28th April 2017, Volume 130 Number 1454

Carol HJ Lee, Isabelle M Duck, Chris G Sibley

Psychological distress and mental illness are prevalent in New Zealand.1,2 The 2014/15 New Zealand Health Survey (NZHS) indicated that around 6% of New Zealand adults experienced high psychological distress in the past month and 17% have been diagnosed with a mood or an anxiety disorder in their lifetime.2 However, there are inconsistencies between assessed mental health and rate of diagnosis of mental illness across ethnic groups. For instance, both Pacific and Māori peoples tend to have poorer mental health,but Pacific peoples exhibit lower rates of actual diagnosis and mental health service use than Māori.1,2 Furthermore, despite generally being found to have good mental health,1,2 some studies reveal that Asian peoples have high psychological distress.3,4

Signal Detection Theory (SDT) is a commonly used framework for categorising inconsistencies in mental illness diagnosis. From the perspective of SDT, correctly diagnosing a patient with mental illness represents a ‘hit’, failing to diagnose a mental illness a ‘miss’, incorrectly diagnosing an absent mental illness a ‘false alarm’ and not diagnosing an absent illness a ‘correct rejection’.5 However, it should be noted that the classification of ‘false alarms’ may not be appropriate when there is a time difference between the diagnosis and psychological distress measure (eg, lifetime diagnosis versus distress in past month). This is because previously diagnosed individuals with no or low current psychological distress may represent those who have received successful treatment or recovered from episodic mental illness after being diagnosed.

Research on ethnic disparities in rate of diagnosis is crucial in order to more accurately identify those who are being ‘missed’ and in need of focused psychiatric healthcare. The current paper aims to contribute to this goal by analysing ethnic group differences in rates of self-reported diagnosis of depression or an anxiety disorder relative to the likelihood of scoring in the ‘at risk’ range of the Kessler-6 measure of psychological distress in a probability sample of New Zealanders.

According to the 2014/15 NZHS, Pacific and Asian New Zealanders have the lowest rate of a mood or an anxiety disorder diagnosis (5% and 8% respectively).2 This contradicts repeated findings that Pacific peoples reported higher psychological distress than non-Pacific peoples,1,2 a factor that should increase their risk for developing mental illness.2 In terms of SDT, these findings suggest that there may be a greater rate of ‘misses’ among Pacific peoples. This high rate of under-diagnosis may be due to their high deprivation, lack of healthcare access and cultural differences in health beliefs.3,6,7

Some suggest that findings that Asian peoples tend to have good health status may reflect a “healthy immigrant effect”.6,8 This refers to the effect where migrants, who usually come from high social classes and are required to be healthy before immigrating, have good health upon their arrival to the host country but experience negative health consequences over time. Such adverse effects can be linked to their feelings of isolation, experiences of racism and low rates of healthcare utilisation.6,8,9 This in turn might increase the likelihood of ‘misses’ in the diagnosis of mental illness among Asian New Zealanders.

The Kessler-scalesare self-report measures used to screen for non-specific psychological distress in the population.10 Previous studies have confirmed the accuracy and utility of these scales,10,11 and hence, they are commonly used in both international12,13 and New Zealand studies.2,3,14 Furthermore, Ministry of Health New Zealand,15 as well as other foreign health organisations,16 have promoted the use of Kessler scales by primary-care practitioners. These scales enable us to estimate the proportion of Māori, Pacific, Asian and European New Zealanders who may be at risk (ie, scoring in the ‘at risk’ scale range) of experiencing non-specific psychological distress. This is possible because the Kessler-6 includes optimal and frequently used cut-points categorising people into broad ‘low-risk’ and ‘high risk’ groups.11–13

Extending previous research, we use a nationally representative sample of New Zealand adults to examine ethnic disparities in rate of mental illness diagnosis. We employ the Kessler-6 scale to assess the proportion of Māori, Pacific, Asian and European New Zealanders who score in the ‘at risk’ scale range over the past month, and compare this to the proportion of those who reported being diagnosed with depression or an anxiety disorder by a doctor in the last five years. Our analysis can be thought of as a broad test in terms of SDT, as we aim to compare the proportion of ‘hits’, ‘misses’, ‘correct rejections’ and previously diagnosed individuals with low current distress (referred to as ‘diagnosed/low distress’) across ethnic groups. Findings from this study will reveal those who are most in need of psychiatric healthcare and provide a framework for future research on ethnicity-specific barriers to healthcare provision.

Method

Sampling procedure

The Time 1 (2009) NZAVS longitudinal panel recruited participants by randomly selecting samples from the New Zealand electoral roll (response rate: 16.6%). A booster sample was recruited at Time 3 (2011) through an unrelated survey posted on a major New Zealand newspaper website. Further booster samples were recruited from the 2012 and 2014 electoral roll in subsequent waves (response rates: 6.2–12.33%, retention rates: around 60% across waves).We used the Time 6 (2014/15) NZAVS sample, containing 15,822 participants, for this study (retention rate: 57.2% over five years, 81.5% from previous year, see technical document).17

Participants

15,822 participants (10,003 female, 5,800 male; 19 missing) completed the Time 6 questionnaire. Participants’ mean age was 49.34 years (SD = 14.04, range 18–95; nine missing). The medians of the annual household income quartile groups were $33,900, $73,000, $110,000 and $190,000 (1,143 missing). Additionally, 74.6% (259 missing) were parents, 74.7% (640 missing) were in a committed romantic relationship and 77% (188 missing) were employed. Education (1,114 missing) was coded as a 10-point ordinal variable ranging from 0 (none) to 10 (PhD/equivalent degree, M = 5.05, SD = 2.85).

Measures

Psychological distress was measured using the Kessler-6 scale.10 This self-report measure includes six items asking participants to rate on a 5-point scale (0 = none of the time, 4 = all of the time) how often, over the last 30 days:

  • … you feel hopeless
  • … you feel so depressed that nothing could cheer you up
  • … you feel restless or fidgety
  • … you feel that everything was an effort
  • … you feel worthless
  • … you feel nervous.

Ratings for each item were summed to create a final Kessler score between 0 and 24. As discussed by Kessler et al,11 we created a categorical ‘Kessler risk score’ where participants who scored between 0–12 were coded as ‘0’ (low risk) and those who scored 13 and above were coded a ‘1’ (high risk). The NZHS also uses findings from Kessler et al11 to define cut-points on the Kessler-10 scale.18 Specifically, they use a score of 12 as the optimal cut-point to identify those with high ‘psychological distress’ and at higher risk of developing anxiety or depression (scores 0–11 indicating no/low distress, 12–40 indicating high distress).18 Participants completed the question “have you been diagnosed with, or treated for, any of the following health conditions by a doctor in the last five years?’ This question contained various response options, including depression and anxiety disorder. Those who selected either or both depression and anxiety disorder were coded as having been diagnosed by a doctor.

Ethnicity was measured using the standard New Zealand Census item, for which participants could select or nominate each ethnic group they identified with. Participants were priority coded into four mutually exclusive ethnic groups. ‘Māori’ had priority coding over all other ethnicities, followed by ‘Pacific’ and ‘Asian’ peoples, then ‘European’ respectively. The ‘European’ category included all those of European descent (eg, New Zealand European, Irish, English). Those who did not fit into these four categories were excluded from this variable.

Statistical analyses

We conducted a log linear and nested Chi-square analysis with three categorical variables: ‘diagnosis’ (no, yes), ‘Kessler risk score’; (low, high) and ‘ethnicity’ (Māori, Pacific, Asian, European) on SPSS. Analyses applied a post-stratification sample weight for men and women from the four primary ethnic groups separately, and region of residence based on the 2013 New Zealand Census.19 Unweighted analyses are reported in tabulated form.

Results

The three-way log linear analysis produced a final model that retained all effects. The highest-order interaction (diagnosis × Kessler risk score × ethnicity) was significant, χ2(3) = 10.97, p=.012. To break down this effect, separate Chi-Square tests on ‘diagnosis’ and ‘Kessler risk scores’ were performed for each ethnicity. For Europeans, there was a significant association between Kessler risk score and diagnosis, χ2(1) = 653.50, p<.001. This was also true for Māori, χ2(1) = 103.74, p<.001, Pacific, χ2(1) = 29.66, p<.001 and Asian peoples, χ2(1) = 65.71, p<.001.

Europeans had a significantly lower rate of high Kessler scores, but higher rate of diagnosis with depression or an anxiety disorder compared to all other ethnic groups (see Table 1). Conversely, Asian peoples had a significantly lower rate of diagnosis compared to other ethnic groups. Here, it is important to note the time difference between the Kessler-6 scale (psychological distress over the last month) and our diagnosis measure (self-reported diagnosis over the last five years). Consequently, the prevalence of diagnosis is higher than that for high Kessler risk scores for most ethnic groups.

Table 1: Rate of high Kessler scores and diagnosis within ethnic groups (with 95% CI).

Ethnicity

High Kessler risk score

Diagnosis

 

Unweighted

 

Weighted on gender, ethnicity and region

Unweighted

 

Weighted on gender, ethnicity and region

European

4.6%

(4.3–5.0%)

4.5%

(4.1–4.9%)

15.2%

(14.6–15.9%)

14.5%

(13.9–15.2%)

Māori

 

7.6%

(6.5–8.8%)

7.5%

(6.4–8.7%)

14.4%

(12.9–16.0%)

12.6%

(11.2–14.2%)

Pacific

 

8.4%

(6.1–11.4%)

8.0%

(6.4–10.0%)

14.0%

(11.1–17.6%)

10.5%

(8.7–12.8%)

Asian

 

8.8%

(6.8–11.3%)

8.5%

(7.3–9.8%)

8.8%

(6.8–11.3%)

7.7%

(6.6–8.9%) 

Table 2: Percentage of correct rejections, diagnosed/low distress, misses and hits within ethnic groups.

Prioritised Ethnicity

SDT categories

Unweighted

(95% CI)

Weighted on gender, ethnicity and region

(95% CI)

European

Correct rejection

82.7%

(82.1–83.4%)

83.4%

(82.7–84.1%)

Diagnosed/low distress

12.7%

(12.1–13.3%)

12.1%

(11.5–12.7%)

Miss

2.0%

(1.8–2.3%)

2.1%

(1.8–2.3%)

Hit

2.6%

(2.3–2.9%)

2.5%

(2.2–2.8%)

Māori

Correct rejection

81.4%

(79.6–83.1%)

82.9%

(81.1–84.5%)

Diagnosed/low distress

11.0%

(9.7–12.5%)

9.7%

(8.4–11.1%)

Miss

4.2%

(3.4–5.2%)

4.5%

(3.7–5.5%)

Hit

3.3%

(2.6–4.2%)

3.0%

(2.3–3.8%)

Pacific

Correct rejection

80.6%

(76.6–84.1%)

83.8%

(81.2–86.1%)

Diagnosed/low distress

11.0%

(8.4–14.3%)

8.2%

(6.5–10.2%)

Miss

5.4%

(3.6–7.9%)

5.7%

(4.3–7.4%)

Hit

3.0%

(1.8–5.1%)

2.4%

(1.6–3.6%)

Asian

Correct rejection

84.6%

(81.6–87.2%)

85.9%

(84.2–87.3%)

Diagnosed/low distress

6.6%

(4.9–8.9%)

5.6%

(4.7–6.8%)

Miss

6.6%

(4.9–8.9%)

6.5%

(5.5–7.7%)

Hit

2.2%

(1.3–3.7%)

2.0%

(1.5–2.8%) 

As illustrated in Table 2, all ethnic groups show high rates of ‘correct rejections’ and low rates of ‘hits’. Asian peoples show the highest rate of ‘misses’, followed by Pacific peoples, Māori and Europeans respectively. The rate of those categorised as ‘diagnosed/low distress’ show the reverse order, with Europeans showing the highest and Asian peoples showing the lowest rate. There were significant differences between ‘diagnosed/low distress’ and ‘misses’ proportions between any two ethnic groups, except for between ‘misses’ proportions of Asian and Pacific peoples (see Figure 1).

Figure 1: The percentage of ‘correct rejections’, ‘diagnosed/low distress’, ‘misses’ and ‘hits’ within each ethnicity (weighted on gender, ethnicity and region).

c 
Note: Proportions with significant differences within Correct rejection (European & Asian, Māori & Asian), diagnosed/low distress (all comparisons), Miss (all comparisons except Asian & Pacific) and Hit (none).  

To further explore our data, a nested Chi-square analysis using ‘diagnosis’, ‘Kessler risk scores’ and ‘ethnicity’ was conducted (see Table 3). Overall comparisons (χ2(3) = 74.02, p<.001), as well as those within low (χ2 (3) = 70.14, p<.001) and high Kessler risk scores (χ2 (3) = 55.98, p<.001) were significant. Of those with low Kessler risk scores in the last month, Europeans (12.6%) showed the highest and Asian peoples (6.2%) showed the lowest rate of within ethnic group diagnosis in the previous five years (Māori: 10.4%, Pacific: 8.9%).

Table 3: Weighted cross-tabulation of diagnosis and priority-coded ethnicity within low and high Kessler risk score groups (with 95% Confidence Intervals).

Kessler risk score

Diagnosis

 

NZ European/ Pakeha

Māori

Pacific

Asian

Total

Low (scored between 0–12)

No

Count

8,938a

1,587b

739b

1,657c

12,921

 

 

% within diagnosis

69.2%

(68.4–70.0%)

12.3%

(11.7–12.9%)

5.7%

(5.3–6.1%)

12.8%

(12.3–13.4%)

100.0%

 

 

% within ethnicity

87.4%

(86.7–88.0%)

89.6%

(88.1–90.9%)

91.1%

(89.0–92.9%)

93.8%

(92.6–94.9%)

88.6%

(88.1–89.1%)

 

 

% of Total

61.3%

(60.5–62.1%)

10.9%

(10.4–11.4%)

5.1%

(4.7–5.4%)

11.4%

(10.7–11.9%)

88.6%

 

Yes

Count

1,293a

185b

72b

109c

1,659

 

 

% within diagnosis

77.9%

(75.9–79.9%)

11.2%

(9.7–12.8%)

4.3%

(3.5–5.4%)

6.6%

(5.5–7.9%)

100.0%

 

 

% within ethnicity

12.6%

(12.0–13.3%)

10.4%

(9.1–12.0%)

8.9%

(7.1–11.0%)

6.2%

(5.1–7.4%)

11.4%

(10.9–11.9%)

 

 

% of Total

8.9%

(8.4–9.3%)

1.3%

(1.1–1.5%)

0.5%

(0.4–0.6%)

0.7%

(0.6–0.9%)

11.4%

High (scores between 13–24)

No

Count

220a

86b

50b,c

125c

481

 

 

% within diagnosis

45.7%

(41.3–50.2%)

17.9%

(14.7–21.6%)

10.4%

(8.0–13.4%)

26.0%

(22.3–30.1%)

100.0%

 

 

% within ethnicity

45.5%

(41.1–49.9%)

60.1%

(52.0–67.8%)

70.4%

(59.0–79.8%)

76.2%

(69.2–82.1%)

55.8%

(52.5–59.1%)

 

 

% of Total

25.5%

(22.7–28.5%)

10.0%

(8.2–12.2%)

5.8%

(4.4–7.6%)

14.5%

(12.3–17.0%)

55.8%

 

Yes

Count

264a

57b

21b,c

39c

381

 

 

% within diagnosis

69.3%

(69.3–64.5%)

15.0%

(11.7–19.0%)

5.5%

(3.6–8.3%)

10.2%

(7.6–13.7%)

100.0%

 

 

% within ethnicity

54.5%

(50.1–58.9%)

39.9%

(32.2–48.1%)

29.6%

(20.2–41.0%)

23.8%

(17.9–30.9%)

44.2%

(40.9–47.5%)

 

 

% of Total

30.6%

(27.6–33.8%)

6.6%

(5.1–8.5%)

2.4%

(1.6–3.7%)

4.5%

(3.3–6.1%)

44.2%

Note: N=15,442, weighted on gender, ethnicity and region of residence, different subscript letters for proportions indicate significant differences across columns (z-test based on standard procedures in SPSS).

Among those with high Kessler risk scores, Europeans (54.5%) again showed the highest and Asian peoples (23.8%) showed the lowest rate of within ethnic group diagnosis (Māori: 39.9%, Pacific: 29.6%; Table 4 presents unweighted proportions).

Table 4: Unweighted cross-tabulation of diagnosis and priority-coded ethnicity within low and high Kessler risk score groups (with 95% Confidence Intervals).

Kessler risk score

Diagnosis

 

NZ European/ Pakeha

Māori

Pacific

Asian

Total

Low (scored between 0–12)

No

Count

10,287a

1,584a

345a

528b

12,744

 

 

% within diagnosis

80.7%

(80.0–81.4%)

12.4%

(11.9–13.0%)

2.7%

(2.4–3.0%)

4.1%

(3.8–4.5%)

100.0%

 

 

% within ethnicity

86.7%

(86.1–87.3%)

88.0%

(86.5–89.5%)

88.0%

(84.4–90.9%)

92.8%

(90.4–94.6%)

87.2%

(86.6–87.7%)

 

 

% of Total

70.4%

(69.6–71.1%)

10.8%

(10.3–11.4%)

2.4%

(2.1–2.6%)

3.6%

(3.3–3.9%)

87.2%

 

Yes

Count

1,573a

215a

47a

41b

1,876

 

 

% within diagnosis

83.8%

(82.1–85.4%)

11.5%

(10.1–13.0%)

2.5%

(1.9–3.3%)

2.2%

(1.6–3.0%)

100.0%

 

 

% within ethnicity

13.3%

(12.7–13.9%)

12.0%

(10.5–13.5%)

12.0%

(9.1–15.6%)

7.2%

(5.4–9.6%)

12.8%

(12.3–13.4%)

 

 

% of Total

10.8%

(10.3–11.3%)

1.5%

(1.3–1.7%)

0.3%

(0.2–0.4%)

0.3%

(0.2–0.4%)

12.8%

High (scores between 13-24)

No

Count

250a

82b

23b,c

41c

396

 

 

% within diagnosis

63.1%

(58.3–67.7%)

20.7%

(17.0–25.0%)

5.8%

(3.9–8.6%)

10.4%

(7.7–13.7%)

100.0%

 

 

% within ethnicity

43.7%

(39.7–47.8%)

55.8%

(47.7–63.6%)

63.9%

(47.6–77.5%)

74.5%

(61.7–84.2%)

48.9%

(45.5–52.3%)

 

 

% of Total

30.9%

(27.8–34.1%)

10.1%

(8.2–12.4%)

2.8%

(1.9–4.2%)

5.1%

(3.8–6.8%)

48.9%

 

Yes

Count

322a

65b

13b,c

14c

414

 

 

% within diagnosis

77.8%

(73.5–81.5%)

15.7%

(12.5–19.5%)

3.1%

(1.8–5.3%)

3.4%

(2.0–5.6%)

100.0%

 

 

% within ethnicity

56.3%

(52.2–60.3%)

44.2%

(36.4–52.3%)

36.1%

(22.5–52.4%)

25.5%

(15.8–38.3%)

51.1%

(47.7–54.5%)

 

 

% of Total

39.8%

(36.4–43.2%)

8.0%

(6.4–10.1%)

1.6%

(0.9–2.7%)

1.7%

(1.0–2.9%)

51.1%

Note: N=15,430, Chi-square overall (χ2(3) = 20.07, p<.001) within low (χ2(3) = 19.57, p<.001) and high Kessler risk scores (χ2(3) = 26.67, p<.001), different subscript letters indicate significant differences across columns (z-test based on standard procedures in SPSS). 

Relative to Europeans, Māori, Pacific and Asian peoples are more likely to score in the ‘high risk’ range over the last month, but are less likely to be diagnosed in the last five years. Of those with high Kessler risk scores, Māori are 1.32 times more likely, Pacific peoples 1.54 times more likely and Asian peoples 1.67 times more likely to be undiagnosed (within ethnic group) compared to Europeans. Alternatively, of those with low Kessler risk scores, Māori are .82 times, Pacific peoples .71 times and Asian peoples .49 times less likely to be diagnosed (within ethnic group) than Europeans.

Discussion

The present study examined ethnic disparities in mental illness diagnosis using a probability sample of New Zealand adults. We used the Kessler-6 scale to identify those who scored in the ‘at risk’ range over the last month, and compared this to the proportion of those actually diagnosed with depression or an anxiety disorder by a doctor in the past five years. Our results reveal ethnic inequalities in rate of diagnosis, which are inconsistent with ethnic differences in the screening risk for mental illness.

Ethnic differences in mental illness diagnosis

After applying sample weighting on gender, ethnicity and region of residence, we found that Europeans (4.5%) were the least likely to score in the high Kessler risk range (Māori: 7.5%, Pacific: 8%, Asian: 8.5%), but showed the highest prevalence of diagnosed depression and anxiety disorder (European: 14.5%, Māori: 12.6%, Pacific: 10.5%, Asian: 7.7%). Most ethnic groups exhibited a higher rate of diagnosis than high Kessler-risk scores, which can be explained by the time difference between the Kessler-6 scale (distress over the last month) and our diagnostic measure (diagnosis over the last five years).

The 2014/15 NZHS found that both Māori (9.6%) and Pacific (10.2%) adults exhibited greater rates of high Kessler-10 risk scores (ie, 12 or above), but Māori individuals showed a much higher rate of lifetime diagnosis with mood or an anxiety disorder (17.4% and 7.5% respectively).2 In comparison to the NZHS, our results indicate a slightly lower prevalence of high Kessler risk scores for Māori (7.5%) and Pacific peoples (8%). Furthermore, the NZHS found that Asian peoples have the lowest (5.5%), whereas our study found that Asian peoples have the highest rate (8.5%) of high psychological distress.2 Regarding mental illness diagnosis, the 2006 New Zealand Mental Health Survey (NZMHS) found that, after adjusting for demographic factors, Māori (5.7%) and non-Māori/non-Pacific peoples (5.8%) exhibited a higher 12-month prevalence of major depressive disorder than Pacific peoples (3.5%).20 However, there were no significant ethnic differences in 12-month prevalence of anxiety disorders (12.9–15.6%).20

In terms of SDT, we found that mental health problems of Māori (4.5%), Asian (6.5%) and Pacific peoples (5.7%) are more likely to be ‘missed’ compared to Europeans (2.1%). Although ethnic minorities exhibited greater rates of high Kessler-risk scores, all ethnic groups showed similar rates of ‘hits’ (2–3%). Additionally, Europeans (12.1%) showed the highest proportion of those categorised as ‘diagnosed/low distress’ (Māori: 9.7%, Pacific: 8.2%, Asian: 5.6%). As the self-reported diagnosis question asked about the last five years, this indicates that Europeans who were diagnosed in the past may have been more likely to successfully manage or be treated for their depression/anxiety disorder within the asked five-year period, relative to people from other ethnic groups.

Under-diagnosis among Pacific and Asian peoples

For Asian peoples, their under-diagnosis is likely to be associated with their low rate of psychiatric healthcare utilisation,1,2 which in turn is linked with language or cultural barriers to healthcare access.6,8 This includes the greater stigma surrounding mental illness in Asian cultures. Among Pacific peoples, factors such as costs, transport and language barriers are likely to contribute to their low utilisation of healthcare services and under-diagnosis.2,6,7 Moreover, perhaps due to the lack of cultural competence among doctors, many Pacific peoples report experiences of low-quality service in primary healthcare.6

Previous American studies indicate that patients’ ethnicities influence the ability of physicians to accurately detect mental health problems.21,22 For instance, Borowsky et al21 found that American physicians were less likely to detect depression in African-American or Hispanic patients. Likewise, New Zealand physicians may be less likely to recognise mental health problems in Pacific and Asian peoples. Alegria and McGuire23 found that ‘cultural, social and contextual’ factors have important influence on how one expresses psychiatric symptoms. Unlike Western perspectives, Pacific peoples place greater emphasis on familial wellbeing as an aspect of their own subjective wellbeing,24 Similarly, Asian peoples tend to work through health problems as a family and often endorse non-Western models of health treatment.9 These cultural beliefs may be influencing how Pacific and Asian peoples express their mental illness in differential ways and hence, medical professionals need to develop cultural competence to understand and accurately diagnose these individuals.

Caveats and future research

The 2006 NZMHS found that the high prevalence of mental illness among Pacific peoples was largely explained by population differences in age and gender.20 Future studies should examine whether population age, in addition to gender and region of residence, may be affecting ethnic differences in psychological health and diagnosis. Furthermore, although the Kessler-6 scale is a widely used and validated measure of psychological distress,optimal scaling rules were found to differ across countries.25 Therefore, it is crucial to investigate potential cultural variations in understandings of ‘depression’ and ‘anxiety’, and how this may influence the accuracy of Kessler-risk scores.

This study analysed data from the broader longitudinal self-report postal NZAVS. It is important to note that the NZAVS has a relatively low initial response rate of 16.6% in 2009. Although general survey response rates have been declining over the years, some medical research has nevertheless achieved much higher response rates. For example, a recent study by Chesang et al26 on contraceptive use in New Zealand achieved a response rate of 47% by recruiting participants through a postal survey with telephone follow-up. The NZAVS did not employ a telephone follow-up at Time 1. Although low response rates do not necessarily indicate non-response bias, high response rates are desirable in probability sample surveys, as this enables more accurate estimation of sampling error and the ability to correct for biases in estimates.27

Applying post-survey adjustments can correct for bias in data collection even without high response rates.27 In support of this idea, the 2012 telephone survey by Pew Centre (with a 9% response rate) report that their sample provided a reliable reflection of the general public on social and economic measures after applying sample weighting.28 Similarly, the NZAVS applies post-stratification sample weighting on demographics, and its validity in monitoring changes in New Zealanders’ political attitudes over time has been demonstrated.29 The point, however, remains that as non-response bias tends to vary depending on the variable of interest,27 we were unable to determine the exact degree of bias in the specific items used in this study.

Concluding comments

The current study investigated ethnic disparities in self-reported diagnosis of depression or an anxiety disorder by a doctor relative to scores on a screening measure for these same mental illnesses using data from the NZAVS. Our results reveal ethnic disparities in diagnosis of mental illness, which are inconsistent with scores on the screening measure. Although a larger proportion of Māori, Pacific and Asian peoples scored in the high Kessler risk range, Europeans reported the highest rate of doctor diagnosed depression or anxiety. Mental health problems of ethnic minorities, especially Pacific and Asian peoples, are more likely to be ‘missed’, while those of Europeans are more likely to be ‘hit’. These findings are likely to reflect ethnic inequalities in access to, expectations from and style of communication with, medical professionals. 

Summary

Our results indicate an ethnic inequality in self-reported diagnosis of depression or an anxiety disorder by a doctor, relative to scores on the screening measure (ie, Kessler-6 scale) for these same forms of mental illness in a nationally representative sample of New Zealand adults. Māori, Pacific and Asian New Zealanders were more likely to score in the ‘at risk’ range of the Kessler-6 scale, indicating an increased likelihood of depression or anxiety, relative to European New Zealanders. However, European New Zealanders reported the highest rate of actual diagnosis with depression or anxiety in the previous five-year period. This inequality may reflect ethnic group differences in access to, expectations from and style of communication with, medical professionals.

Abstract

Aim

This study explored ethnic disparities in self-reported diagnosis of depression or an anxiety disorder by a doctor, relative to scores on the screening measure for these same forms of mental illness in a probability sample of New Zealand adults.

Method

15,822 participants responded to the 2014/15 New Zealand Attitudes and Values Study (NZAVS) longitudinal panel. Participants completed the Kessler-6 scale (a screening measure of non-specific psychological distress over the last month) and reported whether a doctor had diagnosed them with depression or an anxiety disorder any time in the last five years.

Results

Māori, Pacific and Asian New Zealanders were more likely to score in the ‘at risk’ range of the Kessler-6 scale, indicating an increased likelihood of depression or anxiety, relative to European New Zealanders. However, European New Zealanders reported the highest rate of actual diagnosis with depression or anxiety in the previous five-year period.

Conclusion

There is an ethnic inequality in diagnosis received in the last five years relative to population-level screening risk for depression and anxiety disorders over the last month. Māori, Pacific and Asian New Zealanders are more likely to be under-diagnosed with depression and anxiety disorders relative to European New Zealanders. This inequality may reflect ethnic group differences in access to, expectations from and style of communication with, medical professionals.

Author Information

Carol HJ Lee, Psychology, University of Auckland, Auckland; Isabelle M Duck, Westgate Medical Centre, Auckland; Chris G Sibley, Psychology, University of Auckland, Auckland.

Acknowledgements

The research was supported by a Templeton World Charity Foundation Grant awarded to Chris Sibley (ID: 0077). Mplus syntax for the models reported here are available at: www.psych.auckland.ac.nz/uoa/NZAVS 

Correspondence

Carol HJ Lee, Psychology, School of Psychology, University of Auckland, Auckland 1142.

Correspondence Email

clee879@aucklanduni.ac.nz

Competing Interests

Nil.

References

  1. Mental Health Foundation. Quick Facts and Stats 2014. Auckland: Author. 2014. 
  2. Ministry of Health. Annual Update of Key Results 2014/15: New Zealand Health Survey. Wellington: Author. 2015.
  3. Krynen A, Osborne D, Duck IM, et al. Measuring psychological distress in New Zealand: Item response properties and demographic differences in the Kessler-6 screening measure. New Zealand Journal of Psychology. 2013; 42:69–83.
  4. Rasanathan K, Ameratunga S, Chen J, et al. A health profile of young Asian New Zealanders who attend secondary school: findings from Youth2000. Auckland: The University of Auckland. 2006. 
  5. McFall RM, Treat TA. Quantifying the information value of clinical assessments with signal detection theory. Annu Rev Psychol. 1999; 50:215–41.
  6. St George I, editor. Cole’s Medical practice in New Zealand. Wellington: Medical Council of New Zealand. 2013. 
  7. Statistics New Zealand and Ministry of Pacific Island Affairs. Health and Pacific peoples in New Zealand. Wellington: Statistics New Zealand and Ministry of Pacific Island Affairs. 2011.
  8. Wong A. Challenges for Asian health and Asian health promotion in New Zealand. Auckland: Health promotion Forum of New Zealand. 2015. 
  9. Harris R, Cormack D, Tobias M, et al. The pervasive effects of racism: experiences of racial discrimination in New Zealand over time and associations with multiple health domains. Soc Sci Med. 2012; 74:408–15.
  10. Kessler RC, Andrews G, Colpe LJ, et al. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med. 2002; 32:959–76.
  11. Kessler RC, Barker PR, Colpe LJ, et al. Screening for serious mental illness in the general population. Arch Gen Psychiatry. 2003; 60:184–9.
  12. Lee S, Tsang A, Ng KL, et al. Performance of the 6-item Kessler scale for measuring serious mental illness in Hong Kong. Compr Psychiatry. 2012; 53:584–92.
  13. Giallo R, Cooklin A, Nicholson JM. Risk factors associated with trajectories of mothers’ depressive symptoms across the early parenting period: an Australian population-based longitudinal study. Archives of women’s mental health. 2014; 17:115–25.
  14. Foulds J, Wells JE, Mulder R. The association between material living standard and psychological distress: Results from a New Zealand population survey. Int J Soc Psychiatry. 2014; 60:766–71.
  15. New Zealand Guidelines Group. Identification of Common Mental Disorders and Management of Depression in Primary Care. An Evidence-based Best Practice Guideline. Wellington: New Zealand Guidelines Group. 2008. 
  16. M. Johnson. Screening tools for Primary Care. California: California Primary Health Association, Integrated Behavioral Health Project. 2013. 
  17. Sibley CG. Sampling procedure and sample details for the New Zealand Attitudes and Values Study. NZAVS Technical Documents. 2014;e01. http://www.psych.auckland.ac.nz/en/about/our-research/research-groups/new-zealand-attitudes-and-values-study/nzavs-tech-docs.html
  18. Ministry of Health. Content guide 2014/15. Wellington: Author. 2015.
  19. Sibley CG. Procedures for estimating post-stratification NZAVS sample weights. 2014. http://cdn.auckland.ac.nz/assets/psych/about/our-research/nzavs/NZAVSTechnicalDocuments/NZAVS-Technical-Documents-e08-Procedures-for-Estimating-Sample-Weights.pdf
  20. MA Oakley Browne, JE Wells, KM Scott (eds). Te Rau Hinengaro: The New Zealand Mental Health Survey. Wellington: Ministry of Health. 2006.
  21. Borowsky SJ, Rubenstein LV, Meredith LS, et al. Who is at risk of nondetection of mental health problems in primary care?. J Gen Intern Med. 2000; 15:381-8.
  22. Chung H, Teresi J, Guarnaccia P, et al. Depressive symptoms and psychiatric distress in low income Asian and Latino primary care patients: prevalence and recognition. Community Ment Health J. 2003; 39:33–46.
  23. Alegría M, McGuire T. Rethinking a universal framework in the psychiatric symptom-disorder relationship. J Health Soc Behav. 2003; 44:257–74.
  24. Manuela S, Sibley CG. The Pacific Identity and Wellbeing Scale (PIWBS): A culturally-appropriate self-report measure for Pacific peoples in New Zealand. Social indicators research. 2013; 112:83–103.
  25. Kessler RC, Green JG, Gruber MJ, et al. Screening for serious mental illness in the general population with the K6 screening scale: results from the WHO World Mental Health (WMH) survey initiative. International journal of methods in psychiatric research. 2010; 19:4–22.
  26. Chesang J, Richardson A, Potter J, Sneyd MJ, Coope P. Prevalence of contraceptive use in New Zealand women. NZ Med J. 2016; 129:71–80.
  27. Groves RM. Nonresponse rates and nonresponse bias in household surveys. Public opinion quarterly. 2006; 70:646–75.
  28. Pew Research Centre. Assessing the Representativeness of Public Opinion Surveys. Washington D.C: Author. 2012. 
  29. Sibley CG, Robertson A, Osborne D, et al. Bias and tracking accuracy in voting projections using the New Zealand attitudes and values study. Political Science. 2016.