Social Science & Medicine 66 (2008) 1165e1173www.elsevier.com/locate/socscimed

Mentalephysical comorbidity in an ethnicallydiverse population*

Kate Scott a,*, Magnus A. McGee b, David Schaaf c, Joanne Baxter d

a Psychological Medicine, Wellington School of Medicine and Health Sciences, Mein Street,

Newtown P.O. Box 7343, Wellington South, New Zealandb Christchurch School of Medicine and Health Sciences, Christchurch, New Zealand

c University of Auckland, Auckland, New Zealandd Otago University, Dunedin, New Zealand

Available online 26 December 2007

Abstract

The relationship between mental disorders and chronic physical conditions is well established, but the possibility of ethnic groupdifferences in mentalephysical associations has seldom been investigated. This study investigated ethnic differences in associa-tions between four physical conditions (chronic pain, cardiovascular disease, diabetes, and respiratory disease) and 12-monthmood and anxiety disorders. A nationally representative face-to-face household survey was carried out in New Zealand from2003 to 2004 with 12,992 participants aged 16 and older, achieving a response rate of 73.3%. The current study is of the subsampleof 7,435 participants who were assessed for chronic physical conditions (via a standard checklist), and compares Maori, Pacific andOther New Zealanders. DSM-IV mental disorders were measured with the Composite International Diagnostic Interview (CIDI3.0). The ethnic groups differed significantly in prevalences of both physical and mental disorders, but almost no ethnic differencesin mentalephysical associations were found. Independent of ethnicity, associations were observed between chronic pain and moodand anxiety disorders, cardiovascular disease and anxiety disorders, respiratory disease and mood and anxiety disorders. Despitedifferences in mental and physical health status between ethnic groups in New Zealand, mentalephysical disorder associationsoccur with considerable consistency across the groups. These results suggest that whatever factors are conducive to the developmentof a mental disorder from a physical disorder (or vice versa), they are either unaffected by the cultural differences manifest in theseethnic groups, or, any cultural factors operating serve to both increase and decrease comorbidity such that they cancel each other out.� 2007 Elsevier Ltd. All rights reserved.

Keywords: New Zealand; Mental disorders; Chronic physical conditions; Comorbidity; Ethnicity

* Te Rau Hinengaro: The New Zealand Mental Health Survey

(NZMHS) was funded by the Ministry of Health, the Alcohol Advi-

sory Council of New Zealand and the Health Research Council of

New Zealand. Work on this paper was supported in part by the

New Zealand Lottery Grants Board. The survey was carried out in

conjunction with the World Health Organization World Mental

Health (WMH) Survey Initiative.

* Corresponding author.

E-mail address: kate.scott@otago.ac.nz (K. Scott).

0277-9536/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.socscimed.2007.11.022

Introduction

The significant comorbidity of physical conditionswith mental disorders is now well documented (Buist-Bouwman, de Graaf, Vollebergh, & Ormel, 2005; Dew,1998; Harter, Conway, & Merikangas, 2003; Katon &Ciechanowski, 2002; Wells, Golding, & Burnam,

1166 K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

1989a, 1989b). This overlap of physical and mentalconditions is consequential for both the individuals con-cerned and their treatment providers. Mentalephysicalcomorbidity impacts on role impairment (Kessler,Ormel, Demler, & Stang, 2003; Sullivan, LaCroix,Baum, Grothaus, & Katon, 1997), treatment costs andadherence (Ciechanowski, Katon, & Russo, 2000;Simon, Ormel, Von Korff, & Barlow, 1995) and prema-ture mortality risk (Harris & Barraclough, 1998; vanMelle et al., 2004; Zhang et al., 2005). Recent datafrom the World Mental Health surveys have indicatedthat both depressive and anxiety disorders are equally,and independently associated with a range of chronicphysical conditions (Scott et al., 2007).

One aspect of mentalephysical comorbidity that hasseldom been investigated is whether the associationsbetween mental and physical disorders are uniformacross ethnic groups within countries. A recent studylooked at comorbidity of any depressive and any anxi-ety disorder with asthma, diabetes and cardiovasculardisease among four Latino subgroups (Cuban, PuertoRican, Mexican, and Other Latino) (Ortega, Feldman,Canino, Steinman, & Algeria, 2006). However, the au-thors’ claim to have observed unique patterns amongthe subgroups was not supported by any test of interac-tions between ethnic group, mental disorder and diseaseoutcomes. One of the interesting findings to emergefrom that study was that none of the four Latino groupsshowed an increased prevalence of depressive disorderamong those with cardiovascular disease (relative tothose without cardiovascular disease). Ortega et al. sug-gested that this might be explained by protective cul-tural factors that Latinos share, such as close familybonds. It is this kind of possibility e that ethnic groupsmay differ in factors which facilitate or protect againstthe development of comorbid conditions e whichmakes the topic of ethnicity important to study. It mayyield insights into the mechanisms linking mental andphysical disease generally, and it may shed light on in-equalities in health outcomes among ethnic groups,indicating areas of need for targeted health care.

New Zealand has a number of ethnic groups withsubstantial cultural differences, including the indige-nous Maori population, the Pacific population (immi-grants and their New Zealand born descendants froma number of islands in the South Pacific), and theremainder of the New Zealand population (Other)who are mostly descended from 19th century Britishand European colonists. These three groups differ interms of mental and physical health status (Baxter,Kokaua, Wells, McGee, & Oakley Browne, 2006;Ministry of Health, 2004). Te Rau Hinengaro: The

New Zealand Mental Health Survey (NZMHS),which surveyed mental disorders in the general pop-ulation and also screened for a number of chronicphysical conditions, therefore presents a useful oppor-tunity to investigate mentalephysical comorbidity inan ethnically diverse population. Importantly, theNZMHS is a large survey and it oversampled theMaori and Pacific populations to boost precision ofestimates. This is relevant to the logistical issuethat has no doubt hampered investigations of thistopic: stratifying into multiple ethnic groups andthen investigating mentalephysical comorbiditieswithin each group demands a large sample size.

This study uses the NZMHS to assess whether thereare significant differences between the Maori, Pacificand Other ethnic groups in the associations betweenfour chronic physical conditions (chronic pain, cardio-vascular disease, diabetes and respiratory disease) and12-month mood and anxiety disorders. Specifically,we assessed whether there were interactions betweenethnicity and each of the physical conditions in theassociation with mood and anxiety disorders.

Methods

Sample

Te Rau Hinengaro: The New Zealand Mental HealthSurvey (NZMHS) was a nationally representativehousehold survey involving face-to-face interviewswith 12,992 people aged 16 and older. Interviewswere conducted from October 2003 to December2004 with a response rate of 73.3%. Mental disorderswere diagnosed with the Composite International Di-agnostic Interview (CIDI). The CIDI is a long inter-view so to reduce respondent burden a strategy ofshort (Part 1) and long (Part 1þ Part 2) forms of theinterview was adopted. All participants (12,992)completed the first half of the interview (Part 1) whichincluded screening questions for all mental disorders;full diagnostic assessment for the mood disorders, allanxiety disorders except post-traumatic stress disorder(PTSD) and obsessive-compulsive disorder (OCD), thesubstance use disorders; assessment of suicidal behav-iour and health service use. A subsample of partici-pants who completed Part 1 of the interview andreported having no lifetime history of mental disorderwere then terminated at the end of Part 1. All partici-pants who met criteria for any mental disorders, orwho had ever had a suicide plan or attempt, or whohad ever been hospitalised for psychiatric problems,together with a probability subsample of those with

1167K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

no reported psychiatric problems, went on to completePart 2 of the interview. Part 2 included the remainder ofthe mental disorders (PTSD, OCD, and eating disor-ders), assessment of physical conditions, and othertopics related to the survey aims.

Part 2 respondents (n¼ 7,435) were weighted by theinverse of their probability of selection for Part 2 of theinterview to adjust for differential sampling. A more de-tailed description of the survey methods is providedelsewhere (Wells et al., 2006).

Mental disorder status

The interview used to assess mental disorder wasversion 15 of the World Mental Health-CIDI. Version20 is now the CIDI 3.0 (Kessler & Ustun, 2004). Thisascertains lifetime disorder plus recency of episodesor symptoms, which allows 12-month and 1-month di-agnoses to be derived. Disorders were assessed usingthe definitions and criteria of the Diagnostic and Statis-tical Manual of Mental Disorders, Fourth Edition(DSM-IV) (APA, 1994). CIDI organic exclusion ruleswere imposed in making all diagnoses. Because theemphasis in this paper was on co-occurrence of mentaland physical disorders (what is referred to as ‘episode’comorbidity rather than ‘lifetime comorbidity’ (Batstra,Bos, & Neeleman, 2007)), due to its greater relevance toclinicians, we used 12-month prevalences of mental dis-order in this paper rather than lifetime prevalences.Many of those who meet the criteria for lifetime mentaldisorder do not meet the criteria for mental disorder inthe past 12 months (Wells et al., 2006), making lifetimeprevalences less suitable for the investigation of co-occurring mental and physical disorders. One-monthmental disorders were too few in number for the presentinvestigation.

Diagnoses were calculated with hierarchy (so, forexample, those meeting the criteria for both generalizedanxiety disorder (GAD) and major depressive disorderwould only be counted once, as having major depres-sive disorder). The disorder groups included in thispaper are anxiety disorders (panic disorder, agoraphobiawithout panic, specific phobia, social phobia, GAD,PTSD, and OCD) and mood disorders (major depres-sive disorder, dysthymia, bipolar disorder (I, II andany with mania or hypomania)). Substance use disor-ders were not included, in part because they are some-what less reliable in their associations with physicalconditions (Scott, Oakley Browne, McGee, & Wells,2006), and also because their lower prevalence makesinvestigating ethnic differences in their associationwith physical conditions problematic.

Ethnicity status

Ethnicity was determined using the ethnicity ques-tion from the 2001 New Zealand Census of Populationand Dwellings, in which respondents were asked whichethnic group or groups they belonged to, allowingmultiple ethnicities to be reported. The Statistics NewZealand prioritization rule for ethnicity assignmentwas followed. Those classified as Maori were thosewho identified solely as Maori or who identified Maorias one of their ethnic groups. Those who identified asbelonging to any Pacific Island group, but not Maori,were classified as Pacific. All others were combinedinto the composite Other ethnic group. In the 2001 Cen-sus, Maori were 11.2%, Pacific 4.5% and Other 84.3%of the New Zealand population aged 16 years and older.In the NZMHS Maori and Pacific were oversampledusing a mixture of targeting and screening. This dou-bled the number of Maori and quadrupled the numberof Pacific people in the sample relative to their propor-tions in the population.

Chronic physical conditions

Chronic physical conditions were screened for byasking participants in the Part 2 subsample if they hadever had arthritis, rheumatism, chronic back or neckproblems, frequent or severe headaches, other chronicpain, seasonal allergies, stroke, heart attack, or whetherthey had ever been told by a doctor they had heart dis-ease, high blood pressure, asthma, tuberculosis, chroniclung disease, diabetes, ulcer, HIV/AIDS, epilepsy orcancer. For problems that could have remitted (pain,seasonal allergies, high blood pressure, tuberculosis,diabetes or ulcer), participants were also asked if theystill had the condition/s in the past 12 months. The12-month prevalence of chronic conditions reportedhere includes the enduring conditions and those othersstill present in the past 12 months.

The four conditions included in the current analysesare those which are prevalent in New Zealand (chronicpain, cardiovascular disease, respiratory disease) and/or show marked ethnic disparities in prevalence (diabe-tes). Conditions are grouped as follows: Chronic pain(arthritis or rheumatism; chronic back or neck problem;frequent or severe headaches; and any other chronicpain); cardiovascular disease (stroke; heart attack;and heart disease); respiratory disease (asthma; chronicobstructive pulmonary disease; emphysema; and otherchronic lung disease); and diabetes (diabetes). Priorresearch has demonstrated reasonable correspondencebetween self-reported chronic conditions such as

Table 1

Sex, age and ethnicity distribution in the New Zealand Mental Health

Survey (NZMHS)

Whole (Part 1) sample

(n¼ 12,992)

Part 2 subsample

(n¼ 7,435)

Weighted

(%)

Number Unweighted

(%)

Number Unweighted

(%)

Sex

Male 5,634 43.3 3016 40.6 48.0

Female 7,358 56.6 4419 59.4 52.0

Age (years)

16e24 1,535 11.8 1027 13.8 15.7

25e44 5,304 40.8 3215 43.2 39.7

45e64 3,909 30.1 2266 30.5 29.6

65 þ 2,244 17.3 927 12.5 15.0

Prioritised ethnicity

Maori 2,595 20.0 1643 22.1 11.2

Pacific 2,236 17.2 1339 18.0 4.5

Other 8,161 62.8 4453 59.9 84.3

1168 K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

diabetes, heart disease and respiratory disease and gen-eral practitioner medical records (Kriegsman, Penninx,Van Eijk, Boeke, & Deeg, 1996; NCHS, 1994).

Analysis

Estimates wereweighted to take into account the prob-ability of selection (of meshblocks within a stratum, ofhouseholds within a meshblock, and of individuals withina household), to adjust for oversampling of Maori andPacific groups, to adjust for non-response, and to post-stratify by age, sex and ethnicity to the 2001 censuspopulation. Part 2 weights were used for the analyses in-volving physical conditions as only the Part 2 subsamplewas assessed for physical conditions. Analyses were car-ried out using SUDAAN 9.0.1. to allow unbiased varianceestimates from the complex survey design. Taylor serieslinearization (Shah, 1998) was used to approximate thevariance of estimates. For prevalences with less than 30events in the numerator, confidence intervals were calcu-lated according to a method by Korn & Graubard, 1998.

Logistic regression models were used to generate ad-justed prevalence estimates (predictive marginals) thatcontrolled for the differential age and sex distributionof those with and without the specified physical condi-tion, and to generate age and sex-adjusted odds ratios ofthe association between each physical condition andmental disorder outcomes (relative to those withoutthe physical condition). All values of N quoted are thenumber of observations, not the weighted number.

Results

Sample characteristics

Table 1 shows the number and the unweighted andweighted proportions of demographic subgroups inthe NZMHS.

Ethnicity and chronic physical conditions

The prevalences of the four physical conditions byethnic group are shown in Table 2, adjusted for differ-ences in the age and sex distribution across the threeethnic groups. There was a significant effect of ethnicityoverall in the prevalence of chronic pain (p< 0.001),cardiovascular disease (p< 0.01), diabetes (p< 0.0001)and respiratory disease (p< 0.0001). In terms of pairedcontrasts between ethnic groups, after adjusting forage and sex, Maori had a significantly higher preva-lence of chronic pain compared with the Pacific group(p< 0.0001), and compared with the Other group

(p¼ 0.05). The Pacific group had a significantly lowerprevalence of chronic pain than the Other group(p¼ 0.001). For cardiovascular disease, Maori had a sig-nificantly higher prevalence than the Other group(p< 0.01). For diabetes, Maori and Pacific groups didnot differ significantly from each other but Maori weresignificantly higher than the Other group (p< 0.001),as was the Pacific group (p< 0.001). Finally, for respira-tory disease, all three ethnic groups differed significantlyfrom each other (Maori vs Pacific: p< 0.0001; Maorivs Other: p< 0.0001; Pacific vs Other: p< 0.05). Theseresults are consistent with the general pattern of ethnicdifferences in chronic diseases in the most recentnational health survey (Ministry of Health, 2004).

Ethnicity and 12-month mood and anxiety disorders

The prevalences of any 12-month mood and any 12-month anxiety disorder are shown by ethnic group inTable 3, after adjusting for age and sex differences be-tween the ethnic groups. There was a significant effectof ethnicity overall in the prevalences of both mooddisorders (p< 0.01) and anxiety disorders (p¼ 0.02).Maori had a significant higher prevalence of mooddisorder compared with Pacific (p< 0.01) and com-pared with the Other group (p< 0.01). Maori also hada significantly higher prevalence of anxiety disorderscompared with the Other group (p< 0.01).

Chronic pain and mental disordersThe prevalences of any 12-month mood disorder

and any 12-month anxiety disorder among those with

Table 2

Prevalence of chronic physical conditions by ethnicity, adjusted for age and sex, plus unweighted, unadjusted Ns (from the Part 2 subsample)

Ethnicity Chronic physical condition

Chronic paina Cardiovascularb Diabetes Respiratoryc

%

(CI)

N

%

(CI)

N

%

(CI)

N

%

(CI)

N

Maori 42.2

(38.7, 45.7) 701

11.9

(9.1, 14.7) 123

8.0

(5.7, 10.3) 98

26.7

(23.3, 30.0) 425

Pacific 31.0

(26.8, 35.1) 421

8.5

(5.7, 11.4) 80

11.3

(7.6, 15.0) 92

13.4

(10.4, 16.4) 208

Other 38.4

(36.4, 40.4) 2058

7.6

(6.5, 8.6) 415

3.6

(2.8, 4.3) 169

17.1

(15.6, 18.6) 824

a Chronic pain: arthritis or rheumatism; chronic back or neck problem; frequent or severe headaches; and any other chronic pain.b Cardiovascular disease: stroke; heart attack; and heart disease.c Respiratory disease: asthma; chronic obstructive pulmonary disease; emphysema; and other chronic lung disease.

Table 3

Prevalence of 12-month mood and anxiety disorders by ethnicity,

adjusted for age and sex, plus unweighted, unadjusted Nsa

Ethnicity 12-Month mental disorder

Any mood disorder

% (CI) N

Any anxiety disorder

% (CI) N

Maori 10.1 (8.8, 11.5) 319 17.6 (15.4, 19.7) 534

Pacific 7.2 (5.8, 8.7) 200 14.8 (12.4, 17.3) 372

Other 7.7 (6.9, 8.4) 603 14.4 (13.3, 15.5) 1125

a The cell Ns for Any Mood Disorder come from the whole (Part 1)

sample (n¼ 12,992); the cell Ns for Any Anxiety Disorder come from

the Part 2 subsample (n¼ 7,435).

1169K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

and without chronic pain, for each ethnic group, areshow in Table 4. This table also reports the odds ofany mood and any anxiety disorder among those withchronic pain (relative to those without chronic pain)for each ethnic group. Chronic pain was significantlyassociated with both mood disorder and anxiety disor-der in each of the ethnic groups, with ORs for mooddisorder among those with chronic pain (relative tothose without pain) ranging from 1.69 to 2.09, andORs for anxiety disorders among those with chronicpain ranging from 1.89 to 2.23. A test of the interactionof ethnicity and chronic pain status in the associationwith either mood disorder or anxiety disorder foundno significant differences across ethnic groups.

Cardiovascular disease and mental disordersTable 5 shows that the relationship between cardio-

vascular disease and 12-month mood disorder fell justshort of significance for the total population (OR1.42; 95% CI: 0.98, 2.06), and was not significant forany one ethnic group. The relationship between cardio-vascular disease and 12-month anxiety disorder wasa little stronger, being significant for the total populationand the Other ethnic group, with a similar (though non-significant) pattern of higher odds of anxiety disorderamong Maori and Pacific with cardiovascular disease.A test of the interaction of ethnicity and cardiovasculardisease in the odds of either mood disorder or anxietydisorder found no significant differences across ethnicgroups.

Diabetes and mental disordersNo significant associations were found between

diabetes and either mood disorder or anxiety disorderin any ethnic group, or for the total population

(Table 6). No differences were found across ethnic groupsin the relationship between diabetes and mental disorders.

Respiratory disease and mental disordersAs Table 7 shows, the association between respira-

tory disease and mood disorder was significant for thetotal population, for the Other and Pacific ethnic groups,and marginally significant for Maori. A slightly weakerrelationship was apparent between respiratory diseaseand anxiety disorders with significant associationsonly for the total population and the Other ethnic group.A test of the interaction between ethnicity and respira-tory disease in the association with anxiety disorderswas not significant, but it was significant in the associ-ation with mood disorders (Wald F: 3.50, df: 2,p¼ 0.03). This interaction is a function of the factthat while the Pacific group had similar odds of mooddisorder as the Other group among those without respi-ratory disease, and lower odds than Maori, they hadhigher odds of mood disorder relative to the other twoethnic groups among those with respiratory disease.

Table 4

Prevalence of any mood disorder or any anxiety disorder among persons with and without chronic pain,a and odds of mental disorder among those

with the physical condition (relative to those without)

Ethnicity Any 12-month mood disorderb Any 12-month anxiety disorderb

No chronic

pain %

(95% CI)

With chronic

pain %

(95% CI)

OR

(95% CI)

No chronic

pain %

(95% CI)

With chronic

pain %

(95% CI)

OR

(95% CI)

Maori 9.5

(7.7, 11.3)

14.9

(12.0, 17.8)

1.69

(1.22, 2.34)

15.5

(13.0, 18.0)

26.1

(22.2, 30.0)

1.98

(1.52, 2.57)

Pacific 7.2

(5.3, 9.1)

11.4

(7.9, 15.0)

1.68

(1.10, 2.57)

14.0

(11.0, 16.9)

23.3

(18.0, 28.7)

1.89

(1.26, 2.85)

Other 5.7

(4.9, 6.5)

11.0

(9.4, 12.5)

2.09

(1.68, 2.58)

10.6

(9.4, 11.7)

20.4

(18.3, 22.4)

2.23

(1.86, 2.53)

Total Population 6.2

(5.4, 6.9)

11.5

(10.1, 12.3)

2.00

(1.67, 2.40)

11.3

(10.3, 12.3)

21.2

(19.4, 23.0)

2.17

(1.86, 2.53)

All estimates are adjusted for age and sex.a Chronic pain: arthritis or rheumatism; chronic back or neck problem; frequent or severe headaches; and any other chronic pain.b DSM-IV CIDI 3.0 disorders with hierarchy.

1170 K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

Further analyses indicated that the interaction be-tween ethnicity and respiratory disease in associationwith mood disorder was more a function of asthmarather than chronic obstructive pulmonary disease,with the odds of mood disorder amongst Pacific peoplewith asthma (relative to those without asthma) of 3.05(95% CI: 1.73, 5.38) compared to an OR of 1.32(95% CI: 0.91, 1.90) among Maori and 1.59 (95% CI:1.23, 2.05) among the Other group.

Discussion

This study assessed whether there were significantethnic differences in the associations between selectedchronic physical conditions and mental disorders. De-spite ethnic differences in both the prevalence of phys-ical conditions and the prevalence of mental disorders,

Table 5

Prevalence of any mood disorder or any anxiety disorder among persons w

among those with the physical condition (relative to those without)

Ethnicity Any 12-month mood disorderb

No cardiovascular

%

(95% CI)

With cardiovascular

%

(95% CI)

OR

(95%

Maori 11.3

(9.8, 12.9)

12.8

(6.3, 19.4)

1.15

(0.62,

Pacific 8.1

(6.3, 9.9)

12.0

(2.9, 21.0)

1.54

(0.64,

Other 7.4

(6.6, 8.1)

10.2

(6.2, 14.2)

1.44

(0.91,

Total Population 7.8

(7.1, 8.5)

10.7

(7.3, 14.1)

1.42

(0.98,

All estimates are adjusted for age and sex.a Cardiovascular disease: stroke; heart attack; and heart disease.b DSM-IV CIDI 3.0 disorders with hierarchy.

almost no ethnic differences were found in the associ-ation between the two. The one exception was the find-ing that the association between respiratory disease(asthma, principally) and mood disorder was signifi-cantly stronger in the Pacific group relative to the othergroups, a finding which is hard to interpret in isolation.Independent of ethnicity, there were robust associationsbetween chronic pain and both mood and anxiety dis-orders, a significant association between cardiovasculardisease and anxiety disorders, and a significant associ-ation between respiratory disease and both mood andanxiety disorders.

The focus of this paper is on the question of ethnicdifferences in these associations rather than the associ-ations themselves, but it can be noted that the mentalephysical comorbidities observed here are consistentwith other literature on mental disorder associations

ith and without cardiovascular disease,a and odds of mental disorder

Any 12-month anxiety disorderb

CI)

No cardiovascular

%

(95% CI)

With cardiovascular

%

(95% CI)

OR

(95% CI)

2.14)

18.9

(16.5, 21.3)

28.2

(17.3, 39.1)

1.73

(0.96, 3.12)

3.74)

15.9

(13.3, 18.5)

26.8

(13.1, 40.5)

1.96

(0.92, 4.15)

2.29)

13.8

(12.7, 14.9)

19.7

(14.6, 24.8)

1.56

(1.10, 2.21)

2.06)

14.5

(13.5, 15.4)

21.2

(16.7, 25.8)

1.62

(1.21, 2.18

Table 6

Prevalence of any mood disorder or any anxiety disorder among persons with and without diabetes, and odds of mental disorder among those with

the physical condition (relative to those without)

Ethnicity Any 12-month mood disordera Any 12-month anxiety disordera

No diabetes

%

(95% CI)

With diabetes

%

(95% CI)

OR

(95% CI)

No diabetes

%

(95% CI)

With diabetes

%

(95% CI)

OR

(95% CI)

Maori 11.2

(9.7, 12.8)

16.5

(7.3, 25.7)

1.58

(0.78, 3.20)

19.2

(16.9, 21.6)

23.9

(13.9, 33.9)

1.34

(0.75, 2.39)

Pacific 8.2

(6.4, 10.0)

9.6

(2.4, 16.8)

1.19

(0.50, 2.84)

15.9

(13.3, 18.6)

22.5

(6.8, 38.2)

1.54

(0.59, 4.06)

Other 7.4

(6.7, 8.2)

9.3

(4.3, 14.3)

1.29

(0.69, 2.38)

14.0

(12.9, 15.1)

17.5

(11.2, 23.8)

1.31

(0.83, 2.08)

Total Population 7.9

(7.2, 8.6)

10.3

(6.3, 14.4)

1.36

(0.86, 2.14)

14.7

(13.7, 15.6)

18.9

(13.6, 24.1)

1.37

(0.95, 1.97)

All estimates are adjusted for age and sex.a DSM-IV CIDI 3.0 disorders with hierarchy.

1171K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

with chronic pain, cardiovascular and respiratory dis-ease (Evans et al., 2005; Harter et al., 2003; Scottet al., 2007). The cross-sectional nature of this studyprecludes conclusions about the direction of the rela-tionships observed.

The generally consistent patterns of physicalemental associations across ethnic group may seemcounter-intuitive at first, given that these three groupsdiffer in physical health status as observed here and inprior research (Ministry of Health, 2004), in mentalhealth status (Baxter et al., 2006), in socioeconomic sta-tus (Oakley Browne, Wells, & Scott, 2006) and inmodels of health (Scott, Sarfati, Tobias, & Haslett,2000). However, it is important to note that the possibil-ity that mentalephysical comorbidity patterns mightdiffer across ethnic subgroups is not predicated on dif-ferences between ethnic groups in the prevalence ofphysical and mental conditions, as may be supposed.

Table 7

Prevalence of any mood disorder or any anxiety disorder among persons with

those with the physical condition (relative to those without)

Ethnicity Any 12-month mood disorderb

No respiratory

disease %

(95% CI)

With respiratory

disease %

(95% CI)

OR

(95% C

Maori 10.5

(8.8, 12.2)

13.9

(10.3, 17.4)

1.38

(0.96,

Pacific 6.6

(4.9, 8.4)

17.9

(11.2, 24.7)

3.12

(1.78,

Other 6.9

(6.1, 7.7)

10.3

(8.2, 12.3)

1.56

(1.22,

Total Population 7.2

(6.5, 8.0)

11.0

(9.3, 12.7)

1.60

(1.30,

All estimates are adjusted for age and sex.a Respiratory disease: asthma; chronic obstructive pulmonary disease; emb DSM-IV CIDI 3.0 disorders with hierarchy.

The strength of association between physical and men-tal conditions is tested by the comparison of the preva-lence of, for example, mood disorder among those withand without respiratory disease; an ethnic group differ-ence in that association would require the differential inmood disorder prevalence among those with and with-out respiratory disease to vary significantly across eth-nic groups. The ethnic group differences in mood andrespiratory disease prevalences themselves have no nec-essary implication for group differences in the strengthof association between these conditions. This point isexemplified in this study where Maori have higher prev-alences of both respiratory disease and mood disordercompared with the Pacific group, but it is the Pacificgroup which shows the strongest association betweenrespiratory disease and mood disorder.

However, although group differences in disorderprevalence are not necessarily predictive of differences

and without respiratory disease,a and odds of mental disorder among

Any 12-month anxiety disorderb

I)

No respiratory

disease %

(95% CI)

With respiratory

disease %

(95% CI)

OR

(95% CI)

2.00)

18.6

(16.0, 21.1)

21.6

(17.2, 26.1)

1.22

(0.89, 1.66)

5.47)

16.1

(13.2, 19.0)

17.8

(12.0, 23.6)

1.13

(0.71, 1.81)

2.01)

13.2

(12.1, 14.3)

18.3

(15.3, 21.1)

1.49

(1.18, 1.87)

1.97)

13.9

(12.9, 14.9)

18.7

(16.2, 21.2)

1.44

(1.19, 1.74)

physema; and other chronic lung disease.

1172 K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

in physicalemental associations, group differences inthe factors that are influential in whether a mental disor-der develops out of a physical disorder (or vice versa)should lead to differences in the comorbidities. Suchfactors may be either facilitative in the developmentof, say, a comorbid mental disorder among those witha physical disorder (e.g., physical disease severity), orprotective against its development (e.g., close familysupport). The fact that almost no ethnic groups differ-ences were observed in mentalephysical associationsin the present study is therefore consistent with eitherof the following two possibilities: (a) facilitatory andprotective factors are both operating in the same ethnicgroup (e.g., a given ethnic group may be characterizedby both greater physical disease severity at time ofdiagnosis, but also closer family support), effectivelycanceling out the manifestation of ethnic group differ-ences in mentalephysical associations; or (b) whateverfactors are conducive to the development of a mentaldisorder from a physical disorder (or vice versa), theyare unaffected by the cultural differences manifest inthese three ethnic groups.

A limitation of this study is that physical conditionswere ascertained by a standard checklist, rather thanphysician’s examination. A distinction may be drawnbetween those conditions which were ascertained viaself-report of symptoms (the pain conditions) and thoseascertained via self-report of a physician’s diagnosis(respiratory conditions, heart disease, and diabetes).Self-report of chronic pain conditions has reasonablevalidity given that these are largely self-defined. Forthe other conditions, while acknowledging the limita-tion of self-report, methods research indicates thatself-report of diagnosis generally shows good agree-ment with medical records data (Kehoe, Wu, Leske, &Chylack, 1994; Kriegsman et al., 1996; NCHS, 1994),and importantly, the presence of depressive or anxietysymptoms has not been found to bias or inflate theself-report of diagnosed physical conditions (Kolk,Hanewald, Schagen, & Gijsbers van Wijk, 2002).

Another limitation is that the cell sizes for somecomorbidities in the smaller ethnic groups were small.Nonetheless, although some relationships were not sig-nificant within the smaller ethnic groups, the magnitudeand pattern of relationships was generally similar acrossthe ethnic groups. These small cell sizes occurred despitea generally large survey sample size, with substantialoversampling of the smaller ethnic groups, which illus-trates the logistical challenge of examining this topic.

In sum, mentalephysical disorder associations oc-curred with considerable consistency across three ethnicsubgroups in New Zealand, groups which nonetheless

differ in mental and physical health status. The resultssuggest that whatever factors are conducive to the devel-opment of a mental disorder from a physical disorder (orvice versa), they are either unaffected by the cultural dif-ferences manifest in these three ethnic groups; or, anycultural factors operating serve to both increase and de-crease comorbidity such that they cancel each other out.

Acknowledgement

We thank the WMH staff for assistance with instru-mentation, fieldwork and data analysis. These activitieswere supported by the US National Institute of MentalHealth (R01MH070884), the John D and Catherine TMacArthur Foundation, the Pfizer Foundation, theUS Public Health Service (R13-MH066849, R01-MH069864, and R01-DA016558), the Fogarty Inter-national Center (FIRCA R01-TW006481), the PanAmerican Health Organization, Eli Lilly and Company,Ortho-McNeil Pharmaceutical, Inc., GlaxoSmithKline,and Bristol-Myers Squibb. WMH publications arelisted at http://www.hcp.med.harvard.edu/wmh/. Othermembers of the NZMHS Research Team are: M.A.Oakley Browne, J.E. Wells, J. Kokaua, T. K. Kingi,R. Tapsell, S. Foliaki, M. H. Durie, C. Tukuitongaand C. Gale. We thank the Kaitiaki Group and PacificAdvisory Group for their input and support for thissurvey and we thank all the participants.

References

APA. (1994). DSM-IV: Diagnostic and statistical manual of mental

disorders. (4th ed.). Washington, DC: American Psychiatric

Association.

Batstra, L., Bos, E. H., & Neeleman, J. (2007). Quantifying psychi-

atric comorbidity. Social Psychiatry & Psychiatric Epidemiology,

37, 105e111.

Baxter, J., Kokaua, J., Wells, J. E., McGee, M. A., & Oakley

Browne, M. A. (2006). Ethnic comparisons in Te Rau Hinengaro:

the New Zealand Mental Health Survey (NZMHS). Australian

and New Zealand Journal of Psychiatry, 40, 905e913.

Buist-Bouwman, M. A., de Graaf, R., Vollebergh, W. A. M., &

Ormel, J. (2005). Comorbidity of physical and mental disorders

and the effect on work-loss days. Acta Psychiatrica Scandinavica,111, 436e443.

Ciechanowski, P. S., Katon, W. J., & Russo, J. E. (2000). Depression

and diabetes. Archives of Internal Medicine, 160, 3278e3285.

Dew, M. A. (1998). Psychiatric disorder in the context of physical

illness. In B. P. Dohrenwend (Ed.), Adversity, stress and psycho-

pathology. New York, NY: Oxford University Press.

Evans, D. L., Charney, D. S., & Lewis, L., et al. (2005). Mood

disorders in the medically ill: scientific review and recommenda-

tions. Biological Psychiatry, 58, 175e189.

Harris, E. C., & Barraclough, B. (1998). Excess mortality of mental

disorder. British Journal of Psychiatry, 173, 11e53.

1173K. Scott et al. / Social Science & Medicine 66 (2008) 1165e1173

Harter, M. C., Conway, K. P., & Merikangas, K. R. (2003). Associa-

tions between anxiety disorders and physical illness. European

Archives of Psychiatry & Clinical Neuroscience, 253(6), 313e320.

Katon, W. J., & Ciechanowski, P. S. (2002). Impact of major

depression on chronic medical illness. Journal of Psychosomatic

Research, 53, 859e863.

Kehoe, R., Wu, S.-Y., Leske, M. C., & Chylack, L. T. (1994). Com-

paring self-reported and physician reported medical history.

American Journal of Epidemiology, 139, 813e818.

Kessler, R. C., Ormel, J., Demler, O., & Stang, P. E. (2003). Comorbid

mental disorders account for the role impairment of commonly

occurring chronic physical disorders: results from the National

Comorbidity Survey. Journal of Occupational and Environmental

Medicine, 45(12), 1257e1266.

Kessler, R. C., & Ustun, B. (2004). The World Mental Health

(WMH) Survey Initiative version of the World Health Organiza-

tion (WHO) Composite International Diagnostic Interview

(CIDI). International Journal of Methods in Psychiatric Research,13, 93e121.

Kolk, A. M., Hanewald, G. J., Schagen, S., & Gijsbers van

Wijk, C. M. (2002). Predicting medically unexplained physical

symptoms and health care utilization. A symptom-perception

approach. Journal of Psychosomatic Research, 52(1), 35e44.

Korn, E. L., & Graubard, B. I. (1998). Confidence intervals for pro-

portions with small expected number of positive counts estimated

from survey data. Survey Methodology, 24(2), 193e201.

Kriegsman, D. M., Penninx, B. W., Van Eijk, J. T., Boeke, A. J., &

Deeg, D. J. (1996). Self-reports and general practitioner informa-

tion on the presence of chronic diseases in community dwelling

elderly. Journal of Clinical Epidemiology, 49, 1407e1417.

van Melle, J. P., de Jonge, P., & Spijkerman, T. A., et al. (2004). Prog-

nostic association of depression following myocardial infarction

with mortality and cardiovascular events: a meta-analysis. Psy-chosomatic Medicine, 66(6), 814e822.

Ministry of Health. (2004). A portrait of health: Key results of the

2002/03 New Zealand Health Survey. Wellington: Ministry of

Health.

NCHS. (1994). Evaluation of National Health Interview Survey diag-

nostic reporting. Vital and Health Statistics 2, 120, 1e116.

Oakley Browne, M. A., Wells, J. E., & Scott, K. M. (2006). Te RauHinengaro: The New Zealand Mental Health Survey. Wellington:

Ministry of Health.

Ortega, A. N., Feldman, J. M., Canino, G., Steinman, K., & Algeria, M.

(2006). Co-occurence of mental and physical illness in US Latinos.

Social Psychiatry & Psychiatric Epidemiology, 41, 927e934.

Scott, K. M., Bruffaerts, R., & Tsang, A. (2007). Depression-anxiety

relationships with chronic physical conditions: results from the

World Mental Health surveys. Journal of Affective Disorders,

103, 113e120.

Scott, K. M., Oakley Browne, M. A., McGee, M. A., & Wells, J. E.

(2006). Mental-physical comorbidity in Te Rau Hinengaro: the

New Zealand Mental Health Survey (NZMHS). Australian and

New Zealand Journal of Psychiatry, 40, 882e888.

Scott, K. M., Sarfati, D., Tobias, M. I., & Haslett, S. J. (2000). A chal-

lenge to the cross-cultural validity of the SF-36 health survey:

Factor structure in Maori, Pacific and New Zealand European eth-

nic groups. Social Science & Medicine, 51(11), 1655e1664.

Shah, B. V. (1998). Linearization methods of variance estimation. In

P. Armitage, & T. Colton (Eds.), Encyclopedia of biostatistics (pp.

2276e2279). Chichester: John Wiley and Sons.

Simon, G., Ormel, J., Von Korff, M., & Barlow, W. (1995). Health

care costs associated with depressive and anxiety disorders in

primary care. American Journal of Psychiatry, 152, 352e357.

Sullivan, M. D., LaCroix, A. Z., Baum, C., Grothaus, L. C., &

Katon, W. J. (1997). Functional status in coronary artery disease:

a one-year prospective study of the role of anxiety and depres-

sion. American Journal of Medicine, 103, 348e356.

Wells, J. E., Oakley Browne, M. A., Scott, K. M., McGee, M. A.,

Baxter, J., & Kokaua, J. (2006). Te Rau Hinengaro: the New

Zealand Mental Health Survey (NZMHS): overview of methods

and findings. Australian and New Zealand Journal of Psychiatry,40, 835e844.

Wells, K. B., Golding, J. M., & Burnam, M. A. (1989a). Affective,

substance use, and anxiety disorders in persons with arthritis,

diabetes, heart disease, high blood pressure, or chronic lung

conditions. General Hospital Psychiatry, 11(5), 320e327.

Wells, K. B., Golding, J. M., & Burnam, M. A. (1989b). Chronic

medical conditions in a sample of the general population with

anxiety, affective, and substance use disorders. American Journal

of Psychiatry, 146(11), 1440e1446.

Zhang, X., Norris, S. L., Gregg, E. W., Cheng, Y. J., Beckles, G., &

Kahn, H. S. (2005). Depressive symptoms and mortality among

persons with and without diabetes. American Journal of Epidemi-

ology, 161, 652e660.