Common mental disorders and psychological distress among Pacific adults living in Aotearoa New Zealand

“Pacific peoples” is a heterogeneous population in Aotearoa New Zealand, composed of individuals whose homelands and ancestral connections stem from various Pacific nations, predominantly from Melanesia and Polynesia.¹ Pacific peoples make up almost 9% of the total New Zealand population and are predicted to increase to 11% of the total New Zealand population by 2043.² The Pacific population has a young age structure (median age: 23.4 years) and is highly concentrated in urban regions, with three-quarters of Pacific living in the Auckland and Wellington Regions in 2018.³ Pacific peoples is an ethnically and culturally diverse population, which includes at least 17 different Pacific ethnic groups. The six largest groups are Samoan (48%), Tongan (22%), Cook Islands Māori (21%), Niuean (8%), Fijian (5%) and Tokelauan (2%).^3,4 The Cook Islands, Niue and Tokelau fall within the “Realm of New Zealand”, as these Pacific nations were annexed by the British monarch in the early 1900s and colonised under the New Zealand flag. Individuals affiliated with these “Pacific Realm countries” are New Zealand citizens by birth,⁵ and these Pacific subpopulations have greater proportions of people living in New Zealand than in their Pacific homelands, with over 90% of the Cook Islands and Niue populations currently living in New Zealand.⁴ Importantly, the Pacific Realm countries share similar histories of colonisation and migration to New Zealand, both of which contribute to the shaping of mental health inequities in cultural populations through the intergenerational disruption of health-protective social structures and cultural resources.⁶

Pacific peoples experience high rates of mental health conditions and disproportionately low rates of mental health service use. Te Rau Hinengaro (The New Zealand Mental Health Survey) is the only population mental health survey to date that measured mental health with structured clinical interviews, based on assessment criteria for mental disorders in the Diagnostic and Statistical Manual of Mental Disorders (DSM). This survey, conducted in 2003–2004, showed that Pacific adults aged 16 years and over had a higher prevalence of mental disorders than Others (non-Māori, non-Pacific) and low rates of mental health service use.⁷ While elevated crude rates of mental disorders were explained by the young population structure and higher levels of socio-economic deprivation,⁷ Pacific peoples’ underutilisation of mental health care was only partly explained by these, suggesting other factors, such as known care access barriers, underpin Pacific peoples’ mental health service use.^8,9

An analysis of New Zealand Health Survey (NZHS) data from 2014 to 2019 demonstrated that Pacific adults have lower rates of doctor-diagnosed mood and/or anxiety disorders despite higher rates of psychological distress than non-Māori, non-Pacific adults.¹⁰ This same pattern was also demonstrated in the New Zealand Attitudes and Values Study (NZAVS),¹¹ although these surveys are limited by their use of self-report measures of diagnosed disorders, which contrasts with the systematic measurement of diagnosable mental disorders in Te Rau Hinengaro.

The predominant focus of national research and policy involves examining and addressing mental health inequities in the total Pacific population. This pan-Pacific approach is understandable given the way Pacific peoples formed collectively in New Zealand as a result of shared cultural customs, values and beliefs and similar migration histories within the country. However, there is growing evidence that mental health varies within Pacific peoples, particularly in various subpopulations defined by socio-demographic factors, such as Pacific born in New Zealand (“NZ-born”), or those who affiliate with more than one ethnic group (“multi-ethnic”). Nationally representative cross-sectional surveys,¹⁰ including Te Rau Hinengaro,¹² have consistently demonstrated that mental disorders are more prevalent in Pacific peoples born in New Zealand (“NZ-born”) compared with those born overseas (“Overseas-born”). Similarly, evidence from nationally representative cross-sectional surveys show that multi-ethnic Pacific adults have poorer mental health than sole-Pacific adults, with lower self-esteem¹³ and higher rates of personal experiences with mental illness¹⁴ and diagnosed mood and/or anxiety disorders.¹⁰

Examining mental health within Pacific peoples is challenging due to the relatively small size of the population and even smaller sizes of Pacific subpopulations of interest. Specialised statistical techniques, such as Bayesian modelling, are one way of examining subpopulation outcomes, and these methods have previously been used to compare both health and mental health outcomes between Pacific ethnic groups.^15,16 Pooling data across multiple waves of a routine national survey is another method for looking more closely at mental health in Pacific subpopulations. This paper aims to use pooled NZHS data to examine mental health within Pacific adults, by socio-demographic factors and the specific Pacific ethnic groups. Such within-Pacific analyses are useful for identifying subpopulations with higher needs and informing the development of targeted mental health promotion policy and initiatives.

Methods

Study design

This study used data from the NZHS—an annual cross-sectional survey administered by the Ministry of Health – Manatū Hauora. Anonymised NZHS data in the form of confidentialised unit record files (CURFs) were sourced from Stats NZ by standard microdata access application processes. Data were supplied in January 2023.

Participants

The NZHS uses a complex multi-stage sampling approach with a stratified probability-proportional-to-size design to obtain a nationally representative sample. Participants are selected, first by area-based selection of primary sampling units (PSUs) from Stats NZ’s household survey frame, then by selecting households within the PSUs and randomly selecting eligible participants from those households. All survey waves in this study used PSUs for first-stage selection, except for the 2014/2015 wave, which used meshblocks—smaller areas (around 90 people), one or more of which fit into a single PSU. Eligible participants include adults aged 15 years and over, and children aged 0–14 years. Targeted sampling of PSUs where more Pacific people live is undertaken to increase Pacific sample sizes.¹⁷ Only NZHS adult participants were included in this study (N=68,317; mean age=45.31 years, SD=19.03). Overall response rates ranged from 79–80% across survey waves.

Procedures

The NZHS uses trained interviewers to conduct face-to-face computer-assisted interviews from 1 July to 30 June of the following calendar year. Information is collected across various health domains, including long-term conditions, health status, health behaviours and risk factors and health service utilisation and barriers.¹⁶

Measures

Mental health outcomes

Two mental health outcome measures were included in this study: doctor-diagnosed common mental disorders and psychological distress. “Common mental disorders” is a term referring to depressive disorders and anxiety disorders, which are highly prevalent and considered “common” in global populations.¹⁸ Doctor-diagnosed common mental disorders were measured with a composite binary (yes/no) variable, derived from two questions: “Have you ever been told by a doctor that you have depression?” and “Have you ever been told by a doctor that you have an anxiety disorder? This includes panic attacks, phobia, post-traumatic stress disorder and obsessive-compulsive disorder.”

Psychological distress was measured with the Kessler Psychological Distress Scale (K10).¹⁹ The K10 is an internationally validated 10-item scale that measures how often respondents experienced symptoms of psychological distress (e.g., hopelessness, nervousness, restlessness, depression or worthlessness) over the past 4 weeks. Responses are recorded on 5-point scales (“none of the time”, “a little of the time”, “some of the time”, “most of the time” and “all of the time”) and scored 0–4 for each item. Total scores range from 0 to 40, with higher total scores indicating higher psychological distress. In this study, “high psychological distress” was defined as a score total of 12 or more—the cut-off point used in NZHS reporting. K10 scores of 12 or more indicate a high probability of depression or anxiety disorder.²⁰ The K10 has good predictive validity and, for example, was shown to discriminate between cases and non-cases of anxiety and mood disorders in Te Rau Hinengaro.²¹ However, the K10 is not a diagnostic tool, and is therefore not recommended as a measure of the prevalence of mental health conditions in New Zealand population groups.²²

Ethnicity

The NZHS uses the standard ethnicity question from the New Zealand Census of Population and Dwellings. Respondents identify the ethnic group/s they belong to, with a number of options provided, including an “Other” category and accompanying free-text field for written responses. This study used total response ethnicity classification, meaning participants who identified with Pacific as any of their ethnic groups were included in the Pacific ethnic group for analysis.

Level 2 ethnicity data were supplied for the four largest Pacific ethnic groups (Samoan, Cook Islands Māori, Tongan and Niuean). These were used to group participants into four Pacific ethnic groups. Niuean data were not disaggregated because of small strata sizes within each mental health outcome (n=45, common mental disorders; n=36, psychological distress). A residual “Other Pacific” category included Niueans (n=381) and all remaining Pacific participants for whom Level 2 ethnicity data were not supplied (n=481). Participants who self-identified with two or more Pacific ethnic groups (n=225) were preferentially assigned to the smaller Pacific group, creating mutually exclusive Pacific ethnic groups.

Cook Islands Māori and Niueans were combined into a “Realm” Pacific group for analysis by Realm country status. Tokelauans are also part of the Realm of New Zealand, but were assigned to the “non-Realm” group because Tokelauan ethnicity data were not included in CURFs. “Tokelauan” is not a response option in the standard NZHS ethnicity question, and free-text responses associated with the “Other” ethnic group responses were not supplied upon request due to confidentiality concerns and limited NZHS team capacity to provide customised CURFs.

Other covariates

Other covariates entered into the models included age (15–24, 25–44, 45–64 and 65+ years) and sex (male/female). Socio-economic position (SEP) was measured with both an area-based measure (New Zealand Index of Deprivation 2013 [NZDep2013])²³ and an individual-level measure (highest educational qualification attained). NZDep2013 data were collapsed into three categories: low (NZDep deciles 1–3), medium (deciles 4–7) and high deprivation (deciles 8–10). Highest educational qualification data were used to create a binary variable (no secondary qualification vs secondary qualification or higher).

Analysis

Data from the 2014/2015 to 2018/2019 NZHS waves were pooled into a single dataset and analysed using Stata/SE version 18.0. The complex sampling method was accounted for by applying supplied calibrated survey weights, which adjust for inverse sampling weighting, stratification and clustering due to area-based sampling. Supplied jackknife replicate weights were used to estimate variance. All survey weights were adjusted by dividing by the total number of survey waves included in the study. Unweighted frequencies and weighted prevalence estimates were calculated for mental health outcomes in Pacific adults, analysed by socio-demographic factors (age, sex, deprivation and education), Pacific ethnicity (Samoan, Tongan, Cook Islands Māori and Other Pacific) and Realm country status (Realm/non-Realm). Unadjusted and adjusted risk ratios for mental health outcomes and associated 95% confidence intervals (CIs) were calculated using generalised linear regression, specifying a binomial distribution with a log link. Outcomes and covariates were entered into models in a stepwise process: 1) unadjusted, 2) adjusted for age and sex, and 3) adjusted for SEP, entering NZDep and educational qualification sequentially. Post-estimation adjusted Wald tests were run to test for differences between the levels of socio-demographic factors and Pacific ethnic groups.

Results

There were 4,335 Pacific adults aged 15 years and over in the NZHS dataset, pooled from the 2014/2015 to 2018/2019 survey waves (mean age=31.17 years; SD=17.05). Table 1 summarises the socio-demographic characteristics of the participants.

Five-year period prevalence estimates of doctor-diagnosed common mental disorders and psychological distress during 2014–2019 are presented in Table 2. Results showed the prevalence of diagnosed common mental disorders was significantly lower in Pacific males (6.2%) than females (10.9%, ARR=0.57; 95% CI 0.46–0.71). The prevalence of diagnosed common mental disorders was significantly higher in Pacific adults from low- and medium-deprivation areas (ARR_Low=1.77; 95% CI 1.19–2.65; ARR_Medium=1.43; 95% CI 1.11–1.83) than in Pacific adults from high-deprivation areas and was significantly lower in Pacific youth aged 15–24 years (5.5%) than Pacific adults aged 25–64 years (25–44 years, p=0.002; 45–64 years, p=0.003). Table 2 also demonstrates the prevalence of psychological distress was lower in Pacific males (9.3%) than Pacific females (13.2%, ARR=0.70, 0.57–0.86). There were no other significant differences in psychological distress observed by sex, age, deprivation or education.

View Table 1–3.

Table 3 shows that, among the Pacific ethnic groups, the prevalence of diagnosed common mental disorders was higher in Cook Islands Māori (12.1%) than both Samoans (7.2%, ARR=1.67; 95% CI 1.28–2.17) and Tongans (6.5%, p=0.0031; Appendix Table 1). Diagnosed common mental disorders were also significantly more common in Other Pacific (11.4%) than in Samoans (ARR=1.43; 95% CI 1.04–1.98) and Tongans (p=0.037; Appendix Table 1). When analysed by Realm country status, the risk of diagnosed common mental disorders was over 60% higher in Pacific adults affiliated with Realm countries compared with those affiliated with non-Realm countries (ARR=1.65, 95% CI 1.30–2.08). The prevalence of psychological distress was higher in Cook Islands Māori (14.2%) than Samoans (10.4%, ARR=1.41; 95% CI 1.06–1.89), but no other differences between Pacific ethnic groups reached statistical significance in post-tests (Appendix Table 1). Full regression model results are appended (Appendix Table 2). 

Discussion

This study showed that mental health outcomes vary within Pacific peoples, with Cook Islands Māori adults reporting significantly higher rates of doctor-diagnosed common mental disorders than Samoans and Tongans. This is consistent with results from Te Rau Hinengaro, which demonstrated a general patterning of a higher prevalence of mental disorders in Cook Islands Māori adults, although results did not reach statistical significance.⁷ A related finding was that the risk of common mental disorders was collectively 65% higher in Pacific ethnic groups affiliated with nations that fall under the “Realm of New Zealand” (Cook Islands and Niue in this paper; see Methods section). To our knowledge, no other studies have demonstrated variability in diagnosed mental disorders by Realm country status, although one previous study of the NZHS demonstrated hazardous drinking was significantly higher in females from Pacific Realm countries,²⁴ and the cross-sectional national Youth Insights Survey showed past-month substance use was significantly higher in Pacific youth affiliated with the Realm countries.²⁵ More broadly, the New Zealand Census Mortality Study demonstrated that cardiovascular disease mortality rates were significantly higher in Cook Islands Māori and Niueans in 2001–2004.¹⁵ This consistent patterning of poorer health and mental health outcomes in Cook Islands Māori and Niueans indicates an urgent need for research and policy interventions focussing on these Pacific subpopulations.

This study demonstrates the period prevalence of diagnosed common mental disorders was significantly higher in Pacific adults living in low- and medium-deprivation areas. This is consistent with national specialist mental health service use data that show anxiety and depression diagnoses decrease with deprivation in Pacific young people aged 10–24 years.²⁶ This patterning of Pacific mental health by deprivation that differs from what is seen in the total New Zealand population, whereby, for the latter, poorer outcomes are generally observed in more deprived areas.²⁷

There are multiple possible interpretations for these within-Pacific differences in diagnosed mental disorders, and this highlights the limitations of self-report measures of diagnosed disorders. It is possible there are financial barriers for Pacific adults, preventing access to services in those with lower SEP; or, similarly, that the citizenship status of Realm-Pacific individuals affords them easier access to primary and specialist services. It is also possible that Realm Pacific adults or those from low deprivation areas have different concepts of mental health/illness and, therefore, have different help-seeking behaviours and care preferences. In short, the outcome measure conflates both diagnosable and diagnosed mental conditions and, therefore, is not as useful for directing policy as measures derived from structured clinical interviews (e.g., composite international diagnostic interview [CIDI] used in Te Rau Hinengaro), which are more indicative of need. This limitation underscores a need for more frequent routine monitoring of national mental health—with robust measures of diagnosable mental conditions—to enable policymakers to accurately differentiate between mental health needs and service access and track mental health outcomes over time.

We found the elevated risk of diagnosed common mental disorders in Realm Pacific ethnic groups was stable and did not change appreciably after adjusting for demographic or socio-economic factors (Appendix Table 2). This pattern was also demonstrated in an NZHS study on doctor-diagnosed mood and anxiety disorders among New Zealand-born and multi-ethnic Pacific adult subpopulations.¹⁰ Together, these results suggest other causal factors unaccounted for in the models should be considered, highlighting a need to look more broadly at the structural causes of Pacific mental health. Structural causal factors are the underlying causes of mental health that occur in an individual’s surrounding socio-cultural, geopolitical and physical environments.²⁸ These factors shape the circumstances we grow, work, live and age in—the “social determinants of mental health” (e.g., housing, education and employment/income)—which influence our mental health over the life course and across generations.²⁹ For global Indigenous populations, the legacy of colonisation is an important structural factor²⁸ because the impacts of dispossession of ancestral lands, institutional racism, forced assimilation and subjugation through legislative and social policies are intergenerational, and these continue to influence the multiple social determinants that contribute to present-day health inequities.⁶ The higher rate of diagnosed common mental disorders in Realm Pacific observed in this paper, for example, could be explained by lower rates of cultural connectedness among Realm Pacific adults compared with non-Realm Pacific adults,¹⁴ which itself is rooted in post-colonial social policies (e.g., English-only language schools) that led to the gradual decline of Indigenous Pacific health-protective cultural resources (e.g., Realm Pacific languages) over successive generations. While research is needed to understand how cultural connectedness is associated with Pacific mental health, it is likely the within-Pacific mental health differences in this study are underpinned by these broader structural and social determinants of mental health. Indeed, a small body of national cross-sectional and longitudinal studies demonstrate that factors such as migration, acculturation and racism are all associated with Pacific mental health.⁹ Similarly, nationally representative cross-sectional studies demonstrate that social determinants, such as ethnic identity³⁰ and Pacific language,³¹ are associated with better mental health outcomes and, therefore, would be worth considering for clinical- and population-level intervention.

This study showed the prevalence of both psychological distress and doctor-diagnosed common mental disorders was higher in Pacific females than males. This latter result differs from Te Rau Hinengaro, which found no significant sex differences in Pacific adults who met diagnostic criteria for mental disorders or serious mental disorders experienced in the past 12 months.⁷ Given Te Rau Hinengaro used a more comprehensive measure of mental disorders, our result could be reflecting a broad range of factors, such as unconscious gender biases, or male–female differences in self-reporting of mental disorders and/or presentation to services.

The results on psychological distress within Pacific adults are somewhat inconclusive. This is first because psychological distress did not vary significantly by age, deprivation or education, all of which are covariates with known associations with psychological distress.²⁷ Second, psychological distress did not differ by Realm country status. Third, the wide errors around the Pacific ethnic group estimates mean the finding of higher psychological distress in Cook Islands Māori than Samoans should be interpreted with caution, particularly since Cook Islands Māori did not differ from either Tongans or Other Pacific in post-testing (Appendix Table 1). Nevertheless, together these results raise two points about 1) the validity of the K10 and its suitability for measuring acute mental distress in Pacific peoples, and 2) the likelihood that Pacific mental health is driven largely by broader social determinants, which have a cumulative and pervasive impact on mental wellbeing, operating over years and generations, and, therefore, would show up more strongly in long-term measures than short-term measures of distress. These points are supported by a recent NZHS analysis,¹⁰ which demonstrated psychological distress did not differ between New Zealand-born/Overseas-born and multi-ethnic/sole-Pacific subpopulations—results that are incongruous with cross-sectional national surveys that show New Zealand-born and multi-ethnic Pacific sub-groups have significantly higher rates of assessed mental disorders⁷ and doctor-diagnosed mental health conditions.¹⁴ Further research is needed to better understand within-Pacific mental health differences and to examine the validity of current measures used to monitor Pacific mental health.

Strengths and limitations

A key strength of this study is its acknowledgment that Pacific peoples is made up of several (sub)populations, each with different socio-cultural influences on mental health that contribute to diverse outcomes. The use of pooled data from multiple waves of a routine nationally representative survey enabled closer examination of the patterning of mental health in Pacific sub-groups with greater precision than is possible with single survey waves. To our knowledge, this is also one of the first studies to examine mental health outcomes by Pacific Realm country status—a Pacific subpopulation of emerging importance in Pacific health research.^14,24

This study is limited first by the measures of mental health used in the NZHS, particularly self-reported doctor-diagnosed disorders, which may not reflect the true burden of Pacific mental health because it does not separate out mental health needs from access (i.e., it is difficult to ascertain whether apparent increased risk is due to increased burden or better care access). Second, by combining multiple survey waves, this study assumes mental health outcomes are stable across time. Third, not having access to Level 2 ethnicity data, and the consequent inability to identify Tokelauans for analysis, may have introduced misclassification bias, potentially producing underestimated effect sizes in the analysis by Realm country status. However, it is worth noting that Tokelau has a different governance structure within the Realm of New Zealand and was annexed by the Crown slightly later than the Cook Islands and Niue;⁵ therefore, is it possible Tokelauan mental health differs from Cook Islanders’ and Niueans’. Nevertheless, these ethnicity data constraints do signal a need for national survey administrators to consider either modifying the standard ethnicity question by including “Tokelauan” as a response option or ensuring Level 2 ethnicity data are made available to researchers under conditions where data confidentiality is preserved in research outputs (e.g., aggregated Pacific sub-groups).

Conclusions

In conclusion, this study demonstrates that mental health outcomes vary significantly within Pacific peoples, which suggests mental health promotion should be targeted to Pacific ethnic groups with higher needs, particularly Cook Islands Māori and Niueans—both with Pacific homelands that fall under the Realm of New Zealand. To enable a more accurate understanding of Pacific mental health inequities, a routine national survey with a structured diagnostic interview schedule is needed, particularly since the previous survey of this nature—Te Rau Hinengaro—is 20 years old. Data access policies should allow researchers to use Pacific ethnicity data on Tokelauans under conditions where identifiability issues can be mitigated through aggregated research outputs. This study also shows that within-Pacific variability in mental health was not accounted for by socio-demographic or socio-economic differences between the Pacific ethnic groups. This supports a socio-ecological view of Pacific mental health and suggests researchers and policymakers should consider how broader structural and social determinants of mental health influence individual Pacific psychology, and how broad-based interventions could help reduce Pacific mental health inequities.

View Appendix.