Transgender Health: New Zealand’s Innovative Statistical Standard for Gender Identity

Abstract

The implementation of the New Zealand government’s recently developed statistical standard for gender identity has led to, and will stimulate further, collection of gender identity data in administrative records, population surveys, and perhaps the census. This will provide important information about the demographics, health service use, and health outcomes of transgender populations to allow evidence-based policy development and service planning.

However, the standard does not promote the two-question method, risking misclassification and undercounts; does promote the use of the ambiguous response category “gender diverse” in standard questions; and is not intersex inclusive.

Nevertheless, the statistical standard provides a first model for other countries and international organizations, including United Nations agencies, interested in policy tools for improving transgender people’s health.

The 193 United Nations (UN) member states have pledged to “leave no one behind” in the 2030 Sustainable Development Agenda,¹ and this includes them not leaving transgender people behind in realizing goal 3: “to ensure healthy lives and promote well-being for all.” Official data on gender identity that are of high quality (i.e., valid, reliable, accurate, and acceptable to respondents) are crucial for transformative health research, policy, and action aimed at addressing the health disadvantage faced by transgender people. Policy tools for improving the collection of official gender identity data are urgently required.

We describe and critically analyze a novel such policy tool produced by the New Zealand government: a national statistical standard for gender identity data collection. To our knowledge, this is the first of its kind globally. The Australian government is currently also developing a statistical standard for gender identity,^2,3 and transgender health experts have advised the US Department of Health and Human Services to also standardize their official gender identity data collections.⁴ Moreover, UN data initiatives for improving transgender health (e.g., the UN Development Program’s LGBTI Inclusion Index)⁵ will require globally harmonized gender identity data.

THE HEALTH OF TRANSGENDER POPULATIONS

Gender identity is “each person’s deeply felt internal and individual experience of gender, which may or may not correspond with the sex assigned at birth.”^6(p6) People whose sex assigned at birth does not correspond fully with their gender identity (e.g., an assigned female person who identifies as a man) are commonly referred to as “transgender.” People whose sex assigned at birth and gender identity correspond are called “cisgender.” Although a lack of gender identity data from large-scale population surveys and censuses has prevented definitive prevalence and other demographic estimates for transgender populations, the most robust estimates suggest that 0.3% to 0.5% of people identify as transgender.⁷

Compared with cisgender populations, transgender populations face considerable health disadvantage, especially in the domains of mental health, sexual health, health care access, discrimination, and violence.⁷ For example, several studies have highlighted very high rates of suicide attempts and HIV infection among transgender populations.⁷ Transgender people also face barriers when accessing transition-related and -unrelated health care, including discrimination and mistreatment from health practitioners and lack of funding.⁷ They are the target of discrimination in education, employment, and housing⁸ as well as gender-based violence, including murder.⁹ These adverse health outcomes are particularly marked among transgender people who are also members of other marginalized groups, such as ethnic minorities or low-income groups.⁷

With transgender populations’ considerable health disadvantage increasingly being recognized, gender identity is coming to be understood as a social determinant of health.¹⁰ Health policymakers are now motivated to improve transgender people’s health. Consequently, national governments and international organizations such as the US Department of Health and Human Services and the UN Programme on HIV/AIDS have now dedicated policies and action plans for this purpose.^11,12 These initiatives generally acknowledge transgender people’s human rights to freedom from discrimination and to health.^7,13

OFFICIAL DATA ON GENDER IDENTITY

High-quality gender identity data disrupt the familiar mantra of “no data, no problem, no action.”^14(p858) They can be used to determine the population prevalence and other demographics of transgender populations and to monitor their health service use and health outcomes. This evidence can in turn be used to effectively design, plan, implement, and evaluate policies and services for improving transgender health. Collecting high-quality official gender identity data has consequently become a priority for transgender health research.¹⁵

Official gender identity data collection is technically feasible.¹⁶ Researchers have developed and validated standard survey questions¹⁷ and data collection guidelines.¹⁸ Current best practice includes using the two-question method of collecting both sex assigned at birth and current gender identity (e.g., sex assigned at birth reported as “female” and gender identity reported as “male”). The one-question method of collecting just gender identity cannot identify transgender people who strongly identify along the gender binary as “male” or “female” and therefore only select one of these two categories, as opposed to selecting the transgender category.¹⁹ This method therefore misses this transgender subpopulation, likely leading to considerable misclassification and undercounts of the total transgender population.²⁰

Although there may be concerns over the feasibility, quality, and acceptability of official gender identity data collection, analogous concerns over official sexual orientation data collection proved to be unfounded once implemented.^21,22 Moreover, gender identity data collection is accepted among the wider community. For example, nearly all health care center attendees interviewed in a recent study were willing to answer gender identity questions and considered these questions important and acceptable.²³ Several health centers in the United States already successfully collect gender identity data.²³

A few potential technical challenges do remain in question development. These include ensuring that gender identity response categories are understood by cisgender individuals while remaining relevant to diverse transgender people as well as the need to use the two-question method (i.e., to collect both sex assigned at birth and current gender identity),²⁰ to allow write-in response categories, and to permit multiple responses to precisely capture gender identity. The geographic heterogeneity and ongoing evolution of gender identity terminology challenge the development of international standard questions.

International organizations (including the UN Programme on HIV/AIDS, the UN Development Program, and the World Health Organization), national governments, health service providers, human rights institutions, and several transgender individuals and groups have identified the need for high-quality official gender identity data to improve transgender people’s health.^5,16,24,25 Some national governments are required to collect gender identity data to fulfill human rights obligations.²⁶ Yet, despite the urgent need for and technical feasibility of gender identity data, international organizations and national governments still rarely if ever collect, analyze, or report these data.

THE STATISTICAL STANDARD FOR GENDER IDENTITY

New Zealand is a country where transgender communities’ and human rights institutions’ advocacy has resulted in some progressive government policies supporting transgender people’s health. Milestone public policies include implied legal protection from discrimination on the basis of gender identity (since 2006),²⁷ a Human Rights Commission inquiry into discrimination experienced by transgender people and the implementation of the inquiry’s recommendations into action (since 2008),²⁸ Ministry of Health guidelines for gender-affirming health care (2011),²⁹ and discretionary government funding for three male-to-female and one female-to-male gender-affirming surgeries every two years (since 2005; personal communication, David St George, Ministry of Health, November 5, 2015).

The development of New Zealand’s Statistical Standard for Gender Identity²⁶ (hereafter the “gender identity statistical standard”) began in 1995 when New Zealand’s Statistical Standard for Sex³⁰ noted that measuring the diversity and complexity of gender identity concepts required “new and separate definitions, questions, classifications, and statistical standards.”^(p1) In 2010, transgender communities expressed their concern to Statistics New Zealand (the lead government department for official statistics) that the Statistical Standard for Sex³⁰ required them to respond on the basis of their sex assigned at birth, not their gender identity.³¹

Statistics New Zealand responded in 2011 by revising this statistical standard, providing dedicated guidelines that enabled respondents to be classified on the basis of their gender identity, at least after having commenced gender transition.³¹ The department then started considering the development of a statistical standard for gender identity in 2012, publishing a discussion article³² and conducting community consultations in 2014. It finally released the gender identity statistical standard on July 17, 2015, and scheduled a review of the standard for 2018.²⁶

Structure and Content

A statistical standard is “a set of rules used to standardize the way data are collected and statistics are produced.”³³ Statistical standards provide the required information about data collected on a particular topic to aid understanding and interpretation of the data among data producers and consumers. These standards present the officially approved versions of how to “define the underlying concept; define the specific variables of interest; collect data (coding structure, statistical units, recommended question modules); process the data (classification, standard editing); present the data (output categories); [and] interpret the data.”³³ The gender identity statistical standard provides such conceptual, measurement, data collection, analysis, and reporting frameworks for gender identity.

Conceptual framework.

The gender identity statistical standard provides the rationale for its development, defines and conceptualizes key measurement concepts, and discusses key operational issues. It was developed to standardize definitions and measurements of gender identity across the government, to meet human rights requirements,³⁴ and to enable policy and service development for transgender people. It adopts the definitions presented in the box on this page for the key measurement concepts of “gender identity” and “gender diverse.”

Definitions of Gender Identity and Gender Diverse

Concept	Definition
Gender identity	Gender identity is an individual’s internal sense of being wholly female, wholly male, or having aspects of female and male. Gender identity is understood to refer to each person’s deeply felt internal and individual experience of gender, which may or may not correspond with the gender recorded at birth. A person’s gender identity can change over the lifetime and can be expressed in numerous ways and forms. This expression includes outward social markers, such as name, clothing, hairstyles, mannerisms, voice, and other behavior.
Gender diverse	Gender diverse is having a gender identity or gender expression that differs from a particular society’s dominant gender roles.

Source. Statistics New Zealand.^26(p1)

The standard further conceptualizes gender identity as changeable across social context and over time and as a self-defined concept that can be self-reported from adolescence onward and collected from all people, including via proxy. Importantly, it states that the two-question method²⁰ for gender identity data collection is required under certain circumstances only if there is a reason to do so. Thus, it implies that the one-question method is standard practice, but this recommendation diverges from the international best practice³⁵ of always using the two-question method.²⁰

Measurement framework.

The measurement framework presents the mandatory question requirements and examples of standard questions. Question requirements are that any gender identity question must be suitable for self-completion, be appropriate for all gender identities, and elicit a single response at minimum (but allowing multiple responses).

The gender identity statistical standard also notes preferences for questions that include the term “gender” together with “identity” or “identify” and that have a write-in response category that enables a full description of a respondent’s gender identity or identities. It also flags that some respondents may not yet know the relatively new term “gender diverse,” the main response category for transgender people. Four example questions that adhere to these requirements are provided, including “What is your gender identity,” with the three response categories of “male,” “female,” and “gender diverse, please state.”

Data collection, analysis, and reporting.

Data collection guidelines, the standard classification, and the standard output are also provided. The guidelines emphasize several crucial data collection issues. Key elements include that survey respondents should on request receive information on the need for gender identity data and the privacy and confidentiality requirements of these data and that they should have the right to refuse a response. When using the two-question method,²⁰ the standard recommends that the gender identity question follow the question on sex assigned at birth. However, the ordering of questions on gender identity and sex assigned at birth remains controversial, and studies that randomize the ordering are required to determine best practice. Because respondents may have multiple gender identities, questions must enable a multiple response. The standard also recommends that gender identity data not be imputed.

Gender identity categories are classified into the two-level hierarchical classification presented in the box on the next page. The broadest classification level includes the three categories of “female,” “male,” and “gender diverse.” The more detailed classification level includes six categories, disaggregating the “gender diverse” category into four subcategories. All categories also include the indigenous Māori language terms and concepts.

BOX 2—Standard Classification for Gender Identity

1. Male

11. Male, or Tāne

2. Female

21. Female, or Wahine

3. Gender diverse

30. Gender diverse not further defined, or Ira tāngata kōwhiri kore

31. Transgender male to female, or Whakawahine

32. Transgender female to male, or Tāngata ira tāne

39. Gender diverse not elsewhere classified, or Ira tāngata kōwhiri kore

Source. Statistics New Zealand.^26(p4)

Limitations of the Standard

We have identified three main limitations of the gender identity statistical standard. That the standard’s development process and supporting evidence base were not published prevented the identification of what caused these limitations. The standard’s first limitation is that it encourages the one-question method of collecting gender identity data, as opposed to following international best practice³⁵ of always using the two-question method.²⁰ This risks potentially serious misclassification and undercounts of the transgender population, especially in health settings.

Second, in the proposed standard gender identity questions, having only one response category that uses a relatively new identity term (“gender diverse”) for various gender identities may have some limitations. It may obscure some of the complexity within the broader transgender population. For example, important differences in health service needs and health outcomes may exist between, for example, transgender and genderqueer participants, and at the first level of classification, the standard will not enable health researchers to differentiate these identities.

Even the more nuanced second classification level has the limitation of the “gender diverse not elsewhere classified” category encompassing diverse identities, including non-Western identities (e.g., the Samoan fa’afafine) and nonbinary identities (e.g., genderqueer). The term “gender diverse” is new, will be unfamiliar to many people, and has not yet been robustly tested cognitively. We do not know the extent to which diverse transgender people will endorse this category or to which cisgender people may endorse it (e.g., because they do not fully understand it).

Third, although the gender identity statistical standard defines the term “intersex” and although there are recommendations³ for making statistical standards for gender identity include intersex people, the standard does not provide guidance on being intersex inclusive in gender identity data collection, analysis, and reporting (e.g., “gender diverse” is likely to not be intersex inclusive).

These three major limitations to some degree reduce the validity, reliability, accuracy, and acceptability of gender identity data collected using the gender identity statistical standard. However, all statistical standards have some limitations, and such limitations generally do not preclude data collection. For example, the limits of statistical standards for ethnicity are very similar to those of the gender identity statistical standard, and the former statistical standards are applied widely, including in New Zealand and the United States. Nonetheless, Statistics New Zealand should address these limitations in the gender identity statistical standard’s general review in 2018, and equivalent national and international statistical standards should avoid these limitations.

Implications for Research, Policy, and Action

Limitations notwithstanding, that New Zealand has developed the first national statistical standard for gender identity has several implications for New Zealand, other countries, and international organizations. Implementation of the standard in New Zealand’s health records, population surveys, and the census will improve national population prevalence and other demographic estimates and health services and outcomes monitoring, and this evidence will enable the design, cost, implementation, and evaluation of health policies and services for transgender populations.

Although standard implementation is voluntary, Statistics New Zealand is testing the inclusion of a gender identity question in the 2018 New Zealand Census of Population and Dwellings, the Ministry of Youth Development has started requiring its funded providers to report the number of youths accessing their services who identify as “gender diverse,” and the Ministries of Health and Education are planning to implement the standard (personal communication, Alana Lattimore, Statistics New Zealand, June 30, 2016). New Zealand’s and the United States’ experiences demonstrate that statistical standards for similar social identities (e.g., ethnicity) have effectively improved the quality of similar official data.

We predict that the standard’s implementation will catapult New Zealand toward becoming a leader in transgender health data and research. For example, gender identity data from official New Zealand probability surveys can be linked with health data (e.g., cancer registrations) to generate urgently needed⁷ transgender cohort studies.

New Zealand has already shared the gender identity statistical standard with the UN Expert Group on International Statistical Classifications³⁶ and is collaborating bilaterally with other countries (e.g., Australia) on gender identity data collection (personal communication, Alana Lattimore, Statistics New Zealand, July 1, 2016). For countries and international organizations interested in developing national or international statistical standards for gender identity (e.g., the United States, Australia, and UN agencies), the gender identity statistical standard signals that national and international efforts to standardize gender identity data can be fruitful. It also provides a first complete national statistical standard for gender identity to build on, although the limitations we have described must be rigorously tackled to provide robust technical guidance for collecting, analyzing, and reporting gender identity data. Producers of official statistics should be guided in any efforts to develop national or international statistical standards for gender identity data collection by transgender health experts and communities.