View Article PDF

Being counted is an acknowledgement of both existence and value. It means that one matters. It is the hallmark of Treaty promise (Orange 1987:257). History will judge our commitment to the Treaty in part by our ability to ensure that people are counted, that disparities are acknowledged and appropriate policies are put in place, especially those which eliminate disparities between Māori and non-Māori, a solemn commitment by the Treaty of Waitangi.”[[1]]

Ethnicity data matter. Longstanding and significant inequities exist between Māori and non-Māori across most health indicators including life expectancy, health determinants, health outcomes and healthcare. These inequities are rooted in processes of colonisation and colonialism, supported by a system of racism.[[2]] They are a breach of the Te Tiriti o Waitangi and Māori Indigenous rights.[[2]] Furthermore, the lack of urgency towards their elimination, particularly by Crown agencies, further reflects the disregard with which Māori and Māori health are held.

Ethnicity data have been collected for decades in health and population datasets; however, the quality, consistency and completeness continue to be problematic, particularly for Māori.[[3]] This has often been reflected in the systematic under-counting of Māori in datasets including the National Health Index (NHI),[[4,5]] in mortality datasets,[[6,7]] and others.[[8–13]] This then impacts on the ability to estimate health prevalence, rates, inequities and progress (or not) over time. Accurate ethnicity data are also critical to correctly identify individuals eligible for healthcare interventions. In New Zealand, an individual’s ethnicity determines the age of their eligibility for cardiovascular risk assessment[[14]] and diabetes screening,[[15]] sore throat management,[[16]] eligibility for diabetes treatments,[[17]] and the proposed age extension to the Bowel Cancer Screening Programme.[[18]]

Māori health advocates have long argued for high quality ethnicity data in health as a Māori health right;[[1,19,20]] Māori have a “right to be counted”[[2]] and “Māori have the right to monitor the Crown and to evaluate Crown action and inaction”.[[21]] While protocols for the standardised collection of ethnicity data in the health and disability sector[[22]] have existed for nearly 20 years, their implementation seems yet to be fully realised, with ongoing concerns regarding the under-counting of Māori in health and disability sector data. This study examines ethnicity data quality and its implications for Māori health and inequities. We focus on NHI ethnicity as a consistent variable across a range of health datasets. The NHI provides a unique identifier for each person who receives healthcare with associated identifying information including ethnicity.[[4]] NHI ethnicity is the most up-to-date record of an individual’s ethnicity based on their most recent health encounters. We seek to examine in detail NHI ethnicity data quality (particularly under-count) for Māori by comparing aggregate level data for Māori and non-Māori across a range of population and health datasets, and by comparing Māori and non-Māori NHI ethnicity among individuals enrolled with a primary health organisation (PHO) with their linked self-identified ethnicity from the 2018 Census. We also examine differences by age and gender.

Methods

We are guided by kaupapa Māori research methods, particularly kaupapa Māori epidemiology that includes: 1) the right to monitor the Crown; 2) the right to be counted; 3) the right to have a powerful voice; and 4) the right to name racism and colonialism as fundamental causes of inequities for Māori.[[23]] Our kaupapa Māori epidemiological approach supports the use of a Māori/non-Māori analysis as a reflection of Te Tiriti.

Study design

This is a descriptive study that compares the recording of Māori and non-Māori ethnicity across a range of datasets and in two main ways:

1. Population aggregate analysis

2. Individual linked analysis

In these analyses, Māori are defined as anyone who has their ethnicity recorded as Māori, either alone, or in combination with another ethnic group(s).[[22]] Non-Māori are all other people in the dataset (including those with missing ethnicity). Gender could only be examined by the binary categories of male and female due to dataset limitations. The study was approved by the University of Otago Human Ethics Committee (HD 20/079).

1. Population aggregate analysis

At an aggregate level, total population data are compared for Māori and non-Māori across a range of population and health datasets to compare both counts and proportions of Māori and non-Māori. The datasets include the 2018 Estimated Resident Population (ERP); 2018 Census Usually Resident Population (Census); 2018 Health Service User (HSU) population; and 2018 Primary Health Organisation (PHO) population. Appendix 1 provides further details on the datasets. All data are estimated at a single point in time and as close as possible to the same point in time (30 June 2018) for comparability.

The two health datasets (HSU and PHO) are compared with official population estimates (ERP) to examine potential differential under-representation of Māori compared to non-Māori in health data (either from differential access to care or under-counting of Māori). We compare the total numbers and proportions of Māori and non-Māori in each dataset overall, and by age and gender. This aggregate analysis provides an indication of potential under-counting of Māori if proportions of Māori in health data are lower than non-Māori compared to official population data. However, the contribution of differential access to care cannot be teased out from potential under-counting.

2. Individual linked analysis

In order to examine the misclassification of ethnicity and potential systematic under-counting of Māori on NHI data, we undertook a linked analysis at the level of individuals. To be eligible for inclusion, people had to be enrolled in a PHO (as at 1 April 2018), have self-completed the ethnicity question in the 2018 Census, linked in the NZ Integrated Data Infrastructure (IDI) spine and to Ministry of Health (MoH) NHI data.

Analyses were undertaken in the Statistics NZ (Stats NZ) datalab using the September 2021 IDI Refresh (available from 27 October 2021). Data were linked between the IDI spine and MoH.[[24]] 2018 Census data were restricted to only those who had self-identified their ethnicity in this Census i.e., ethnicity supplemented from other administrative data were not used (Appendix 1). Ethnicity data from the 2018 Census were considered the standard against which PHO NHI ethnicity data were compared. This is in line with the Ethnicity Data Protocols for the Health Sector and the Census, whereby a standard question should be used and people should self-identify their own ethnicity whenever possible, with guidance on instances where this may not be possible, e.g., young children, people who are incapacitated.[[ 22]] Those with missing NHI ethnicity (75,243 individuals; 2%) were categorised as non-Māori.

Among linked individuals, two main measures were used to examine data quality. First, we calculated the proportion whose ethnicity was the same in both datasets using the Māori/non-Māori population denominators from the Census (proportion matched). This provides a measure of misclassification. Secondly, the net under-count of Māori on the NHI was estimated by 1 minus (the number of people classified as Māori on the NHI as a percentage of those classified as Māori on the Census). This indicates the extent to which the misclassification is differential.

Results

View Tables & Figures.

Population aggregate results

Table 1 shows the population counts across population and health datasets with numbers (and proportions) for Māori and non-Māori. Of the four datasets examined, the ERP has the most people as well as the highest proportion of Māori (16.7%). The total numbers in the HSU dataset are closest to the ERP, although lower for Māori and (correspondingly) slightly higher for non-Māori. The proportion of Māori in the HSU is lower than the ERP at 15.5%. The PHO dataset has the least people and the lowest proportion of Māori (14.6%).

Figure 1 shows the number of people recorded as Māori and non-Māori in the datasets by 5-year age group. The ERP has the highest number of Māori across almost all ages. The exception is the youngest age group where ERP and HSU numbers are similar. PHO data are lowest for Māori across all ages. Differences between datasets are less obvious with increasing age.

For non-Māori, PHO and Census data have lower counts across most ages compared to ERP. For PHO data, this is more obvious for adolescents and young adults. The HSU data has higher numbers of non-Māori than ERP in children, lower numbers in young adults and similar numbers among adults and older people.

Compared to the ERP, there are fewer Māori males and females in every age group in the Census, HSU (except 0–4 years) and PHO datasets and the percentage differences are disproportionately larger than for non-Māori (Figure 2). The percentage differences for males tend to be disproportionately larger than for females in the datasets (Census, HSU, PHO) compared to the ERP, for both Māori and non-Māori. Within the HSU data, Māori males are under-represented across most age groups. For non-Māori, there is under-representation of young males and over-representation of females in multiple age groups.

Individual linked results

3,498,789 people on the PHO dataset were linked to a 2018 Census record. This equates to 78% linkage of people on the PHO dataset. Linkage varied by ethnic grouping (based on NHI ethnicity on the PHO dataset), with 64% linkage for Māori and 80% for non-Māori. Within the PHO linked sample, 491,928 (14.1%) of the population were Māori (using Census ethnicity data), compared to 414,846 (11.9%) using NHI data (Table 2). Both of these proportions are lower than the total population percentage of Māori in the ERP of 16.7% (Table 1).

Of the 491,928 individuals who identified as Māori on the Census, 386,976 (78.7%) were also recorded as Māori on the NHI (Table 2). Of the 3 million individuals categorised as non-Māori ethnicity from the Census, 2,978,991 (99.1%) were categorised as non-Māori on the NHI. The high match of NHI and Census ethnicity for non-Māori was seen across all age groups.

For Māori (on the Census) there is some variation by age, with the youngest age group (0–4 years) having a higher match (approx. 90%) and a lower level of match in those aged 20–24 years (Figure 3). The level of match is lower for Māori males than Māori females, and it is more apparent in adults over 20 years old where matching for Māori males remains consistently low.

Figure 4 examines the net under-counting of Māori ethnicity on the NHI by age and gender. Net under-count tells us overall how much the NHI data under-counts Māori compared to the 2018 Census (assuming the Census is the “correct” classification). The net under-count captures the proportion of Māori missed from the NHI (i.e., those identified as Māori on the Census but not on NHI) minus non-Māori incorrectly classified as Māori in NHI (those identified as non-Māori on the Census but Māori on NHI).

The net under-count in the NHI for Māori is 15.7% overall. The net under-count for 0–4-year-olds  is less than 5%, but increases to between 13–23% for all other age groups. When this is examined separately by gender, the net under-count for Māori males is higher than Māori females from 20 years on, and is greater than 20% in most age groups (20–59 years). For Māori women, there is more variation in net under-count by age, with peaks in the under-count in the age groups 20–24 years (22.7% under-count) 40–44 years (18.4% under-count), and 80+ years (19.2% under-count).

Discussion

High quality ethnicity data are fundamental for understanding and monitoring Māori health and health inequities as well as in the provision of targeted services and interventions that are responsive to Māori aspirations and needs. This study shows that Māori are under-represented in health datasets (HSU and PHO), compared to official population numbers (ERP), and are systematically under-counted in NHI ethnicity data. At a population aggregate level, all four datasets had different total population and Māori numbers, reflecting differences in their composition. Additionally, there were differences in the proportion of Māori in each dataset ranging from 16.7% on the ERP to 14.6% on the PHO dataset. There were lower numbers of Māori in the HSU and PHO datasets compared to the ERP across almost all ages. This under-representation of Māori in health data is likely to be due to under-counting of Māori in health datasets as well as ongoing differential access to healthcare. This was demonstrated in the linked analysis of individuals whereby NHI ethnicity (from PHO data) was more likely to misclassify Māori as non-Māori (than non-Māori as Māori), resulting in a 16% net under-count of Māori on the NHI (with variation by age and gender).

It is clear that the health and disability sector continues to fail in its responsibility to achieve expected standards of ethnicity data collection,[[22]] with important implications for our ability to monitor Māori health and equity, and to target health services. In the monitoring of Māori health and equity, the under-counting of Māori in the NHI can lead to a numerator/denominator bias when rates using population data from other sources such as ERP are used, as is common practice. This will lead to underestimation of the Māori rate for any health variable of interest. In contrast, the impact of any misclassification for non-Māori will be negligible because of the much larger population. This can make inequities look better or worse than they really are, depending on the health variable concerned.

The HSU population is being used as one method to try and minimise the numerator/denominator bias by using it as the population denominator.[[4]] The 2020 HSU has been used recently in the estimation of COVID-19 vaccination coverage.[[25]] However, this may allow the inclusion of people in the numerator that may not be in the denominator (e.g., people who have been vaccinated but had otherwise not had a health interaction in the period of the HSU). This can potentially overestimate rates. Our study shows that the HSU should not be used to estimate actual numbers of Māori eligible for services (e.g., the number of Māori to be vaccinated), as it significantly under-represents Māori.

Other methods have also been used to mitigate the numerator/denominator bias for Māori.[[26]] These include the calculation of ratios to adjust for Māori under-count[[6,10,12]] that use similar methods as here to estimate the net under-counting of Māori. More recently, administrative datasets are being linked to create new populations in efforts to improve population estimates by ethnicity, e.g., the experimental Administrative Population Census derived from data in the IDI.[[27]] However, these still rely on the underlying quality of ethnicity data in the linked sources. It is important that Māori have the opportunity to self-identify their ethnicity. This is in line with Indigenous rights, as an expression of rangatiratanga and Māori rights to determine our own identities.[[20]]

There are implications with using ethnicity data to determine an individual’s eligibility for health and disability services. A number of health services already prioritise Māori in the assessment of risk or service delivery.[[14–18]] Using the example of bowel cancer screening, where a lower age (from 50 years) of invitation for Māori has been proposed, a large number of Māori who are eligible would be missed if invitations to screening drew on NHI ethnicity. A smaller number of non-Māori would also be mistakenly invited.

The under-counting of Māori in health data is not new and has been examined in various settings over time. Particularly relevant to our study are other analyses that have linked individual health data with Census data. Stats NZ linked 2013 Census data to a combined NHI dataset and found Māori were under-counted by 21%.[[5]] More recently, the MoH has undertaken similar analyses linking people from the 2018 Census to 2019 NHI data and shown a 16% net under-count of Māori in the NHI compared to Census.[[4]] Our analysis also shows a 16% net under-count of Māori on the NHI from the PHO database compared to individuals’ corresponding Census ethnicity.

There are a number of limitations that should be considered in the interpretation of our findings. Firstly, our analysis of linked data is a subset of Māori in the 2018 Census and 2018 PHO datasets with lower linkage for Māori (based on NHI ethnicity). We cannot determine if the level of mismatch and under-counting of those included in this analysis is the same as those not included. While we examined data by gender and age, we were unable to explore ethnicity data for other important groups, for example those who are gender diverse and/or tāngata whaikaha. Secondly, it is possible that people self-report their ethnicity differently, in different settings, and that they can change their ethnicity.[[28]] However, there is also evidence from health settings that ethnicity data protocols are not always adhered to,[[29]] e.g., guessing someone’s ethnicity based on name or appearance, disagreeing with patient’s self-identified ethnicity, non-standardised forms and data recording. In addition, there is evidence of improvements in ethnicity data collection in some health-related datasets, e.g., mortality,[[6,7]] cancer registrations,[[10]] and examples where there is no (or minor) net under-count of Māori.[[9,30–32]] These findings would suggest that (non)adherence to ethnicity data protocols are the likely drivers of the systematic under-counting of Māori. This is a complex issue to tease out in detail, and we have chosen to focus on implications for Māori using a non-Māori comparator. However, we acknowledge that the non-Māori group is made up of multiple ethnicities combined, and that the quality of ethnicity data for specific ethnic groups, and the mismatch of ethnicity within this large grouping are not examined. In particular, similar data issues and implications are likely to be present for Pacific peoples.[[4,5]] We have also not looked at all potential measures of data quality here such as multiple ethnicities, missing data, and consistency with health and disability sector standards.[[5,33]]

The ongoing systematic under-counting of Māori in health data is a reflection that Māori are not valued and that Māori health and the elimination of inequities are not the priorities they are claimed to be. The ongoing unjust and inequitable healthcare experiences and outcomes faced by Māori require the whole of the health and disability sector to commit to the collection of high-quality ethnicity data, in order to understand inequities; to deliver equitable services to Māori; to act to address inequity; and to monitor progress on eliminating inequities. This commitment needs to come from the highest levels and be attached to accountability mechanisms and ongoing quality assessments.[[34]] Quality ethnicity data are fundamental to the elimination of Māori health inequities. The under-counting of Māori remains a breach of Te Tiriti o Waitangi and Indigenous rights. The ongoing acceptance of poor-quality ethnicity data for Māori and the inadequate progress towards high quality ethnicity data in the health and disability sector is itself evidence of racism as “inaction in the face of need”.[[35]]

View Appendices.

Summary

Abstract

Aim

To examine ethnicity data quality; in particular, the representation and potential under-counting of Māori in health and disability sector data, as well as implications for inequities.

Method

Māori and non-Māori ethnicity data are analysed at: 1) a population aggregate level across multiple 2018 datasets (Estimated Resident Population, Census Usually Resident Population, Health Service User (HSU) population and Primary Health Organisation (PHO) enrolments); and 2) an individual level for those linked in PHO and 2018 Census datasets. Ethnicity is drawn from the National Health Index (NHI) in health datasets and variations by age and gender are explored.

Results

Aggregate analyses show that Māori are considerably under-represented in HSU and PHO data. In linked analysis Māori were under-counted on the NHI by 16%. Under-representation in data and under-counting occur across both genders but are more pronounced for Māori men with variations by age.

Conclusion

High quality ethnicity data are fundamental for understanding and monitoring Māori health and health inequities as well as in the provision of targeted services and interventions that are responsive to Māori aspirations and needs. The continued under-counting of Māori in health and disability sector data is a breach of Te Tiriti o Waitangi and must be addressed with urgency.

Author Information

Ricci Harris: Associate Professor/Public Health Physician, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Sarah-Jane Paine: Senior Lecturer, Te Kupenga Hauora Māori, University of Auckland, New Zealand. June Atkinson: Senior Data Analyst, Department of Public Health, University of Otago, Wellington, New Zealand. Bridget Robson: Associate Professor and Director of Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Paula Toko King: Public Health Physician/Senior Research Fellow, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Jennifer Randle: Public Health Medicine Registrar, Department of Public Health, University of Otago, Wellington, New Zealand. Anja Mizdrak: Senior Research Fellow, Department of Public Health, University of Otago, Wellington, New Zealand. Melissa McLeod: Public Health Physician/Senior Lecturer, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand.

Acknowledgements

The study was funded in part by the grant from the Health Research Council of New Zealand (HRC 20/152). We would like to thank Associate Professor Donna Cormack for comments on a draft of the paper. Data in the paper were provided from Stats New Zealand and the New Zealand Ministry of Health. The Ministry of Health retains intellectual property rights to the data they provided. Stats NZ disclaimer regarding IDI data: The results in this paper are not official statistics, they have been created for research purposes from the Integrated Data Infrastructure (IDI) managed by Statistics New Zealand. The opinions, findings, recommendations and conclusions expressed in this paper are those of the author(s) not Statistics NZ or Ministry of Health. Access to the anonymised data used in this study was provided by Statistics NZ in accordance with security and confidentiality provisions of the Statistics Act 1975. Only people authorised by the Statistics Act 1975 are allowed to see data about a particular person, household, business or organisation and the results in this paper have been confidentialised to protect these groups from identification. Careful consideration has been given to the privacy, security and confidentiality issues associated with using administrative and survey data in the IDI. Further detail can be found in the Privacy impact assessment for the Integrated Data Infrastructure available from www.stats.govt.nz.

Correspondence

Associate Professor Ricci Harris: Te Rōpū Rangahau Hauora a Eru Pōmare, Department of Public Health, University of Otago, Wellington PO Box 7343, Wellington 6242, New Zealand.

Correspondence Email

Ricci.harris@otago.ac.nz

Competing Interests

The authors have no conflicts of interest or financial disclosures.

1) Te Rōpū Rangahau Hauora a Eru Pōmare. Counting for nothing: understanding the issues in monitoring disparities in health. Soc Policy J N Z. 2000;14:1-16.

2) Reid P, Cormack D, Paine SJ. Colonial histories, racism and health—The experience of Māori and Indigenous peoples. Public Health. 2019 Jul;172:119-24.

3) Cormack D, McLeod M. Improving and maintaining quality in ethnicity data collections in the health and disability sector [Internet]. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2010 [cited 2022 Mar 17]. Available from: https://www.otago.ac.nz/wellington/otago600098.pdf.

4) Cleary L. Using ethnicity data in Health Statistics. Wellington: Ministry of Health; 2021. Report No.: 1.1.

5) Reid G, Bycroft C, Gleisner F. Comparison of ethnicity information in administrative data and the census. Wellington: Stats NZ; 2016.

6) Ajwani S, Blakely T, Robson B, Atkinson J, Kiro C. Unlocking the numerator-denominator bias III: adjustment ratios by ethnicity for 1981-1999 mortality data. The New Zealand Census-Mortality Study. N Z Med J. 2003;116(1175):U456.

7) Tan L, Blakely T, Atkinson J. Ethnic counts on mortality and census data 2001-06: New Zealand census-mortality study update. N Z Med J. 2010;123(1320):37-44.

8) Bramley D, Latimer S. The accuracy of ethnicity data in primary care. N Z Med J. 2007 Oct 26;120(1264):U2779.

9) Scott N, Clark H, Kool B, Ameratunga S, Christey G, Cormack D. Audit of ethnicity data in the Waikato Hospital Patient Management System and Trauma Registry: pilot of the Hospital Ethnicity Data Audit Toolkit. N Z Med J. 2018 Oct 5;131(1483):21-9.

10) Shaw C, Atkinson J, Blakely T. (Mis)classification of ethnicity on the New Zealand Cancer Registry: 1981-2004. N Z Med J. 2009 May 8;122(1294):10-22.

11) Swan J, Lillis S, Simmons D. Investigating the accuracy of ethnicity data in New Zealand hospital records: still room for improvement. N Z Med J. 2006 Aug 4;119(1239):U2103.

12) Robson B, Harris R, editors. Hauora: Māori Standards of Health IV. A study of the years 2000–2005. Wellington: Te Rōpū Rangahau Hauora Eru Pōmare; 2007.

13) King P. Māori with lived experience of disability: Part I, Wai 2575, #B22 [Internet]. Waitangi Tribunal; 2019 [cited 2022 May 12]. Available from: https://forms.justice.govt.nz/search/Documents/WT/wt_DOC_150437272/Wai%202575%2C%20B022.pdf.

14) Ministry of Health. Cardiovascular Disease Risk Assessment and Management for Primary Care [Internet]. Wellington: Ministry of Health; 2018 [cited 2022 Mar 18]. Available from: https://www.health.govt.nz/publication/cardiovascular-disease-risk-assessment-and-management-primary-care.

15) Ministry of Health, NZ Society for the Study of Diabetes. Screening for diabetes in asymptomatic adults - Type 2 Diabetes Management Guidelines [Internet]. Ministry of Health; 2022 [cited 2022 May 17]. Available from: https://t2dm.nzssd.org.nz/Section-112-Screening-for-diabetes-in-asymptomatic-adults.

16) National Heart Foundation of New Zealand. Evidence-based, best practice New Zealand Guidelines for Rheumatic Fever [Internet]. Auckland: National Heart Foundation of New Zealand; 2019 [cited 2022 Mar 18]. Available from: https://assets.heartfoundation.org.nz/documents/shop/marketing/non-stock-resources/diagnosis-management-rheumatic-fever-guideline.pdf.

17) BPAC. New diabetes medicines funded: empagliflozin and dulaglutide [Internet]. New Zealand: BPAC; 2021 [cited 2022 Mar 17]. Available from: https://bpac.org.nz/2021/docs/diabetes.pdf.

18) McLeod M, Harris R, Crengle S, Cormack D, Scott N, Robson B. Bowel cancer screening age range for Māori: what is all the fuss about? N Z Med J. 2021 May 21;134(1535):71-77.

19) Kilgour R, Keefe V. Kia Piki Te Ora: The collection of Maori health statistics. Wellington: Department of Health; 1992.

20) Robson B, Reid P. Ethnicity Matters: Review of the Measurement of Ethnicity in Official Statistics Māori perspectives paper for consultation [Internet]. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2001 [cited 2022 Mar 22]. Available from: https://www.otago.ac.nz/healthsciences/otago742784.pdf.

21) Reid P, Robson B. Understanding health inequities. In Robson B, Harris R, editors. Hauora: Māori Standards of Health IV. A study of the years 2000–2005. Wellington: Te Rōpū Rangahau Hauora Eru Pōmare; 2007. p.3-10.

22) Ministry of Health. HISO 10001:2017 Ethnicity Data Protocols [Internet]. Wellington, N.Z.: Ministry of Health; 2017 [cited 2022 Mar 22]. Available from: https://www.health.govt.nz/publication/hiso-100012017-ethnicity-data-protocols.

23) Paine SJ, Cormack D, Reid P, Harris R, Robson B. Kaupapa Māori-informed approaches to support data rights and self-determination 1. In: Indigenous Data Sovereignty and Policy. Routledge; 2020. p.187-203.

24) Statistics New Zealand. Integrated Data Infrastructure (IDI) refresh: linking report Statistical Methods, September 2021 refresh. Wellington: Statistics New Zealand; 2021.

25) Ministry of Health. COVID-19: Vaccine data. Spreadsheet covid_vaccinations_15_03_2022 [Internet]. Ministry of Health; 2022 [cited 2022 Mar 17]. Available from: https://www.health.govt.nz/covid-19-novel-coronavirus/covid-19-data-and-statistics/covid-19-vaccine-data.

26) Cormack D, Harris R. Issues in monitoring Māori health and ethnic disparities: an update. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2009.

27) Statistics New Zealand. Experimental administrative population census [Internet]. Statistics New Zealand; 2021 [cited 2022 Apr 4]. Available from: https://www.stats.govt.nz/experimental/experimental-administrative-population-census.

28) Carter KN, Hayward M, Blakely T, Shaw C. How much and for whom does self-identified ethnicity change over time in New Zealand? Results from a longitudinal study. Soc Policy J N Z. 2009;36:32-45.

29) Neuwelt P, Crengle S, Cormack D, McLeod M, Bramley D. General practice ethnicity data: evaluation of a tool. J Prim Health Care. 2014;6(1):49-55.

30) Randle J, Inns S, Harris R, McLeod M. Ethnicity data audit in a secondary care gastroenterology service. N Z Med J. 2022 Nov 11;15(1565):51-59.

31) Page M, Wyeth EH, Samaranayaka A, McNoe B, Walker R, Schollum J, et al. Accuracy of ethnicity data recorded in hospital-based patient clinical records and the Australia and New Zealand dialysis and transplant registry. N Z Med J. 2017 Apr 28;130(1454):65-71.

32) Riddell T, Lindsay G, Kenealy T, Jackson R, Crengle S, Bramley D, et al. The accuracy of ethnicity data in primary care and its impact on cardiovascular risk assessment and management–PREDICT CVD-8. N Z Med J. 2008 Sep 5;121(1281):40-8.

33) Ministry of Health. Primary Care Ethnicity Data Audit Toolkit [Internet]. Wellington: Mininstry of Health; 2021 [cited 2022 Mar 22]. Available from: https://www.health.govt.nz/system/files/documents/publications/primary-care-ethnicity-data-audit-toolkit-16dec21.pdf.

34) Curtis E, Loring B, Harris R, McLeod M, Mills C, Scott N, et al. Māori Health Priorities: A report commissioned by the interim Māori Health Authority (iMHA) to inform development of the interim New Zealand Health Plan (iNZHP). Wellington: interim Māori Health Authority; 2022.

35) Jones CP. Confronting institutionalized racism. Phylon (1960-). 2002;50:7-22.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Being counted is an acknowledgement of both existence and value. It means that one matters. It is the hallmark of Treaty promise (Orange 1987:257). History will judge our commitment to the Treaty in part by our ability to ensure that people are counted, that disparities are acknowledged and appropriate policies are put in place, especially those which eliminate disparities between Māori and non-Māori, a solemn commitment by the Treaty of Waitangi.”[[1]]

Ethnicity data matter. Longstanding and significant inequities exist between Māori and non-Māori across most health indicators including life expectancy, health determinants, health outcomes and healthcare. These inequities are rooted in processes of colonisation and colonialism, supported by a system of racism.[[2]] They are a breach of the Te Tiriti o Waitangi and Māori Indigenous rights.[[2]] Furthermore, the lack of urgency towards their elimination, particularly by Crown agencies, further reflects the disregard with which Māori and Māori health are held.

Ethnicity data have been collected for decades in health and population datasets; however, the quality, consistency and completeness continue to be problematic, particularly for Māori.[[3]] This has often been reflected in the systematic under-counting of Māori in datasets including the National Health Index (NHI),[[4,5]] in mortality datasets,[[6,7]] and others.[[8–13]] This then impacts on the ability to estimate health prevalence, rates, inequities and progress (or not) over time. Accurate ethnicity data are also critical to correctly identify individuals eligible for healthcare interventions. In New Zealand, an individual’s ethnicity determines the age of their eligibility for cardiovascular risk assessment[[14]] and diabetes screening,[[15]] sore throat management,[[16]] eligibility for diabetes treatments,[[17]] and the proposed age extension to the Bowel Cancer Screening Programme.[[18]]

Māori health advocates have long argued for high quality ethnicity data in health as a Māori health right;[[1,19,20]] Māori have a “right to be counted”[[2]] and “Māori have the right to monitor the Crown and to evaluate Crown action and inaction”.[[21]] While protocols for the standardised collection of ethnicity data in the health and disability sector[[22]] have existed for nearly 20 years, their implementation seems yet to be fully realised, with ongoing concerns regarding the under-counting of Māori in health and disability sector data. This study examines ethnicity data quality and its implications for Māori health and inequities. We focus on NHI ethnicity as a consistent variable across a range of health datasets. The NHI provides a unique identifier for each person who receives healthcare with associated identifying information including ethnicity.[[4]] NHI ethnicity is the most up-to-date record of an individual’s ethnicity based on their most recent health encounters. We seek to examine in detail NHI ethnicity data quality (particularly under-count) for Māori by comparing aggregate level data for Māori and non-Māori across a range of population and health datasets, and by comparing Māori and non-Māori NHI ethnicity among individuals enrolled with a primary health organisation (PHO) with their linked self-identified ethnicity from the 2018 Census. We also examine differences by age and gender.

Methods

We are guided by kaupapa Māori research methods, particularly kaupapa Māori epidemiology that includes: 1) the right to monitor the Crown; 2) the right to be counted; 3) the right to have a powerful voice; and 4) the right to name racism and colonialism as fundamental causes of inequities for Māori.[[23]] Our kaupapa Māori epidemiological approach supports the use of a Māori/non-Māori analysis as a reflection of Te Tiriti.

Study design

This is a descriptive study that compares the recording of Māori and non-Māori ethnicity across a range of datasets and in two main ways:

1. Population aggregate analysis

2. Individual linked analysis

In these analyses, Māori are defined as anyone who has their ethnicity recorded as Māori, either alone, or in combination with another ethnic group(s).[[22]] Non-Māori are all other people in the dataset (including those with missing ethnicity). Gender could only be examined by the binary categories of male and female due to dataset limitations. The study was approved by the University of Otago Human Ethics Committee (HD 20/079).

1. Population aggregate analysis

At an aggregate level, total population data are compared for Māori and non-Māori across a range of population and health datasets to compare both counts and proportions of Māori and non-Māori. The datasets include the 2018 Estimated Resident Population (ERP); 2018 Census Usually Resident Population (Census); 2018 Health Service User (HSU) population; and 2018 Primary Health Organisation (PHO) population. Appendix 1 provides further details on the datasets. All data are estimated at a single point in time and as close as possible to the same point in time (30 June 2018) for comparability.

The two health datasets (HSU and PHO) are compared with official population estimates (ERP) to examine potential differential under-representation of Māori compared to non-Māori in health data (either from differential access to care or under-counting of Māori). We compare the total numbers and proportions of Māori and non-Māori in each dataset overall, and by age and gender. This aggregate analysis provides an indication of potential under-counting of Māori if proportions of Māori in health data are lower than non-Māori compared to official population data. However, the contribution of differential access to care cannot be teased out from potential under-counting.

2. Individual linked analysis

In order to examine the misclassification of ethnicity and potential systematic under-counting of Māori on NHI data, we undertook a linked analysis at the level of individuals. To be eligible for inclusion, people had to be enrolled in a PHO (as at 1 April 2018), have self-completed the ethnicity question in the 2018 Census, linked in the NZ Integrated Data Infrastructure (IDI) spine and to Ministry of Health (MoH) NHI data.

Analyses were undertaken in the Statistics NZ (Stats NZ) datalab using the September 2021 IDI Refresh (available from 27 October 2021). Data were linked between the IDI spine and MoH.[[24]] 2018 Census data were restricted to only those who had self-identified their ethnicity in this Census i.e., ethnicity supplemented from other administrative data were not used (Appendix 1). Ethnicity data from the 2018 Census were considered the standard against which PHO NHI ethnicity data were compared. This is in line with the Ethnicity Data Protocols for the Health Sector and the Census, whereby a standard question should be used and people should self-identify their own ethnicity whenever possible, with guidance on instances where this may not be possible, e.g., young children, people who are incapacitated.[[ 22]] Those with missing NHI ethnicity (75,243 individuals; 2%) were categorised as non-Māori.

Among linked individuals, two main measures were used to examine data quality. First, we calculated the proportion whose ethnicity was the same in both datasets using the Māori/non-Māori population denominators from the Census (proportion matched). This provides a measure of misclassification. Secondly, the net under-count of Māori on the NHI was estimated by 1 minus (the number of people classified as Māori on the NHI as a percentage of those classified as Māori on the Census). This indicates the extent to which the misclassification is differential.

Results

View Tables & Figures.

Population aggregate results

Table 1 shows the population counts across population and health datasets with numbers (and proportions) for Māori and non-Māori. Of the four datasets examined, the ERP has the most people as well as the highest proportion of Māori (16.7%). The total numbers in the HSU dataset are closest to the ERP, although lower for Māori and (correspondingly) slightly higher for non-Māori. The proportion of Māori in the HSU is lower than the ERP at 15.5%. The PHO dataset has the least people and the lowest proportion of Māori (14.6%).

Figure 1 shows the number of people recorded as Māori and non-Māori in the datasets by 5-year age group. The ERP has the highest number of Māori across almost all ages. The exception is the youngest age group where ERP and HSU numbers are similar. PHO data are lowest for Māori across all ages. Differences between datasets are less obvious with increasing age.

For non-Māori, PHO and Census data have lower counts across most ages compared to ERP. For PHO data, this is more obvious for adolescents and young adults. The HSU data has higher numbers of non-Māori than ERP in children, lower numbers in young adults and similar numbers among adults and older people.

Compared to the ERP, there are fewer Māori males and females in every age group in the Census, HSU (except 0–4 years) and PHO datasets and the percentage differences are disproportionately larger than for non-Māori (Figure 2). The percentage differences for males tend to be disproportionately larger than for females in the datasets (Census, HSU, PHO) compared to the ERP, for both Māori and non-Māori. Within the HSU data, Māori males are under-represented across most age groups. For non-Māori, there is under-representation of young males and over-representation of females in multiple age groups.

Individual linked results

3,498,789 people on the PHO dataset were linked to a 2018 Census record. This equates to 78% linkage of people on the PHO dataset. Linkage varied by ethnic grouping (based on NHI ethnicity on the PHO dataset), with 64% linkage for Māori and 80% for non-Māori. Within the PHO linked sample, 491,928 (14.1%) of the population were Māori (using Census ethnicity data), compared to 414,846 (11.9%) using NHI data (Table 2). Both of these proportions are lower than the total population percentage of Māori in the ERP of 16.7% (Table 1).

Of the 491,928 individuals who identified as Māori on the Census, 386,976 (78.7%) were also recorded as Māori on the NHI (Table 2). Of the 3 million individuals categorised as non-Māori ethnicity from the Census, 2,978,991 (99.1%) were categorised as non-Māori on the NHI. The high match of NHI and Census ethnicity for non-Māori was seen across all age groups.

For Māori (on the Census) there is some variation by age, with the youngest age group (0–4 years) having a higher match (approx. 90%) and a lower level of match in those aged 20–24 years (Figure 3). The level of match is lower for Māori males than Māori females, and it is more apparent in adults over 20 years old where matching for Māori males remains consistently low.

Figure 4 examines the net under-counting of Māori ethnicity on the NHI by age and gender. Net under-count tells us overall how much the NHI data under-counts Māori compared to the 2018 Census (assuming the Census is the “correct” classification). The net under-count captures the proportion of Māori missed from the NHI (i.e., those identified as Māori on the Census but not on NHI) minus non-Māori incorrectly classified as Māori in NHI (those identified as non-Māori on the Census but Māori on NHI).

The net under-count in the NHI for Māori is 15.7% overall. The net under-count for 0–4-year-olds  is less than 5%, but increases to between 13–23% for all other age groups. When this is examined separately by gender, the net under-count for Māori males is higher than Māori females from 20 years on, and is greater than 20% in most age groups (20–59 years). For Māori women, there is more variation in net under-count by age, with peaks in the under-count in the age groups 20–24 years (22.7% under-count) 40–44 years (18.4% under-count), and 80+ years (19.2% under-count).

Discussion

High quality ethnicity data are fundamental for understanding and monitoring Māori health and health inequities as well as in the provision of targeted services and interventions that are responsive to Māori aspirations and needs. This study shows that Māori are under-represented in health datasets (HSU and PHO), compared to official population numbers (ERP), and are systematically under-counted in NHI ethnicity data. At a population aggregate level, all four datasets had different total population and Māori numbers, reflecting differences in their composition. Additionally, there were differences in the proportion of Māori in each dataset ranging from 16.7% on the ERP to 14.6% on the PHO dataset. There were lower numbers of Māori in the HSU and PHO datasets compared to the ERP across almost all ages. This under-representation of Māori in health data is likely to be due to under-counting of Māori in health datasets as well as ongoing differential access to healthcare. This was demonstrated in the linked analysis of individuals whereby NHI ethnicity (from PHO data) was more likely to misclassify Māori as non-Māori (than non-Māori as Māori), resulting in a 16% net under-count of Māori on the NHI (with variation by age and gender).

It is clear that the health and disability sector continues to fail in its responsibility to achieve expected standards of ethnicity data collection,[[22]] with important implications for our ability to monitor Māori health and equity, and to target health services. In the monitoring of Māori health and equity, the under-counting of Māori in the NHI can lead to a numerator/denominator bias when rates using population data from other sources such as ERP are used, as is common practice. This will lead to underestimation of the Māori rate for any health variable of interest. In contrast, the impact of any misclassification for non-Māori will be negligible because of the much larger population. This can make inequities look better or worse than they really are, depending on the health variable concerned.

The HSU population is being used as one method to try and minimise the numerator/denominator bias by using it as the population denominator.[[4]] The 2020 HSU has been used recently in the estimation of COVID-19 vaccination coverage.[[25]] However, this may allow the inclusion of people in the numerator that may not be in the denominator (e.g., people who have been vaccinated but had otherwise not had a health interaction in the period of the HSU). This can potentially overestimate rates. Our study shows that the HSU should not be used to estimate actual numbers of Māori eligible for services (e.g., the number of Māori to be vaccinated), as it significantly under-represents Māori.

Other methods have also been used to mitigate the numerator/denominator bias for Māori.[[26]] These include the calculation of ratios to adjust for Māori under-count[[6,10,12]] that use similar methods as here to estimate the net under-counting of Māori. More recently, administrative datasets are being linked to create new populations in efforts to improve population estimates by ethnicity, e.g., the experimental Administrative Population Census derived from data in the IDI.[[27]] However, these still rely on the underlying quality of ethnicity data in the linked sources. It is important that Māori have the opportunity to self-identify their ethnicity. This is in line with Indigenous rights, as an expression of rangatiratanga and Māori rights to determine our own identities.[[20]]

There are implications with using ethnicity data to determine an individual’s eligibility for health and disability services. A number of health services already prioritise Māori in the assessment of risk or service delivery.[[14–18]] Using the example of bowel cancer screening, where a lower age (from 50 years) of invitation for Māori has been proposed, a large number of Māori who are eligible would be missed if invitations to screening drew on NHI ethnicity. A smaller number of non-Māori would also be mistakenly invited.

The under-counting of Māori in health data is not new and has been examined in various settings over time. Particularly relevant to our study are other analyses that have linked individual health data with Census data. Stats NZ linked 2013 Census data to a combined NHI dataset and found Māori were under-counted by 21%.[[5]] More recently, the MoH has undertaken similar analyses linking people from the 2018 Census to 2019 NHI data and shown a 16% net under-count of Māori in the NHI compared to Census.[[4]] Our analysis also shows a 16% net under-count of Māori on the NHI from the PHO database compared to individuals’ corresponding Census ethnicity.

There are a number of limitations that should be considered in the interpretation of our findings. Firstly, our analysis of linked data is a subset of Māori in the 2018 Census and 2018 PHO datasets with lower linkage for Māori (based on NHI ethnicity). We cannot determine if the level of mismatch and under-counting of those included in this analysis is the same as those not included. While we examined data by gender and age, we were unable to explore ethnicity data for other important groups, for example those who are gender diverse and/or tāngata whaikaha. Secondly, it is possible that people self-report their ethnicity differently, in different settings, and that they can change their ethnicity.[[28]] However, there is also evidence from health settings that ethnicity data protocols are not always adhered to,[[29]] e.g., guessing someone’s ethnicity based on name or appearance, disagreeing with patient’s self-identified ethnicity, non-standardised forms and data recording. In addition, there is evidence of improvements in ethnicity data collection in some health-related datasets, e.g., mortality,[[6,7]] cancer registrations,[[10]] and examples where there is no (or minor) net under-count of Māori.[[9,30–32]] These findings would suggest that (non)adherence to ethnicity data protocols are the likely drivers of the systematic under-counting of Māori. This is a complex issue to tease out in detail, and we have chosen to focus on implications for Māori using a non-Māori comparator. However, we acknowledge that the non-Māori group is made up of multiple ethnicities combined, and that the quality of ethnicity data for specific ethnic groups, and the mismatch of ethnicity within this large grouping are not examined. In particular, similar data issues and implications are likely to be present for Pacific peoples.[[4,5]] We have also not looked at all potential measures of data quality here such as multiple ethnicities, missing data, and consistency with health and disability sector standards.[[5,33]]

The ongoing systematic under-counting of Māori in health data is a reflection that Māori are not valued and that Māori health and the elimination of inequities are not the priorities they are claimed to be. The ongoing unjust and inequitable healthcare experiences and outcomes faced by Māori require the whole of the health and disability sector to commit to the collection of high-quality ethnicity data, in order to understand inequities; to deliver equitable services to Māori; to act to address inequity; and to monitor progress on eliminating inequities. This commitment needs to come from the highest levels and be attached to accountability mechanisms and ongoing quality assessments.[[34]] Quality ethnicity data are fundamental to the elimination of Māori health inequities. The under-counting of Māori remains a breach of Te Tiriti o Waitangi and Indigenous rights. The ongoing acceptance of poor-quality ethnicity data for Māori and the inadequate progress towards high quality ethnicity data in the health and disability sector is itself evidence of racism as “inaction in the face of need”.[[35]]

View Appendices.

Summary

Abstract

Aim

To examine ethnicity data quality; in particular, the representation and potential under-counting of Māori in health and disability sector data, as well as implications for inequities.

Method

Māori and non-Māori ethnicity data are analysed at: 1) a population aggregate level across multiple 2018 datasets (Estimated Resident Population, Census Usually Resident Population, Health Service User (HSU) population and Primary Health Organisation (PHO) enrolments); and 2) an individual level for those linked in PHO and 2018 Census datasets. Ethnicity is drawn from the National Health Index (NHI) in health datasets and variations by age and gender are explored.

Results

Aggregate analyses show that Māori are considerably under-represented in HSU and PHO data. In linked analysis Māori were under-counted on the NHI by 16%. Under-representation in data and under-counting occur across both genders but are more pronounced for Māori men with variations by age.

Conclusion

High quality ethnicity data are fundamental for understanding and monitoring Māori health and health inequities as well as in the provision of targeted services and interventions that are responsive to Māori aspirations and needs. The continued under-counting of Māori in health and disability sector data is a breach of Te Tiriti o Waitangi and must be addressed with urgency.

Author Information

Ricci Harris: Associate Professor/Public Health Physician, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Sarah-Jane Paine: Senior Lecturer, Te Kupenga Hauora Māori, University of Auckland, New Zealand. June Atkinson: Senior Data Analyst, Department of Public Health, University of Otago, Wellington, New Zealand. Bridget Robson: Associate Professor and Director of Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Paula Toko King: Public Health Physician/Senior Research Fellow, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Jennifer Randle: Public Health Medicine Registrar, Department of Public Health, University of Otago, Wellington, New Zealand. Anja Mizdrak: Senior Research Fellow, Department of Public Health, University of Otago, Wellington, New Zealand. Melissa McLeod: Public Health Physician/Senior Lecturer, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand.

Acknowledgements

The study was funded in part by the grant from the Health Research Council of New Zealand (HRC 20/152). We would like to thank Associate Professor Donna Cormack for comments on a draft of the paper. Data in the paper were provided from Stats New Zealand and the New Zealand Ministry of Health. The Ministry of Health retains intellectual property rights to the data they provided. Stats NZ disclaimer regarding IDI data: The results in this paper are not official statistics, they have been created for research purposes from the Integrated Data Infrastructure (IDI) managed by Statistics New Zealand. The opinions, findings, recommendations and conclusions expressed in this paper are those of the author(s) not Statistics NZ or Ministry of Health. Access to the anonymised data used in this study was provided by Statistics NZ in accordance with security and confidentiality provisions of the Statistics Act 1975. Only people authorised by the Statistics Act 1975 are allowed to see data about a particular person, household, business or organisation and the results in this paper have been confidentialised to protect these groups from identification. Careful consideration has been given to the privacy, security and confidentiality issues associated with using administrative and survey data in the IDI. Further detail can be found in the Privacy impact assessment for the Integrated Data Infrastructure available from www.stats.govt.nz.

Correspondence

Associate Professor Ricci Harris: Te Rōpū Rangahau Hauora a Eru Pōmare, Department of Public Health, University of Otago, Wellington PO Box 7343, Wellington 6242, New Zealand.

Correspondence Email

Ricci.harris@otago.ac.nz

Competing Interests

The authors have no conflicts of interest or financial disclosures.

1) Te Rōpū Rangahau Hauora a Eru Pōmare. Counting for nothing: understanding the issues in monitoring disparities in health. Soc Policy J N Z. 2000;14:1-16.

2) Reid P, Cormack D, Paine SJ. Colonial histories, racism and health—The experience of Māori and Indigenous peoples. Public Health. 2019 Jul;172:119-24.

3) Cormack D, McLeod M. Improving and maintaining quality in ethnicity data collections in the health and disability sector [Internet]. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2010 [cited 2022 Mar 17]. Available from: https://www.otago.ac.nz/wellington/otago600098.pdf.

4) Cleary L. Using ethnicity data in Health Statistics. Wellington: Ministry of Health; 2021. Report No.: 1.1.

5) Reid G, Bycroft C, Gleisner F. Comparison of ethnicity information in administrative data and the census. Wellington: Stats NZ; 2016.

6) Ajwani S, Blakely T, Robson B, Atkinson J, Kiro C. Unlocking the numerator-denominator bias III: adjustment ratios by ethnicity for 1981-1999 mortality data. The New Zealand Census-Mortality Study. N Z Med J. 2003;116(1175):U456.

7) Tan L, Blakely T, Atkinson J. Ethnic counts on mortality and census data 2001-06: New Zealand census-mortality study update. N Z Med J. 2010;123(1320):37-44.

8) Bramley D, Latimer S. The accuracy of ethnicity data in primary care. N Z Med J. 2007 Oct 26;120(1264):U2779.

9) Scott N, Clark H, Kool B, Ameratunga S, Christey G, Cormack D. Audit of ethnicity data in the Waikato Hospital Patient Management System and Trauma Registry: pilot of the Hospital Ethnicity Data Audit Toolkit. N Z Med J. 2018 Oct 5;131(1483):21-9.

10) Shaw C, Atkinson J, Blakely T. (Mis)classification of ethnicity on the New Zealand Cancer Registry: 1981-2004. N Z Med J. 2009 May 8;122(1294):10-22.

11) Swan J, Lillis S, Simmons D. Investigating the accuracy of ethnicity data in New Zealand hospital records: still room for improvement. N Z Med J. 2006 Aug 4;119(1239):U2103.

12) Robson B, Harris R, editors. Hauora: Māori Standards of Health IV. A study of the years 2000–2005. Wellington: Te Rōpū Rangahau Hauora Eru Pōmare; 2007.

13) King P. Māori with lived experience of disability: Part I, Wai 2575, #B22 [Internet]. Waitangi Tribunal; 2019 [cited 2022 May 12]. Available from: https://forms.justice.govt.nz/search/Documents/WT/wt_DOC_150437272/Wai%202575%2C%20B022.pdf.

14) Ministry of Health. Cardiovascular Disease Risk Assessment and Management for Primary Care [Internet]. Wellington: Ministry of Health; 2018 [cited 2022 Mar 18]. Available from: https://www.health.govt.nz/publication/cardiovascular-disease-risk-assessment-and-management-primary-care.

15) Ministry of Health, NZ Society for the Study of Diabetes. Screening for diabetes in asymptomatic adults - Type 2 Diabetes Management Guidelines [Internet]. Ministry of Health; 2022 [cited 2022 May 17]. Available from: https://t2dm.nzssd.org.nz/Section-112-Screening-for-diabetes-in-asymptomatic-adults.

16) National Heart Foundation of New Zealand. Evidence-based, best practice New Zealand Guidelines for Rheumatic Fever [Internet]. Auckland: National Heart Foundation of New Zealand; 2019 [cited 2022 Mar 18]. Available from: https://assets.heartfoundation.org.nz/documents/shop/marketing/non-stock-resources/diagnosis-management-rheumatic-fever-guideline.pdf.

17) BPAC. New diabetes medicines funded: empagliflozin and dulaglutide [Internet]. New Zealand: BPAC; 2021 [cited 2022 Mar 17]. Available from: https://bpac.org.nz/2021/docs/diabetes.pdf.

18) McLeod M, Harris R, Crengle S, Cormack D, Scott N, Robson B. Bowel cancer screening age range for Māori: what is all the fuss about? N Z Med J. 2021 May 21;134(1535):71-77.

19) Kilgour R, Keefe V. Kia Piki Te Ora: The collection of Maori health statistics. Wellington: Department of Health; 1992.

20) Robson B, Reid P. Ethnicity Matters: Review of the Measurement of Ethnicity in Official Statistics Māori perspectives paper for consultation [Internet]. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2001 [cited 2022 Mar 22]. Available from: https://www.otago.ac.nz/healthsciences/otago742784.pdf.

21) Reid P, Robson B. Understanding health inequities. In Robson B, Harris R, editors. Hauora: Māori Standards of Health IV. A study of the years 2000–2005. Wellington: Te Rōpū Rangahau Hauora Eru Pōmare; 2007. p.3-10.

22) Ministry of Health. HISO 10001:2017 Ethnicity Data Protocols [Internet]. Wellington, N.Z.: Ministry of Health; 2017 [cited 2022 Mar 22]. Available from: https://www.health.govt.nz/publication/hiso-100012017-ethnicity-data-protocols.

23) Paine SJ, Cormack D, Reid P, Harris R, Robson B. Kaupapa Māori-informed approaches to support data rights and self-determination 1. In: Indigenous Data Sovereignty and Policy. Routledge; 2020. p.187-203.

24) Statistics New Zealand. Integrated Data Infrastructure (IDI) refresh: linking report Statistical Methods, September 2021 refresh. Wellington: Statistics New Zealand; 2021.

25) Ministry of Health. COVID-19: Vaccine data. Spreadsheet covid_vaccinations_15_03_2022 [Internet]. Ministry of Health; 2022 [cited 2022 Mar 17]. Available from: https://www.health.govt.nz/covid-19-novel-coronavirus/covid-19-data-and-statistics/covid-19-vaccine-data.

26) Cormack D, Harris R. Issues in monitoring Māori health and ethnic disparities: an update. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2009.

27) Statistics New Zealand. Experimental administrative population census [Internet]. Statistics New Zealand; 2021 [cited 2022 Apr 4]. Available from: https://www.stats.govt.nz/experimental/experimental-administrative-population-census.

28) Carter KN, Hayward M, Blakely T, Shaw C. How much and for whom does self-identified ethnicity change over time in New Zealand? Results from a longitudinal study. Soc Policy J N Z. 2009;36:32-45.

29) Neuwelt P, Crengle S, Cormack D, McLeod M, Bramley D. General practice ethnicity data: evaluation of a tool. J Prim Health Care. 2014;6(1):49-55.

30) Randle J, Inns S, Harris R, McLeod M. Ethnicity data audit in a secondary care gastroenterology service. N Z Med J. 2022 Nov 11;15(1565):51-59.

31) Page M, Wyeth EH, Samaranayaka A, McNoe B, Walker R, Schollum J, et al. Accuracy of ethnicity data recorded in hospital-based patient clinical records and the Australia and New Zealand dialysis and transplant registry. N Z Med J. 2017 Apr 28;130(1454):65-71.

32) Riddell T, Lindsay G, Kenealy T, Jackson R, Crengle S, Bramley D, et al. The accuracy of ethnicity data in primary care and its impact on cardiovascular risk assessment and management–PREDICT CVD-8. N Z Med J. 2008 Sep 5;121(1281):40-8.

33) Ministry of Health. Primary Care Ethnicity Data Audit Toolkit [Internet]. Wellington: Mininstry of Health; 2021 [cited 2022 Mar 22]. Available from: https://www.health.govt.nz/system/files/documents/publications/primary-care-ethnicity-data-audit-toolkit-16dec21.pdf.

34) Curtis E, Loring B, Harris R, McLeod M, Mills C, Scott N, et al. Māori Health Priorities: A report commissioned by the interim Māori Health Authority (iMHA) to inform development of the interim New Zealand Health Plan (iNZHP). Wellington: interim Māori Health Authority; 2022.

35) Jones CP. Confronting institutionalized racism. Phylon (1960-). 2002;50:7-22.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Being counted is an acknowledgement of both existence and value. It means that one matters. It is the hallmark of Treaty promise (Orange 1987:257). History will judge our commitment to the Treaty in part by our ability to ensure that people are counted, that disparities are acknowledged and appropriate policies are put in place, especially those which eliminate disparities between Māori and non-Māori, a solemn commitment by the Treaty of Waitangi.”[[1]]

Ethnicity data matter. Longstanding and significant inequities exist between Māori and non-Māori across most health indicators including life expectancy, health determinants, health outcomes and healthcare. These inequities are rooted in processes of colonisation and colonialism, supported by a system of racism.[[2]] They are a breach of the Te Tiriti o Waitangi and Māori Indigenous rights.[[2]] Furthermore, the lack of urgency towards their elimination, particularly by Crown agencies, further reflects the disregard with which Māori and Māori health are held.

Ethnicity data have been collected for decades in health and population datasets; however, the quality, consistency and completeness continue to be problematic, particularly for Māori.[[3]] This has often been reflected in the systematic under-counting of Māori in datasets including the National Health Index (NHI),[[4,5]] in mortality datasets,[[6,7]] and others.[[8–13]] This then impacts on the ability to estimate health prevalence, rates, inequities and progress (or not) over time. Accurate ethnicity data are also critical to correctly identify individuals eligible for healthcare interventions. In New Zealand, an individual’s ethnicity determines the age of their eligibility for cardiovascular risk assessment[[14]] and diabetes screening,[[15]] sore throat management,[[16]] eligibility for diabetes treatments,[[17]] and the proposed age extension to the Bowel Cancer Screening Programme.[[18]]

Māori health advocates have long argued for high quality ethnicity data in health as a Māori health right;[[1,19,20]] Māori have a “right to be counted”[[2]] and “Māori have the right to monitor the Crown and to evaluate Crown action and inaction”.[[21]] While protocols for the standardised collection of ethnicity data in the health and disability sector[[22]] have existed for nearly 20 years, their implementation seems yet to be fully realised, with ongoing concerns regarding the under-counting of Māori in health and disability sector data. This study examines ethnicity data quality and its implications for Māori health and inequities. We focus on NHI ethnicity as a consistent variable across a range of health datasets. The NHI provides a unique identifier for each person who receives healthcare with associated identifying information including ethnicity.[[4]] NHI ethnicity is the most up-to-date record of an individual’s ethnicity based on their most recent health encounters. We seek to examine in detail NHI ethnicity data quality (particularly under-count) for Māori by comparing aggregate level data for Māori and non-Māori across a range of population and health datasets, and by comparing Māori and non-Māori NHI ethnicity among individuals enrolled with a primary health organisation (PHO) with their linked self-identified ethnicity from the 2018 Census. We also examine differences by age and gender.

Methods

We are guided by kaupapa Māori research methods, particularly kaupapa Māori epidemiology that includes: 1) the right to monitor the Crown; 2) the right to be counted; 3) the right to have a powerful voice; and 4) the right to name racism and colonialism as fundamental causes of inequities for Māori.[[23]] Our kaupapa Māori epidemiological approach supports the use of a Māori/non-Māori analysis as a reflection of Te Tiriti.

Study design

This is a descriptive study that compares the recording of Māori and non-Māori ethnicity across a range of datasets and in two main ways:

1. Population aggregate analysis

2. Individual linked analysis

In these analyses, Māori are defined as anyone who has their ethnicity recorded as Māori, either alone, or in combination with another ethnic group(s).[[22]] Non-Māori are all other people in the dataset (including those with missing ethnicity). Gender could only be examined by the binary categories of male and female due to dataset limitations. The study was approved by the University of Otago Human Ethics Committee (HD 20/079).

1. Population aggregate analysis

At an aggregate level, total population data are compared for Māori and non-Māori across a range of population and health datasets to compare both counts and proportions of Māori and non-Māori. The datasets include the 2018 Estimated Resident Population (ERP); 2018 Census Usually Resident Population (Census); 2018 Health Service User (HSU) population; and 2018 Primary Health Organisation (PHO) population. Appendix 1 provides further details on the datasets. All data are estimated at a single point in time and as close as possible to the same point in time (30 June 2018) for comparability.

The two health datasets (HSU and PHO) are compared with official population estimates (ERP) to examine potential differential under-representation of Māori compared to non-Māori in health data (either from differential access to care or under-counting of Māori). We compare the total numbers and proportions of Māori and non-Māori in each dataset overall, and by age and gender. This aggregate analysis provides an indication of potential under-counting of Māori if proportions of Māori in health data are lower than non-Māori compared to official population data. However, the contribution of differential access to care cannot be teased out from potential under-counting.

2. Individual linked analysis

In order to examine the misclassification of ethnicity and potential systematic under-counting of Māori on NHI data, we undertook a linked analysis at the level of individuals. To be eligible for inclusion, people had to be enrolled in a PHO (as at 1 April 2018), have self-completed the ethnicity question in the 2018 Census, linked in the NZ Integrated Data Infrastructure (IDI) spine and to Ministry of Health (MoH) NHI data.

Analyses were undertaken in the Statistics NZ (Stats NZ) datalab using the September 2021 IDI Refresh (available from 27 October 2021). Data were linked between the IDI spine and MoH.[[24]] 2018 Census data were restricted to only those who had self-identified their ethnicity in this Census i.e., ethnicity supplemented from other administrative data were not used (Appendix 1). Ethnicity data from the 2018 Census were considered the standard against which PHO NHI ethnicity data were compared. This is in line with the Ethnicity Data Protocols for the Health Sector and the Census, whereby a standard question should be used and people should self-identify their own ethnicity whenever possible, with guidance on instances where this may not be possible, e.g., young children, people who are incapacitated.[[ 22]] Those with missing NHI ethnicity (75,243 individuals; 2%) were categorised as non-Māori.

Among linked individuals, two main measures were used to examine data quality. First, we calculated the proportion whose ethnicity was the same in both datasets using the Māori/non-Māori population denominators from the Census (proportion matched). This provides a measure of misclassification. Secondly, the net under-count of Māori on the NHI was estimated by 1 minus (the number of people classified as Māori on the NHI as a percentage of those classified as Māori on the Census). This indicates the extent to which the misclassification is differential.

Results

View Tables & Figures.

Population aggregate results

Table 1 shows the population counts across population and health datasets with numbers (and proportions) for Māori and non-Māori. Of the four datasets examined, the ERP has the most people as well as the highest proportion of Māori (16.7%). The total numbers in the HSU dataset are closest to the ERP, although lower for Māori and (correspondingly) slightly higher for non-Māori. The proportion of Māori in the HSU is lower than the ERP at 15.5%. The PHO dataset has the least people and the lowest proportion of Māori (14.6%).

Figure 1 shows the number of people recorded as Māori and non-Māori in the datasets by 5-year age group. The ERP has the highest number of Māori across almost all ages. The exception is the youngest age group where ERP and HSU numbers are similar. PHO data are lowest for Māori across all ages. Differences between datasets are less obvious with increasing age.

For non-Māori, PHO and Census data have lower counts across most ages compared to ERP. For PHO data, this is more obvious for adolescents and young adults. The HSU data has higher numbers of non-Māori than ERP in children, lower numbers in young adults and similar numbers among adults and older people.

Compared to the ERP, there are fewer Māori males and females in every age group in the Census, HSU (except 0–4 years) and PHO datasets and the percentage differences are disproportionately larger than for non-Māori (Figure 2). The percentage differences for males tend to be disproportionately larger than for females in the datasets (Census, HSU, PHO) compared to the ERP, for both Māori and non-Māori. Within the HSU data, Māori males are under-represented across most age groups. For non-Māori, there is under-representation of young males and over-representation of females in multiple age groups.

Individual linked results

3,498,789 people on the PHO dataset were linked to a 2018 Census record. This equates to 78% linkage of people on the PHO dataset. Linkage varied by ethnic grouping (based on NHI ethnicity on the PHO dataset), with 64% linkage for Māori and 80% for non-Māori. Within the PHO linked sample, 491,928 (14.1%) of the population were Māori (using Census ethnicity data), compared to 414,846 (11.9%) using NHI data (Table 2). Both of these proportions are lower than the total population percentage of Māori in the ERP of 16.7% (Table 1).

Of the 491,928 individuals who identified as Māori on the Census, 386,976 (78.7%) were also recorded as Māori on the NHI (Table 2). Of the 3 million individuals categorised as non-Māori ethnicity from the Census, 2,978,991 (99.1%) were categorised as non-Māori on the NHI. The high match of NHI and Census ethnicity for non-Māori was seen across all age groups.

For Māori (on the Census) there is some variation by age, with the youngest age group (0–4 years) having a higher match (approx. 90%) and a lower level of match in those aged 20–24 years (Figure 3). The level of match is lower for Māori males than Māori females, and it is more apparent in adults over 20 years old where matching for Māori males remains consistently low.

Figure 4 examines the net under-counting of Māori ethnicity on the NHI by age and gender. Net under-count tells us overall how much the NHI data under-counts Māori compared to the 2018 Census (assuming the Census is the “correct” classification). The net under-count captures the proportion of Māori missed from the NHI (i.e., those identified as Māori on the Census but not on NHI) minus non-Māori incorrectly classified as Māori in NHI (those identified as non-Māori on the Census but Māori on NHI).

The net under-count in the NHI for Māori is 15.7% overall. The net under-count for 0–4-year-olds  is less than 5%, but increases to between 13–23% for all other age groups. When this is examined separately by gender, the net under-count for Māori males is higher than Māori females from 20 years on, and is greater than 20% in most age groups (20–59 years). For Māori women, there is more variation in net under-count by age, with peaks in the under-count in the age groups 20–24 years (22.7% under-count) 40–44 years (18.4% under-count), and 80+ years (19.2% under-count).

Discussion

High quality ethnicity data are fundamental for understanding and monitoring Māori health and health inequities as well as in the provision of targeted services and interventions that are responsive to Māori aspirations and needs. This study shows that Māori are under-represented in health datasets (HSU and PHO), compared to official population numbers (ERP), and are systematically under-counted in NHI ethnicity data. At a population aggregate level, all four datasets had different total population and Māori numbers, reflecting differences in their composition. Additionally, there were differences in the proportion of Māori in each dataset ranging from 16.7% on the ERP to 14.6% on the PHO dataset. There were lower numbers of Māori in the HSU and PHO datasets compared to the ERP across almost all ages. This under-representation of Māori in health data is likely to be due to under-counting of Māori in health datasets as well as ongoing differential access to healthcare. This was demonstrated in the linked analysis of individuals whereby NHI ethnicity (from PHO data) was more likely to misclassify Māori as non-Māori (than non-Māori as Māori), resulting in a 16% net under-count of Māori on the NHI (with variation by age and gender).

It is clear that the health and disability sector continues to fail in its responsibility to achieve expected standards of ethnicity data collection,[[22]] with important implications for our ability to monitor Māori health and equity, and to target health services. In the monitoring of Māori health and equity, the under-counting of Māori in the NHI can lead to a numerator/denominator bias when rates using population data from other sources such as ERP are used, as is common practice. This will lead to underestimation of the Māori rate for any health variable of interest. In contrast, the impact of any misclassification for non-Māori will be negligible because of the much larger population. This can make inequities look better or worse than they really are, depending on the health variable concerned.

The HSU population is being used as one method to try and minimise the numerator/denominator bias by using it as the population denominator.[[4]] The 2020 HSU has been used recently in the estimation of COVID-19 vaccination coverage.[[25]] However, this may allow the inclusion of people in the numerator that may not be in the denominator (e.g., people who have been vaccinated but had otherwise not had a health interaction in the period of the HSU). This can potentially overestimate rates. Our study shows that the HSU should not be used to estimate actual numbers of Māori eligible for services (e.g., the number of Māori to be vaccinated), as it significantly under-represents Māori.

Other methods have also been used to mitigate the numerator/denominator bias for Māori.[[26]] These include the calculation of ratios to adjust for Māori under-count[[6,10,12]] that use similar methods as here to estimate the net under-counting of Māori. More recently, administrative datasets are being linked to create new populations in efforts to improve population estimates by ethnicity, e.g., the experimental Administrative Population Census derived from data in the IDI.[[27]] However, these still rely on the underlying quality of ethnicity data in the linked sources. It is important that Māori have the opportunity to self-identify their ethnicity. This is in line with Indigenous rights, as an expression of rangatiratanga and Māori rights to determine our own identities.[[20]]

There are implications with using ethnicity data to determine an individual’s eligibility for health and disability services. A number of health services already prioritise Māori in the assessment of risk or service delivery.[[14–18]] Using the example of bowel cancer screening, where a lower age (from 50 years) of invitation for Māori has been proposed, a large number of Māori who are eligible would be missed if invitations to screening drew on NHI ethnicity. A smaller number of non-Māori would also be mistakenly invited.

The under-counting of Māori in health data is not new and has been examined in various settings over time. Particularly relevant to our study are other analyses that have linked individual health data with Census data. Stats NZ linked 2013 Census data to a combined NHI dataset and found Māori were under-counted by 21%.[[5]] More recently, the MoH has undertaken similar analyses linking people from the 2018 Census to 2019 NHI data and shown a 16% net under-count of Māori in the NHI compared to Census.[[4]] Our analysis also shows a 16% net under-count of Māori on the NHI from the PHO database compared to individuals’ corresponding Census ethnicity.

There are a number of limitations that should be considered in the interpretation of our findings. Firstly, our analysis of linked data is a subset of Māori in the 2018 Census and 2018 PHO datasets with lower linkage for Māori (based on NHI ethnicity). We cannot determine if the level of mismatch and under-counting of those included in this analysis is the same as those not included. While we examined data by gender and age, we were unable to explore ethnicity data for other important groups, for example those who are gender diverse and/or tāngata whaikaha. Secondly, it is possible that people self-report their ethnicity differently, in different settings, and that they can change their ethnicity.[[28]] However, there is also evidence from health settings that ethnicity data protocols are not always adhered to,[[29]] e.g., guessing someone’s ethnicity based on name or appearance, disagreeing with patient’s self-identified ethnicity, non-standardised forms and data recording. In addition, there is evidence of improvements in ethnicity data collection in some health-related datasets, e.g., mortality,[[6,7]] cancer registrations,[[10]] and examples where there is no (or minor) net under-count of Māori.[[9,30–32]] These findings would suggest that (non)adherence to ethnicity data protocols are the likely drivers of the systematic under-counting of Māori. This is a complex issue to tease out in detail, and we have chosen to focus on implications for Māori using a non-Māori comparator. However, we acknowledge that the non-Māori group is made up of multiple ethnicities combined, and that the quality of ethnicity data for specific ethnic groups, and the mismatch of ethnicity within this large grouping are not examined. In particular, similar data issues and implications are likely to be present for Pacific peoples.[[4,5]] We have also not looked at all potential measures of data quality here such as multiple ethnicities, missing data, and consistency with health and disability sector standards.[[5,33]]

The ongoing systematic under-counting of Māori in health data is a reflection that Māori are not valued and that Māori health and the elimination of inequities are not the priorities they are claimed to be. The ongoing unjust and inequitable healthcare experiences and outcomes faced by Māori require the whole of the health and disability sector to commit to the collection of high-quality ethnicity data, in order to understand inequities; to deliver equitable services to Māori; to act to address inequity; and to monitor progress on eliminating inequities. This commitment needs to come from the highest levels and be attached to accountability mechanisms and ongoing quality assessments.[[34]] Quality ethnicity data are fundamental to the elimination of Māori health inequities. The under-counting of Māori remains a breach of Te Tiriti o Waitangi and Indigenous rights. The ongoing acceptance of poor-quality ethnicity data for Māori and the inadequate progress towards high quality ethnicity data in the health and disability sector is itself evidence of racism as “inaction in the face of need”.[[35]]

View Appendices.

Summary

Abstract

Aim

To examine ethnicity data quality; in particular, the representation and potential under-counting of Māori in health and disability sector data, as well as implications for inequities.

Method

Māori and non-Māori ethnicity data are analysed at: 1) a population aggregate level across multiple 2018 datasets (Estimated Resident Population, Census Usually Resident Population, Health Service User (HSU) population and Primary Health Organisation (PHO) enrolments); and 2) an individual level for those linked in PHO and 2018 Census datasets. Ethnicity is drawn from the National Health Index (NHI) in health datasets and variations by age and gender are explored.

Results

Aggregate analyses show that Māori are considerably under-represented in HSU and PHO data. In linked analysis Māori were under-counted on the NHI by 16%. Under-representation in data and under-counting occur across both genders but are more pronounced for Māori men with variations by age.

Conclusion

High quality ethnicity data are fundamental for understanding and monitoring Māori health and health inequities as well as in the provision of targeted services and interventions that are responsive to Māori aspirations and needs. The continued under-counting of Māori in health and disability sector data is a breach of Te Tiriti o Waitangi and must be addressed with urgency.

Author Information

Ricci Harris: Associate Professor/Public Health Physician, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Sarah-Jane Paine: Senior Lecturer, Te Kupenga Hauora Māori, University of Auckland, New Zealand. June Atkinson: Senior Data Analyst, Department of Public Health, University of Otago, Wellington, New Zealand. Bridget Robson: Associate Professor and Director of Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Paula Toko King: Public Health Physician/Senior Research Fellow, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand. Jennifer Randle: Public Health Medicine Registrar, Department of Public Health, University of Otago, Wellington, New Zealand. Anja Mizdrak: Senior Research Fellow, Department of Public Health, University of Otago, Wellington, New Zealand. Melissa McLeod: Public Health Physician/Senior Lecturer, Te Rōpū Rangahau Hauora a Eru Pōmare, University of Otago, Wellington, New Zealand.

Acknowledgements

The study was funded in part by the grant from the Health Research Council of New Zealand (HRC 20/152). We would like to thank Associate Professor Donna Cormack for comments on a draft of the paper. Data in the paper were provided from Stats New Zealand and the New Zealand Ministry of Health. The Ministry of Health retains intellectual property rights to the data they provided. Stats NZ disclaimer regarding IDI data: The results in this paper are not official statistics, they have been created for research purposes from the Integrated Data Infrastructure (IDI) managed by Statistics New Zealand. The opinions, findings, recommendations and conclusions expressed in this paper are those of the author(s) not Statistics NZ or Ministry of Health. Access to the anonymised data used in this study was provided by Statistics NZ in accordance with security and confidentiality provisions of the Statistics Act 1975. Only people authorised by the Statistics Act 1975 are allowed to see data about a particular person, household, business or organisation and the results in this paper have been confidentialised to protect these groups from identification. Careful consideration has been given to the privacy, security and confidentiality issues associated with using administrative and survey data in the IDI. Further detail can be found in the Privacy impact assessment for the Integrated Data Infrastructure available from www.stats.govt.nz.

Correspondence

Associate Professor Ricci Harris: Te Rōpū Rangahau Hauora a Eru Pōmare, Department of Public Health, University of Otago, Wellington PO Box 7343, Wellington 6242, New Zealand.

Correspondence Email

Ricci.harris@otago.ac.nz

Competing Interests

The authors have no conflicts of interest or financial disclosures.

1) Te Rōpū Rangahau Hauora a Eru Pōmare. Counting for nothing: understanding the issues in monitoring disparities in health. Soc Policy J N Z. 2000;14:1-16.

2) Reid P, Cormack D, Paine SJ. Colonial histories, racism and health—The experience of Māori and Indigenous peoples. Public Health. 2019 Jul;172:119-24.

3) Cormack D, McLeod M. Improving and maintaining quality in ethnicity data collections in the health and disability sector [Internet]. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2010 [cited 2022 Mar 17]. Available from: https://www.otago.ac.nz/wellington/otago600098.pdf.

4) Cleary L. Using ethnicity data in Health Statistics. Wellington: Ministry of Health; 2021. Report No.: 1.1.

5) Reid G, Bycroft C, Gleisner F. Comparison of ethnicity information in administrative data and the census. Wellington: Stats NZ; 2016.

6) Ajwani S, Blakely T, Robson B, Atkinson J, Kiro C. Unlocking the numerator-denominator bias III: adjustment ratios by ethnicity for 1981-1999 mortality data. The New Zealand Census-Mortality Study. N Z Med J. 2003;116(1175):U456.

7) Tan L, Blakely T, Atkinson J. Ethnic counts on mortality and census data 2001-06: New Zealand census-mortality study update. N Z Med J. 2010;123(1320):37-44.

8) Bramley D, Latimer S. The accuracy of ethnicity data in primary care. N Z Med J. 2007 Oct 26;120(1264):U2779.

9) Scott N, Clark H, Kool B, Ameratunga S, Christey G, Cormack D. Audit of ethnicity data in the Waikato Hospital Patient Management System and Trauma Registry: pilot of the Hospital Ethnicity Data Audit Toolkit. N Z Med J. 2018 Oct 5;131(1483):21-9.

10) Shaw C, Atkinson J, Blakely T. (Mis)classification of ethnicity on the New Zealand Cancer Registry: 1981-2004. N Z Med J. 2009 May 8;122(1294):10-22.

11) Swan J, Lillis S, Simmons D. Investigating the accuracy of ethnicity data in New Zealand hospital records: still room for improvement. N Z Med J. 2006 Aug 4;119(1239):U2103.

12) Robson B, Harris R, editors. Hauora: Māori Standards of Health IV. A study of the years 2000–2005. Wellington: Te Rōpū Rangahau Hauora Eru Pōmare; 2007.

13) King P. Māori with lived experience of disability: Part I, Wai 2575, #B22 [Internet]. Waitangi Tribunal; 2019 [cited 2022 May 12]. Available from: https://forms.justice.govt.nz/search/Documents/WT/wt_DOC_150437272/Wai%202575%2C%20B022.pdf.

14) Ministry of Health. Cardiovascular Disease Risk Assessment and Management for Primary Care [Internet]. Wellington: Ministry of Health; 2018 [cited 2022 Mar 18]. Available from: https://www.health.govt.nz/publication/cardiovascular-disease-risk-assessment-and-management-primary-care.

15) Ministry of Health, NZ Society for the Study of Diabetes. Screening for diabetes in asymptomatic adults - Type 2 Diabetes Management Guidelines [Internet]. Ministry of Health; 2022 [cited 2022 May 17]. Available from: https://t2dm.nzssd.org.nz/Section-112-Screening-for-diabetes-in-asymptomatic-adults.

16) National Heart Foundation of New Zealand. Evidence-based, best practice New Zealand Guidelines for Rheumatic Fever [Internet]. Auckland: National Heart Foundation of New Zealand; 2019 [cited 2022 Mar 18]. Available from: https://assets.heartfoundation.org.nz/documents/shop/marketing/non-stock-resources/diagnosis-management-rheumatic-fever-guideline.pdf.

17) BPAC. New diabetes medicines funded: empagliflozin and dulaglutide [Internet]. New Zealand: BPAC; 2021 [cited 2022 Mar 17]. Available from: https://bpac.org.nz/2021/docs/diabetes.pdf.

18) McLeod M, Harris R, Crengle S, Cormack D, Scott N, Robson B. Bowel cancer screening age range for Māori: what is all the fuss about? N Z Med J. 2021 May 21;134(1535):71-77.

19) Kilgour R, Keefe V. Kia Piki Te Ora: The collection of Maori health statistics. Wellington: Department of Health; 1992.

20) Robson B, Reid P. Ethnicity Matters: Review of the Measurement of Ethnicity in Official Statistics Māori perspectives paper for consultation [Internet]. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2001 [cited 2022 Mar 22]. Available from: https://www.otago.ac.nz/healthsciences/otago742784.pdf.

21) Reid P, Robson B. Understanding health inequities. In Robson B, Harris R, editors. Hauora: Māori Standards of Health IV. A study of the years 2000–2005. Wellington: Te Rōpū Rangahau Hauora Eru Pōmare; 2007. p.3-10.

22) Ministry of Health. HISO 10001:2017 Ethnicity Data Protocols [Internet]. Wellington, N.Z.: Ministry of Health; 2017 [cited 2022 Mar 22]. Available from: https://www.health.govt.nz/publication/hiso-100012017-ethnicity-data-protocols.

23) Paine SJ, Cormack D, Reid P, Harris R, Robson B. Kaupapa Māori-informed approaches to support data rights and self-determination 1. In: Indigenous Data Sovereignty and Policy. Routledge; 2020. p.187-203.

24) Statistics New Zealand. Integrated Data Infrastructure (IDI) refresh: linking report Statistical Methods, September 2021 refresh. Wellington: Statistics New Zealand; 2021.

25) Ministry of Health. COVID-19: Vaccine data. Spreadsheet covid_vaccinations_15_03_2022 [Internet]. Ministry of Health; 2022 [cited 2022 Mar 17]. Available from: https://www.health.govt.nz/covid-19-novel-coronavirus/covid-19-data-and-statistics/covid-19-vaccine-data.

26) Cormack D, Harris R. Issues in monitoring Māori health and ethnic disparities: an update. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2009.

27) Statistics New Zealand. Experimental administrative population census [Internet]. Statistics New Zealand; 2021 [cited 2022 Apr 4]. Available from: https://www.stats.govt.nz/experimental/experimental-administrative-population-census.

28) Carter KN, Hayward M, Blakely T, Shaw C. How much and for whom does self-identified ethnicity change over time in New Zealand? Results from a longitudinal study. Soc Policy J N Z. 2009;36:32-45.

29) Neuwelt P, Crengle S, Cormack D, McLeod M, Bramley D. General practice ethnicity data: evaluation of a tool. J Prim Health Care. 2014;6(1):49-55.

30) Randle J, Inns S, Harris R, McLeod M. Ethnicity data audit in a secondary care gastroenterology service. N Z Med J. 2022 Nov 11;15(1565):51-59.

31) Page M, Wyeth EH, Samaranayaka A, McNoe B, Walker R, Schollum J, et al. Accuracy of ethnicity data recorded in hospital-based patient clinical records and the Australia and New Zealand dialysis and transplant registry. N Z Med J. 2017 Apr 28;130(1454):65-71.

32) Riddell T, Lindsay G, Kenealy T, Jackson R, Crengle S, Bramley D, et al. The accuracy of ethnicity data in primary care and its impact on cardiovascular risk assessment and management–PREDICT CVD-8. N Z Med J. 2008 Sep 5;121(1281):40-8.

33) Ministry of Health. Primary Care Ethnicity Data Audit Toolkit [Internet]. Wellington: Mininstry of Health; 2021 [cited 2022 Mar 22]. Available from: https://www.health.govt.nz/system/files/documents/publications/primary-care-ethnicity-data-audit-toolkit-16dec21.pdf.

34) Curtis E, Loring B, Harris R, McLeod M, Mills C, Scott N, et al. Māori Health Priorities: A report commissioned by the interim Māori Health Authority (iMHA) to inform development of the interim New Zealand Health Plan (iNZHP). Wellington: interim Māori Health Authority; 2022.

35) Jones CP. Confronting institutionalized racism. Phylon (1960-). 2002;50:7-22.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.
Login, subscribe or email nzmj@nzma.org.nz to purchase this article.

LOGINSUBSCRIBE