towards vaccination and providing patients with current and well‐informed guidance have a significant influence on the choices made by the communities they serve (Nowlan et al. 2019; Te Whatu Ora Health New Zealand 2023).
In Aotearoa New Zealand, 24% (1,062,000) of the total population had a disability, compared to 20% in 2001 (Stats NZ 2014). People with a disability face more obstacles in achieving their goals than people without a disability and often experience lower living standards, increasing their likelihood of living in poverty (Murray and Loveless 2021). We found that people with a disability had 1.63 times greater odds of being unvaccinated against COVID‐19. This is further compounded by the fact that people with a disability in Aotearoa New Zealand experience inequities when accessing health services, including poorer health outcomes compared to people without a disability. Inequities in healthcare access and adverse health outcomes are further exacerbated among disabled Māori individuals. For instance, it has been observed that Māori people with a disability are twice as likely to be hospitalised or experience fatalities after testing positive for COVID‐19 (Satherley et al. 2023). These findings are consistent with previous research (Ryerson et al. 2021). In the United States, it has been observed that 15.2% of adults exhibit one or more documented functional disabilities. This population group is at a heightened risk of developing chronic conditions, including coronary heart disease, cancer, diabetes, obesity and hypertension when compared to individuals without disabilities (Dixon‐Ibarra and Horner‐Johnson 2014; Ryerson et al. 2021). This suggests that disability is likely to contribute to an increased risk of developing health problems, including COVID–19. A recent study involving 56,749 adults found that individuals with a disability were less likely than those without a disability to receive the COVID‐19 vaccine (Ryerson et al. 2021).
We found that greater residential mobility (having a higher number of addresses recorded between 2018 and 2020) was associated with marginally higher odds of being unvaccinated or partially vaccinated. Also, people in unstable living arrangements are more likely to experience negative COVID‐19 outcomes (Satherley et al. 2023). Recent studies have revealed a correlation between residential mobility and individuals’ vaccination patterns regarding infectious diseases. For instance, an examination of public health data demonstrated that residential mobility in the United States was associated with lower COVID‐19 vaccination rates, even after accounting for a wide range of cultural and sociodemographic factors (Zhang et al. 2023). These findings have been corroborated in a cross‐sectional study conducted in France in 2021 to determine COVID‐19 vaccine coverage and the factors influencing it among people living in precarious housing conditions. This study revealed that the standardised vaccination rate was 80%, which is significantly lower than the coverage observed in the general population. Furthermore, the uptake of vaccines among this group was delayed by approximately two months compared to the general population across all age groups and strata (Roederer et al. 2023).
The odds of being unvaccinated or only partially vaccinated decreased steadily as household income increased from the third‐highest (NZ$30,001 to NZ$50,000) to the highest (NZ$100,001+) bracket. Low vaccine uptake may possibly be attributed to the heightened financial hardship experienced during the pandemic, as well as the additional costs associated with travelling to vaccination centres. A 2021 web‐based cross‐sectional survey (n = 1284 adults aged 18 years and over) aimed to examine sociodemographic correlates of vaccine hesitancy and reasons for likelihood/hesitancy. The findings revealed that individuals with a household income lower than NZ$150,000 reported lower rates of vaccine acceptance (very likely to accept = 54.4%) compared to those with a household income higher than NZ$150,000 (very likely to accept = 70.2%). However, in multinomial regressions, annual household income did not emerge as a statistically significant predictor of COVID‐19 vaccine acceptance (Prickett et al. 2021). A recent report has indicated that there is a significant correlation between household income levels and COVID‐19 outcomes within the Māori community. Specifically, Māori in the lowest‐income households exhibited more unfavourable outcomes, such as an increased likelihood of testing positive, requiring hospitalisation or succumbing to severe COVID‐19 outcomes, compared to those with higher household incomes (Roederer et al. 2023). Data from the National Center for Immunization and Respiratory Diseases (U.S.). (2023) indicates that adults with incomes below the poverty line were vaccinated at a rate approximately half that of individuals with incomes exceeding $75,000 (Centers for Disease Control and Prevention 2023). A recent cross‐sectional survey conducted in the United States (n = 241,645,704 adults aged 18 years and older) found that as income levels decreased, the likelihood of receiving the COVID‐19 vaccine also decreased. This was particularly evident among individuals in the poor income group, who had a 55% lower chance of being vaccinated compared to other income groups (Ben‐Umeh and Kim 2024).
We have observed that Māori individuals residing in housing of inferior quality (characterised by persistent dampness or the presence of mould exceeding the size of an A4 sheet) exhibited a reduced likelihood of being vaccinated against COVID‐19. A study conducted in Aotearoa New Zealand in 2023 found that individuals residing in substandard housing were at a higher risk of contracting COVID‐19, requiring hospitalisation and experiencing mortality within a month after testing positive for the virus (Satherley et al. 2023). Across all our analyses, the presence of a child under five, as indicated by the age of the youngest child in a family or household, was associated with increased odds of being unvaccinated. This may be due to time constraints, caregiving responsibilities, concerns about potential side effects that could hinder their ability to care for their children, misinformation, reduced perception of risk and barriers to access. This finding is likely to have a significant impact on vaccine uptake, as our recent study reported that 9.7% of the Māori population comprised children aged 0–4 years (Mbinta et al. 2024).
5. Strengths, Limitations and Future Directions
To our knowledge, our study represents one of the first comprehensive population analyses of COVID‐19 vaccinations utilising the IDI. The IDI serves as the most extensive and reliable data source for COVID‐19 and household information, enabling our analysis to encompass significant portions of the population and various predictor variables of relevance. Our study presents a variety of policy‐amenable factors associated with COVID‐19 vaccination, providing a quantified understanding of how to address existing disparities.
Although data in the IDI can theoretically cover the whole Aotearoa New Zealand population, there is missing data across numerous variables that should be taken into account when interpreting the results. Variables sourced from the 2018 Census are particularly important to consider, as many have been impacted by the poor response rate to the Census. As the 2018 Census undercounted Māori in particular, missing data is most likely to affect analysis for this group. These issues, particularly for poorer quality variables, will create biases in the results, although their exact nature is difficult to assess. Extended family type and household composition are considered of poor and moderate quality, respectively, due to potential errors in misclassifying individuals to households/families. However, the error rates are higher in larger households and extended families. As Māori are more likely to live in larger households, errors in these variables will disproportionately affect Māori (2018 Census External Data Quality Panel 2020). For the present study, estimates of household and extended family effects are potentially underestimates for Māori in particular, as larger households (or broader extended families) may be misclassified as smaller.
An important aspect of our study to bear in mind is that we have provided a highly general assessment of a broad range of factors related to COVID‐19 vaccination status. While this study gives a preliminary indication of relevant factors, it does not capture the complexities in these associations. For example, some effects, such as that for age, are likely nonlinear, and some of the factors examined may interact with other factors, or their effects may be partially or fully mediated through other independent variables. For example, household income may be associated with vaccination status through an effect on household crowding, while the effect of residential mobility may be smaller at higher income levels. Moreover, our models do not account for all factors relevant to vaccination status, and the inclusion of these variables (many of which are likely not measured in the IDI) may change or eliminate independent variable associations seen here. Future research is required to better account for these complex associations.
Residential mobility relies on individuals providing accurate information to government agencies, but frequent movers may not consistently update their details, leading to potential gaps in the data used for these analyses. Underreporting of these factors in the existing data means our results are likely to underestimate their impact on vaccine uptake. Data on housing quality, household crowding, composition, income, and children under five may have changed since the 2018 census, leading to increased uncertainty in the effect estimates for these variables. These results are based on the use of historical time frames (i.e. up to May 2023). This may lead to discrepancies when compared to alternative analyses or official data (e.g. those that encompass the most up‐to‐date COVID‐19 data).
Another limitation of this study relates to the classification of household income bands. We set NZ$100,001 as the upper threshold for higher‐income households and did not adjust income levels for the number of dependents. While this threshold is common in some contexts, it may not fully capture the economic realities of different household compositions. For instance, a single adult earning NZ$100,001 is in a substantially different financial situation compared to a household where the same income is shared among multiple adults and children. Additionally, other factors, such as level of education, chronic diseases, previous vaccinations, healthcare‐seeking behaviour, region of residence, vaccination selection, previous COVID‐19 infection and COVID‐19‐related hospitalisation or mortality in the household, may influence vaccine uptake, but were not included in the current analysis.
In this study, we used logistic regression models to examine the association between individual and household‐level factors and COVID‐19 vaccination status. While ORs are a standard measure in logistic regression and are widely used in epidemiological research, we acknowledge that they may overstate the magnitude of associations when the outcome is prevalent, as is the case with COVID‐19 vaccination. The ORs tend to diverge from relative risks as the prevalence of the outcome increases, particularly when the initial risk exceeds 20% (Davies et al. 1998).
These findings highlight disparities in COVID‐19 vaccination rates between Māori and non‐Māori in Aotearoa New Zealand. Factors such as overcrowded and poor‐quality housing, lower household income and disability status contribute to lower vaccination uptake. PHO enrolment is a significant predictor of vaccination status, highlighting the importance of ongoing engagement with healthcare services. It is important to note that we examined vaccination coverage among non‐Māori as a broad population group, but that vaccination rates will differ between non‐Māori ethnic groups, and equity gaps may be larger or smaller for different, specific comparison ethnic groups. These findings emphasise the necessity of targeted public health interventions to address these disparities and ensure equitable access to vaccination and healthcare resources for vulnerable populations. Future research should investigate the mechanisms underlying disparities in vaccination rates, examine trends for ordinal variables and evaluate targeted interventions aimed at enhancing vaccine uptake in the Māori population. Longitudinal studies are needed to assess the impact of policy changes and public health initiatives over time.
Supporting Information
Additional supporting information can be found online in the Supporting Information section. Supporting Fig. S1: Directed Acyclic Graph of factors that influence vaccine uptake. Supporting Fig. S2: Predicted probability of COVID‐19 vaccination status across age among Māori. Supporting Table S1: Conceptual directed edge index. Supporting Table S2: Counts for predictor levels for the samples used for each regression. Supporting Table S3: Logistic regression models of vaccination status as of May 2023 among non‐Māori. Supporting Table S4: Logistic regression models of vaccination status as of May 2023 among Māori.
Author Contributions
C.R.S conceived the study. All authors contributed to the study design. N.S. cleaned, analysed and verified the underlying data with input from all authors. J.F.M., N.S., A.A.S., J.S., A.S.‐T., N.F. and C.R.S. interpreted the data. J.F.M. wrote the first draft of the paper with input from A.A.S., J.S., N.F. and C.R.S. All authors revised the manuscript for important intellectual content and approved the final version. C.R.S. supervised the entire work and is the guarantor.
Acknowledgments
Open access publishing facilitated by University of Otago, as part of the Wiley ‐ University of Otago agreement via the Council of Australian University Librarians.
Funding
This work was supported by the Ministry of Health (PROP–053).
Disclosure
C. R. Simpson, A. A. Sporle, J. Shephard, A. Su‘a‐Tavila, B. P. Nguyen, and N. French report support for the present manuscript from the Ministry of Health (NZ). C.R. Simpson reports funding from MBIE (NZ), HRC (NZ), Ministry of Health (NZ), MRC (UK), and CSO (UK); Leadership or fiduciary role in other board, society, committee, or advocacy group, paid or unpaid with the NZ Government Data Ethics Advisory Group as the Chair; outside the submitted work. N French reports grants and consulting fees from the Ministry of Business Innovation and Employment, the Ministry for Primary Industries, the Ministry of Health, the NZ Health Research Council, Te Niwha, the UK Fleming Fund, Operational Solutions for Primary Industries, the Institute of Environmental Science and Research, and International Commission on Microbial Specifications for Foods; outside the submitted work. B Nguyen reports grants from the NZ Ministry of Business Innovation and Employment and NZ Health Research Council; outside the submitted work. The remaining authors declare no competing interests.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supplementary Material
Contributor Information
James F. Mbinta, Email: james.mbinta@vuw.ac.nz.
Colin R. Simpson, Email: colin.simpson@vuw.ac.nz.
Data Availability Statement
Access to the data used in this manuscript was provided by Stats NZ under conditions designed to give effect to the security and confidentiality provisions of the Data and Statistics Act 2022. The results presented in this manuscript are the work of the authors, not Stats NZ or individual data suppliers. These results are not official statistics. They have been created for research purposes from the Integrated Data Infrastructure (IDI), which is carefully managed by Stats NZ. For more information about the IDI, please visit https://www.stats.govt.nz/integrated‐data/.
References
- 2018 Census External Data Quality Panel . 2020. Final report of the 2018 Census External Data Quality Panel.
- Barrett, M. 2025. “ggdag: Analyze and Create Elegant Directed Acyclic Graphs.” R package version 0.2.13.9000. https://github.com/r-causal/ggdag.
- Barron, E. , Bakhai C., Kar P., et al. 2020. “Associations of Type 1 and Type 2 diabetes with COVID‐19‐Related Mortality in England: A Whole‐Population Study.” The Lancet Diabetes & Endocrinology 8, no. 10: 813–822. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ben‐Umeh, K. C. , and Kim J.. 2024. “Income Disparities in COVID‐19 Vaccine and Booster Uptake in the United States: An Analysis of Cross‐Sectional Data from the Medical Expenditure Panel Survey.” PLoS One 19, no. 2: e0298825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Centers for Disease Control and Prevention . 2023. “COVID‐19 Vaccine Uptake and CDC's Commitment to Vaccine Equity.” https://www.cdc.gov/respiratory‐viruses/whats‐new/vaccine‐equity.html.
- COVID‐19 Group . 2023. “COVID‐19: System Planning. Department of the Prime Minister and Cabinet.” https://www.dpmc.govt.nz/publications/covid‐19‐system‐planning.
- Davies, H. T. O. , Crombie I. K., and Tavakoli M.. 1998. “When Can Odds Ratios Mislead?” BMJ 316, no. 7136: 989–991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dixon‐Ibarra, A. , and Horner‐Johnson W.. 2014. “Disability Status as an Antecedent to Chronic Conditions: National Health Interview Survey, 2006–2012.” Preventing Chronic Disease 11: 130251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ferguson, K. D. , McCann M., Katikireddi S. V., et al. 2020. “Evidence Synthesis for Constructing Directed Acyclic Graphs (ESC‐DAGs): A Novel and Systematic Method for Building Directed Acyclic Graphs.” International Journal of Epidemiology 49, no. 1: 322–329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Goodyear‐Smith, F. , and Ashton T.. 2019. “New Zealand Health System: Universalism Struggles with Persisting Inequities.” The Lancet 394, no. 10196: 432–442. [DOI] [PubMed] [Google Scholar]
- Hanley, J. A. , Negassa A., Edwardes MDd., and Forrester J. E.. 2003. “Statistical Analysis of Correlated Data Using Generalized Estimating Equations: An Orientation.” American Journal of Epidemiology 157, no. 4: 364–375. [DOI] [PubMed] [Google Scholar]
- Health Quality & Safety Commission . 2015. “A Window on the Quality of Aotearoa New Zealand's Health Care.” Wellington. https://www.hqsc.govt.nz/assets/window-on-quality-of-NZ-health-care-Nov-2015.pdf?hash=31b56b9eb1657df288bdc4ec2ca0276ea415bfbd.
- Hitchon, E. G. , Eggleston K., Mulder R., Porter R., and Douglas K.. 2024. “The Aotearoa New Zealand Doctor Shortage: Current Context and Strategies for Retention.” New Zealand Medical Journal 137, no. 1592: 9–13. [DOI] [PubMed] [Google Scholar]
- Irurzun‐Lopez, M. , Jeffreys M., and Cumming J.. 2021. “The Enrolment Gap: Who Is Not Enrolling with Primary Health Organizations in Aotearoa New Zealand and What Are the Implications? An Exploration of 2015–2019 Administrative Data.” International Journal for Equity in Health 20, no. 1: 93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jefferies, S. , French N., Gilkison C., et al. 2020. “COVID‐19 in New Zealand and the Impact of the National Response: A Descriptive Epidemiological Study.” The Lancet Public Health 5, no. 11: e612–e623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mbinta, J. F. , Sporle A. A., Sheppard J., et al. 2024. “Effectiveness of COVID‐19 Vaccines against Hospitalisation, Death and Infection over Time in Aotearoa New Zealand: A Retrospective Cohort Study.” The New Zealand Medical Journal 137, no. 1602: 65–101. [DOI] [PubMed] [Google Scholar]
- Ministry of Health . 2021. “Follow-up of Performance Audit and Request for Progress Report: Preparations for the Nationwide Roll-Out of the COVID-19 Vaccine.” https://www.health.govt.nz/system/files/2022-05/oag-cvip-response-letter-15nov21-redacted.pdf.
- Ministry of Health . 2024a. “Immunisation Handbook 2024, version 1. New Zealand: Ministry of Health.” Accessed April 2024. https://www.tewhatuora.govt.nz/for‐health‐professionals/clinical‐guidance/immunisation‐handbook.
- Ministry of Health . 2024b. “COVID-19: Protecting Aotearoa New Zealand.” https://nationdatesnz.org/wp-content/uploads/2024/10/MOH-2024d.pdf.
- Murray, S. , and Loveless R.. 2021. “Disability: The Living Standards Framework and Wellbeing in New Zealand.” Policy Quarterly 17, no. 1: 64–71. [Google Scholar]
- National Center for Immunization and Respiratory Diseases (U.S.) . 2023. COVID-19 Vaccine Uptake and CDC's Commitment to Vaccine Equity. https://stacks.cdc.gov/view/cdc/135797.
- Nowlan, M. , Willing E., and Turner N.. 2019. “Influences and Policies that Affect Immunisation Coverage‐a Summary Review of Literature.” The New Zealand Medical Journal 132, no. 1501: 79–88. [PubMed] [Google Scholar]
- Prickett, K. C. , Habibi H., and Carr P. A.. 2021. “COVID‐19 Vaccine Hesitancy and Acceptance in a Cohort of Diverse New Zealanders.” The Lancet Regional Health ‐ Western Pacific 14: 100241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reddy, R. K. , Charles W. N., Sklavounos A., Dutt A., Seed P. T., and Khajuria A.. 2021. “The Effect of Smoking on COVID‐19 Severity: A Systematic Review and Meta‐Analysis.” Journal of Medical Virology 93, no. 2: 1045–1056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rodrigues, C. M. C. , and Plotkin S. A.. 2020. “Impact of Vaccines; Health, Economic and Social Perspectives [Review].” Frontiers in Microbiology 11, no. 1526: 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Roederer, T. , Mollo B., Vincent C., et al. 2023. “Estimating COVID‐19 Vaccine Uptake and Its Drivers among Migrants, Homeless and Precariously Housed People in France.” Communications Medicine 3, no. 1: 30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ryerson, A. B. , Rice C. E., Hung M. C., et al. 2021. “Disparities in COVID‐19 Vaccination Status, Intent, and Perceived Access for Noninstitutionalized Adults, by Disability Status ‐ National Immunization Survey Adult COVID Module, United States, May 30‐June 26, 2021.” MMWR Morbidity and Mortality Weekly Report 70, no. 39: 1365–1371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Satherley, N. , Diamond T., and Sporle A.. 2023. “More than Just Living in a Deprived Area: An Equity‐Focused Analysis of Policy Amenable Factors Associated with Māori COVID‐19 Outcomes. INZight Analytics Ltd. The National Hauora Coalition. Tāmaki Makaurau.” https://inzight.co.nz/projects/maori-covid-19-outcomes.
- Sonder, G. , Grey C., Mischewski B., Ryan D., and Ryan D.. 2022. “Pacific COVID‐19 Intelligence Gathering and Analysis Project: Evaluation of the August 2020 outbreak.” https://www.health.govt.nz/information-releases/pacific-covid-19-intelligence-gathering-and-analysis-project-evaluation-of-the-august-2020-outbreak.
- Stats, N. Z. 2012. Crowded Housing in New Zealand 1986–2006. Statistics New Zealand. [Google Scholar]
- Stats, N. Z. 2014. “One in Four New Zealanders Identified as Disabled.” https://www.stats.govt.nz/news/one-in-four-new-zealanders-identified-as-disabled/.
- Stats, N. Z. 2022. “Experimental Administrative Population Census: Data Sources, Methods, and Quality (second Iteration).” https://www.stats.govt.nz/research/experimental-administrative-population-census-data-sources-methods-and-quality-second-iteration/.
- Stats, N. Z. 2024. “Data in the IDI. New Zealand Government.” Accessed May 2024. https://www.stats.govt.nz/integrated‐data/integrated‐data‐infrastructure/data‐in‐the‐idi/.
- Te Whatu Ora Health New Zealand . 2023. “Initial Priorities for the National Immunisation Programme in Aotearoa.” https://www.tewhatuora.govt.nz/publications/initial-priorities-for-the-national-immunisation-programme-in-aotearoa.
- Te Whatu Ora Health New Zealand . 2024a. “COVID‐19 vaccine data.” Accessed February 12, 2024. https://www.tewhatuora.govt.nz/our‐health‐system/data‐and‐statistics/covid‐vaccine‐data/.
- Te Whatu Ora Health New Zealand 2024b. “COVID‐19: Case demographics.” Accessed February 12, 2024. https://www.tewhatuora.govt.nz/our‐health‐system/data‐and‐statistics/covid‐19‐data/covid‐19‐case‐demographics/#details‐of‐covid‐19‐deaths.
- Te Whatu Ora Health New Zealand . 2025. “Getting COVID‐19 Vaccines.” Accessed January 30, 2025. https://info.health.nz/immunisations/vaccines‐aotearoa/covid‐19‐vaccines/getting‐covid‐19‐vaccines.
- Williamson, E. J. , Walker A. J., Bhaskaran K., et al. 2020. “Factors Associated with COVID‐19‐Related Death Using OpenSAFELY.” Nature 584, no. 7821: 430–436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- World Health Organization . 2023. “Coronavirus Disease (COVID‐19).” Accessed October 10, 2023. https://www.who.int/emergencies/diseases/novel‐coronavirus‐2019.
- Zhang, N. , Jiang T., Zhang Y., and Zhao G.. 2023. “Residential Mobility Predicts COVID‐19 and Seasonal Flu Vaccination Behaviors in the United States [Original Research].” Frontiers in Public Health 10: 1064962. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material
Data Availability Statement
Access to the data used in this manuscript was provided by Stats NZ under conditions designed to give effect to the security and confidentiality provisions of the Data and Statistics Act 2022. The results presented in this manuscript are the work of the authors, not Stats NZ or individual data suppliers. These results are not official statistics. They have been created for research purposes from the Integrated Data Infrastructure (IDI), which is carefully managed by Stats NZ. For more information about the IDI, please visit https://www.stats.govt.nz/integrated‐data/.