Social Psychiatry and Psychiatric Epidemiology

, Volume 43, Issue 2, pp 87–95

Ethnic Norwegian and ethnic minority adolescents in Oslo, Norway

A longitudinal study comparing changes in mental health

Authors

    • Centre for Child and Adolescent Mental Health, Eastern and Southern Norway
  • Lars Lien
    • Institute of General Practise and Community MedicineUniversity of Oslo
  • Anne Johanne Søgaard
    • Institute of General Practise and Community MedicineUniversity of Oslo
    • Norwegian Institute of Public Health
  • Espen Bjertness
    • Institute of General Practise and Community MedicineUniversity of Oslo
  • Sonja Heyerdahl
    • Centre for Child and Adolescent Mental Health, Eastern and Southern Norway
ORIGINAL PAPER

DOI: 10.1007/s00127-007-0275-z

Cite this article as:
Sagatun, Å., Lien, L., Søgaard, A.J. et al. Soc Psychiat Epidemiol (2008) 43: 87. doi:10.1007/s00127-007-0275-z

Abstract

Background

Little is known about ethnic disparities in mental health during late teens. The aim of this study was to compare changes in self reported mental health between adolescents with ethnic Norwegian and ethnic minority background aged 15-16 years followed for three years.

Methods

The youth part of the Oslo Health Study constituted the baseline of this self-reported longitudinal study, carried out in schools in 2001 (n = 3811). The follow-up in 2004 was conducted partly in school and partly through mail. A total of 2489 (1112 boys and 1377 girls) participated in the follow-up. Twenty percent of the participants had an ethnic minority background. Mental health was measured by the Strengths and Difficulties Questionnaire (SDQ) and The Hopkins Symptom Checklist (HSCL-10).

Results

Ethnic minority boys and girls reported poorer mental health than ethnic Norwegians of the same sex, both at baseline and follow-up. Exceptions were hyperactivity-inattention problems and prosocial behaviours where no differences were found. Consistent changes from baseline to follow-up were; an increase in mental distress and prosocial behaviour. No ethnic disparities were found for changes in mental health from ages 15 to 18 between the two groups. There was no different effect of perceived family economy, parents’ marital status and socioeconomic region of residence in Oslo on change in mental health between ethnic Norwegian and ethnic minority boys and girls from age 15 to 18 years.

Conclusions

Ethnic disparities in mental health remained the same from age 15-16 and throughout teenage years. Demographic factors adjusted for had no different impact on changes in mental health between host and immigrant adolescents.

Key words

adolescentmental healthSDQimmigrantslongitudinal study

Introduction

Psychosocial problems, including emotional, behavioural and peer problems are highly prevalent among adolescents and may severely interfere with everyday functioning [9, 11, 34]. A common feature in most studies is the great gender variation in the development of mental health problems among adolescents, with girls reporting more emotional symptoms and distress and boys more behavioural problems [2, 34, 36, 44, 49].

Prospective studies show that mental distress and emotional problems are increasing from puberty up to early adulthood in girls, with a more stable trend in boys [10, 36], while behavioural problems are peaking in mid- and late-adolescence for both boys and girls somewhat depending on type of behaviour [35, 44].

With an increasing influx of immigrants to the Western parts of the world, cultural/ethnic background has become an important variable when studying mental health in a public health perspective [7, 22].

In some cross sectional studies, using self-report measures, a higher prevalence of mental health problems among immigrants compared to the host population has been reported [19, 26, 29]; while others have found no difference [20, 25, 41, 48]; or a lower frequency among immigrant compared to host adolescents [12, 24, 41]. Even when including studies examining both parental and teachers’ report, no general conclusion of the psychological development and health of migration can be drawn [42, 48]. Beyond methodological dissimilarities in the studies, the various associations reported might reflect different challenges for immigrants in the host countries. Variability between different immigrant groups also appears to be considerable [6]. However, adolescent immigrants have some common tasks: They have to function in different cultural contexts, i.e. those of their family, ethnic community, peers and institutions of the wider society, which can lead to a feeling of alienation [6]. Such difficulties in a period of life when also facing the developmental challenges of adolescence, may induce behavioural and emotional problems [3].

Population based longitudinal studies of adolescents focusing on mental health in different ethnic groups are scarce [18, 27]. One study explored trends among different ethnic groups in a representative US sample from adolescence into young adulthood [18]. Another study, also following adolescents into young adulthood, compared development of mental health problems between the Dutch host population and the Turkish immigrants [27]. Both studies found that ethnic disparities in mental health problems decreased as the adolescents entered young adulthood. We do not know of any population based longitudinal study comparing changes in mental health during late teens between immigrants and the host population.

In a longitudinal population based study following approximately 2,500 adolescents, of which 20% had an ethnic minority background, we collected data about mental health—i.e. mental distress, emotional symptoms, conduct problems, hyperactivity–inattention problems, peer problems and prosocial behaviours. The participants also reported information about known confounders like family economy [8, 23], parents marital status [43] and neighbourhood socioeconomic level [30, 39].

The aim of this study was to compare changes in self reported mental health; between adolescents with ethnic Norwegian and ethnic minority background aged 15–16 years followed for 3 years.

Methods

Baseline study

All 10th graders in Oslo during the school years 1999–2000 and 2000–2001 were invited to enter the youth part of the Oslo Health Study. All parents received written information and the students completed a consent form before participation. The students filled in two questionnaires during two school classes. A project assistant was present in the classroom to inform the students about the survey and to administer the questionnaires. For students not present on the day of the survey, questionnaires were left at school with instructions for the teachers on how to administer them. Those who still did not respond were sent a copy by mail to their home address, together with a prestamped, return envelope. A more detailed description has been published elsewhere [46]. From the total population of 15–16-year-olds, 7343 (88%) participated. Those participating in the 2000–2001 cohort (n = 3811) constituted the baseline of our longitudinal study and were invited again in 2004 (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00127-007-0275-z/MediaObjects/127_2007_275_Fig1_HTML.gif
Fig. 1

Flow-chart of the study population

Follow-up study

The follow-up study was carried out partly as a school-based survey and partly through mail. The procedure of the school-based part of the study was similar to the baseline. All secondary high schools in Oslo took part and the 13th grade students were asked to fill out one questionnaire during one school class.

The participants in the baseline study (2000–2001) who were not enrolled in the 13th grade of secondary high school in Oslo and who at baseline had consented to participate in a future follow-up, were invited to participate by mail. The invitation included an invitation letter, an information brochure, a consent form, the questionnaire and a prestamped return envelope. Two reminders were sent to those who did not respond. The study is described more thoroughly elsewhere [38].

Study population

Of the participants in the baseline study 2,489 (65%) participated in the follow-up and consented a linkage between the two datasets (Fig. 1). The response rate was 58% in boys (n = 1,112) and 74% in girls (n = 1,377). Most of the participants in the follow-up took part in school (85%). Among ethnic Norwegians 70% participated in the follow-up, while 54% participated among ethnic minorities. Among those participating at both time points, 20% (n = 505) had an ethnic minority background (ethnic minority; definition see below) and 95% of these were from non-Western countries.

Measures

Mental health

We used the self-report version of the Strengths and Difficulties Questionnaire (SDQ) [37]. The SDQ is a multi informant 25-item wide-angle screening questionnaire with five scales, each consisting of five items, generating scores for emotional symptoms, conduct problems, hyperactivity–inattention, peer problems, and prosocial behaviour. Each item can be answered with ‘not true’ (0), ‘somewhat true’ (1) or ‘certainly true’ (2). The first four problem scales are summed to generate a total difficulties score ranging from 0 to 40. The prosocial subscale measures the adolescent’s ability to act prosocially, independent of the difficulties measured by the other scales. This subscale assesses positive behaviours (with a range of scores 0–10), like caring and helpful behaviour. SDQ is designed and validated for youngsters (11–16 years), but SDQ has also been used for older youths [33]. In the follow-up questionnaire minor linguistic changes were made in accordance with the approved Norwegian translation. The internal consistency (Cronbach’s alpha) of the various SDQ scales at baseline and follow-up were: 0.73, 0.77 for the total difficulties score; 0.70, 0.73 for emotional symptoms; 0.47, 0.38 for conduct problems; 0.54, 0.65 for hyperactivity–inattention; 0.53, 0.57 for peer problems; and 0.64, 0.61 for prosocial behaviour. The internal consistency was about the same for ethnic Norwegians and ethnic minorities at both time points. In accordance with other studies, the Cronbach’s alpha values were low for some of the subscales [15], particularly for conduct problems [32]. The problem scales are based on current nosological concepts [15]. Conduct problem items cover selected essential criteria for oppositional defiant disorder and conduct disorder [1]. A large SDQ validation study found that a high score on the self-report conduct problems was associated with an odds ratio of 7.1 for having conduct or oppositional-defiant disorder [15]. The self-report version of SDQ has also shown satisfactory discrimination between community and clinical samples [16].

The Hopkins Symptom Checklist, 10 item version (HSCL-10) comprises ten questions regarding psychological symptoms of depression and anxiety (mental distress) experienced the previous week [45]. For each question there are four possible answers; ranging from “not troubled” (1) to “heavily troubled” (4). Missing values were replaced with the sample mean value for each item. Records with three or more missing items were, however, excluded. The average item score is used as a measure of mental distress [45]. The HSCL-25 version has proven to have satisfactory validity and reliability as a measure of mental distress in adults [13, 14], and the ten questions version performs almost as well as the longer versions, also among subjects aged 16–24 years [45]. The consistency (Cronbach’s alpha) of the HSCL—10 scale at baseline and follow-up were 0.86 and 0.88, respectively. The internal consistency was about the same for ethnic Norwegians and ethnic minorities at both time points.

Ethnic background

Ethnicity was determined on basis of the parents’ country of birth. Statistics Norway’s definition of ethnic minorities, as those having both parents born in a country other than Norway, has been applied [21]. Failure to report parents country of birth resulted in exclusion from analysis.

Covariates

The “perceived family economy” question had four options and was dichotomized into those answering “very bad/bad” and “good/very good” based on a question comparing the family economy with other families in Norway.

Parents’ marital status” categorizes those having parents who were married/cohabitant versus the others (i.e. divorced/separated, one or both dead).

Socioeconomic region of residence” in Oslo (neighbourhood socioeconomic level) has previously been classified into four regions based on a social index taking the distribution of unemployment, education, non-Western immigrants and lonely parent into account [31]. The major dividing line concerning health status and health related behaviour was, however, found between East and West [31]. Because of small numbers of participants in some of the four regions in our study when doing separate analysis by gender and ethnicity, we decided to use two regions (East and West) based on the same classification [31] in our analyses.

Since the follow-up part of the study was conducted partly in schools and partly by mail, we constructed a variable called, ‘Invitation group’, dichotomizing between mail and school participation.

Lost to follow-up

Ethnic Norwegian boys and girls who did not participate in the follow-up were characterized at baseline by more mental health problems than those who participated in the follow-up; SDQ total difficulties, (SD) [boys: 10.06 (4.7) vs. 8.26 (4.5) p < 0.001, girls: 11.38 (4.9) vs. 9.96 (4.5) p < 0.001], and HSCL-10 mean score (SD) [boys: 1.38 (0.43) vs. 1.29 (0.35) p < 0.001, girls: 1.68 (0.52) vs. 1.60 (0.53) p = 0.024]. In ethnic minority boys and girls there were no statistical difference at baseline between non-responders and responders in the follow-up with regard to SDQ total difficulties [boys: 10.76 (4.9) vs. 10.02 (5.1) p = 0.124, girls: 11.85 (5.1) vs. 11.71(5.2) p = 0.797] and HSCL-10 score [boys: 1.34 (0.40) vs. 1.48 (0.6) p = 0.514, girls: 1.64 (0.58) vs. 1.69 (0.59) p = 0.341].

Thus, we are comparing changes in a healthier selection of ethnic Norwegians than the selection of ethnic minorities regarding mental health problems/mental distress. Others have shown that subjects who participate after reminders are fairly similar to the non-responders [40]. In an attempt to predict how the attrition might have biased our results we compared change in mental health in ethnic Norwegians who participated after reminders with ethnic Norwegians participating after the primary invitation. Those participating after reminders had a slightly poorer development in mental health than those participating in the primary invitation (mean difference (SE): SDQ total; 1.08 (0.52), p = 0.041: HSCl-10; 0.09 (0.06), p = 0.130).

Ethics

Both protocols were evaluated by the Regional Committee for Medical Research Ethics and were approved by the Norwegian Data Inspectorate. The baseline study, and the part of the follow-up study carried out in the schools, received approval from the school authorities in Oslo.

Statistical analysis

Descriptive statistics were performed by chi-square and independent- and paired sampled t-testes. We used the mean difference between baseline (T1) and follow-up (T2) in SDQ and HSCL-10 (T2 − T1 = Δ) as dependent variable when conducting variance analysis to compare changes in mental health between ethnic Norwegians and ethnic minorities. Invitation group, perceived family economy, parents’ marital status and socioeconomic region of residence were collectively entered as covariates. To study if there was a different effect of the covariates on change in mental health between ethnic Norwegian and ethnic minorities interaction terms were entered one by one (ethnicity* covariate) in MANOVA analysis. Level of statistical significance was set to p < 0.01 because of the high number of analyses. The analyses were done separately for boys and girls. Calculations were performed in SPSS 13.

Results

Descriptive

At baseline ethnic minorities in Oslo were mainly living in the less affluent eastern regions (Table 1). Perceived family economy was lower in ethnic minority adolescents than in Norwegians, but the difference was only statistically significant in girls. Larger proportions of ethnic minority boys and girls had parents that were married/partners compared with Norwegians (Table 1).
Table 1

Baseline characteristics of the study population, by ethnic Norwegian and ethnic minorities boys and girls

Characteristic at T1 among participants at T2

Boys (1112*)

Girls (1377*)

n

Ethnic Norwegian (n = 893) %

Ethnic minority (n = 213) %

p

n

Ethnic Norwegian (n = 1079) %

Ethnic minority (n = 292) %

p

Socioeconomic region of residence

   

<0.001

   

<0.001

    East

548

42.6

84.5

 

746

47.7

87.5

 

    West

535

57.4

15.5

 

583

52.3

12.5

 

Perceived family economy

   

0.077

   

0.001

    Very bad/Bad

289

25.3

31.3

 

410

28.3

38.5

 

    Good/Very good

806

74.7

68.8

 

937

71.7

61.5

 

Parents’ marital status

   

<0.001

   

0.001

    Married/Cohabitants

800

69.5

84.9

 

960

67.9

78.4

 

    Other

304

30.5

15.1

 

409

32.1

21.6

 

* Ethnic background in 6 boys and 6 girls are missing, all participants did not answer every question

Ethnic minority boys and girls reported more emotional symptoms/mental distress, conduct problems and peer problems than Norwegian boys and girls at both time points. Ethnic differences were not present for hyperactivity–inattention problems and prosocial behaviour (Table 2, p ethnic).
Table 2

Mean (M) and standard deviation (SD) in strengths and difficulties questionnaire (SDQ) subscales and Hopkins symptoms checklist (HSCL-10) at baseline (age 15–16) and follow up (age 18–19), by ethnic Norwegian and ethnic minority boys and girls

Measure and time point

Boys*

Girls*

Ethnic Norwegian

Ethnic minority

p, ethnic

Ethnic Norwegian

Ethnic minority

p, ethnic

n**

M

(SD)

p, time

n**

M

(SD)

p, time

n**

M

(SD)

p, time

n**

M

(SD)

p, time

SDQ—emotional symptoms

   

0.128

   

0.654

    

0.024

   

0.582

 

    Baseline

889

1.51

(1.59)

 

207

2.29

(2.00)

 

<0.001

1074

3.21

(2.18)

 

288

4.21

(2.55)

 

<0.001

    Follow up

892

1.61

(1.64)

 

211

2.27

(1.78)

 

<0.001

1076

3.37

(2.36)

 

290

4.27

(2.55)

 

<0.001

SDQ—conduct problems

   

<0.001

   

0.002

    

<0.001

   

0.215

 

    Baseline

889

2.03

(1.64)

 

209

2.38

(1.82)

 

0.010

1076

1.82

(1.33)

 

290

2.06

(1.46)

 

0.011

    Follow up

892

1.60

(1.30)

 

212

1.99

(1.67)

 

0.002

1076

1.62

(1.20)

 

290

1.96

(1.30)

 

<0.001

SDQ—hyperactivity–inattention

   

0.147

   

0.040

    

0.001

   

0.001

 

    Baseline

889

3.24

(1.92)

 

207

3.14

(1.95)

 

0.483

1074

3.65

(1.91)

 

287

3.54

(1.93)

 

0.381

    Follow up

892

3.35

(2.05)

 

212

3.42

(2.11)

 

0.652

1076

3.86

(2.09)

 

291

3.93

(2.20)

 

0.578

SDQ—peer problems

   

0.011

   

0.532

    

0.015

   

0.360

 

    Baseline

888

1.48

(1.50)

 

208

2.23

(1.69)

 

<0.001

1074

1.29

(1.37)

 

288

1.92

(1.48)

 

<0.001

    Follow up

892

1.35

(1.48)

 

212

2.20

(1.66)

 

<0.001

1075

1.40

(1.50)

 

292

2.03

(1.62)

 

<0.001

SDQ—prosocial behavior

   

<0.001

   

<0.001

    

<0.001

   

0.002

 

    Baseline

890

6.95

(1.85)

 

209

6.82

(2.01)

 

0.373

1076

7.94

(1.55)

 

290

8.14

(1.52)

 

0.052

    Follow up

892

7.67

(1.69)

 

212

7.71

(1.80)

 

0.736

1076

8.36

(1.36)

 

291

8.45

(1.56)

 

0.383

SDQ—total difficulties

   

0.112

   

0.512

    

0.038

   

0.133

 

    Baseline

887

8.26

(4.48)

 

207

10.03

(5.11)

 

<0.001

1074

9.96

(4.49)

 

287

11.72

(5.16)

 

<0.001

    Follow up

892

7.91

(4.32)

 

211

9.85

(5.04)

 

<0.001

1075

10.25

(5.01)

 

290

12.17

(5.43)

 

<0.001

HSCL-10

   

<0.001

   

0.005

    

<0.001

   

<0.001

 

    Baseline

890

1.30

(0.35)

 

209

1.37

(0.50)

 

0.059

1074

1.60

(0.53)

 

287

1.70

(0.50)

 

0.017

    Follow up

890

1.35

(0.41)

 

211

1.47

(0.50)

 

0.001

1074

1.71

(0.57)

 

288

1.86

(0.66)

 

<0.001

* General linear model repeated measure showed a statistical interaction between change in time and gender (p < 0,001)

** All did not answer every question

Both ethnic Norwegian and ethnic minority boys and girls who perceived their family economy to be very bad/bad, reported more mental health problems (SDQ-total) and distress (HSCL-10) at baseline than those perceiving their family economy as good/very good (p < 0.001) (data not shown). There was no difference in mental health scores at baseline according to socioeconomic region for either ethnic group (data not shown). Adolescents having parents that were marred/cohabitants reported less mental health problems and distress than the others, but associations were only statistically significant in ethnic Norwegian girls and ethnic minority boys (data not shown).

When simultaneously entering all the covariates in the model, perceived family economy was the only covariate associated with mental health in both ethnic Norwegian and ethnic minorities at baseline. In addition, ethnic minority boys without married/cohabiting parents reported more distress (HSCL-10) than those with married/cohabiting parents.

Changes in mental health in ethnic Norwegians and ethnic minorities

The most consistent statistically significant changes in mental health from 15 to 18 years of age were the increase in HSCL-10 score and in prosocial behaviour, regardless of ethnic background and gender (Table 2, p time and Table 3, crude analysis). Another fairly consistent trend, reaching statistically significance for all groups except ethnic minority girls, was the decrease in conduct problems. A third finding was that hyperactivity–inattention problems increased among girls, in both ethnic Norwegians and ethnic minorities. Peer problems were reduced among ethnic Norwegian boys, but increased in Norwegian girls. This gender pattern was the same in ethnic minorities but not statistically significant.
Table 3

Mean change (T2 − T1 = Δ) with Confidence Interval (CI), both crude and adjusted for invitation group, perceived family economy, socioeconomic region of residence and parents marital status

Measure and Time Point

Boys*

Girls*

Ethnic Norwegian (866**)

Ethnic minority (196**)

p, ethnic

Ethnic Norwegian (1061**)

Ethnic minority (262**)

p, ethnic

Δ

95%CI

Δ

95%CI

Δ

95%CI

Δ

95%CI

SDQ—emotional symptoms

    Crude delta

0.09

−0.03 to 0.21

−0.07

−0.32 to 0.18

0.242

0.15

0.01 to 0.29

0.02

−0.25 to 0.30

0.410

    Adjusted delta

0.09

−0.03 to 0.21

−0.08

−0.34 to 0.18

0.245

0.17

0.03 to 0.31

−0.03

−0.32 to 0.26

0.241

SDQ—conduct problems

    Crude delta

−0.41

−0.51 to −0.30

−0.41

−0.63 to −0.19

0.964

−0.21

−0.29 to −0.12

−0.17

−0.34 to −0.01

0.694

    Adjusted delta

−0.43

−0.54 to −0.33

−0.30

−0.54 to −0.07

0.337

−0.21

−0.30 to −0.13

−0.15

−0.32 to 0.02

0.528

SDQ—hyperactivity–inattention

    Crude delta

0.12

−0.01 to 0.26

0.34

0.05 to 0.62

0.192

0.18

0.05 to 0.30

0.43

0.18 to 0.67

0.071

    Adjusted delta

0.13

−0.01 to 0.27

0.30

0.00 to 0.60

0.326

0.17

0.04 to 0.29

0.45

0.19 to 0.70

0.062

SDQ—peer problems

    Crude delta

−0.13

−0.24 to −0.02

−0.05

−0.28 to 0.18

0.551

0.11

0.02 to 0.21

0.12

−0.07 to 0.31

0.969

    Adjusted delta

−0.13

−0.24 to −0.02

−0.04

−0.29 to 0.20

0.527

0.11

0.02 to 0.21

0.12

−0.08 to 0.32

0.968

SDQ—prosocial behavior

    Crude delta

0.71

0.58 to 0.85

0.90

0.63 to 1.17

0.231

0.45

0.35 to 0.55

0.32

0.13 to 0.51

0.219

    Adjusted delta

0.73

0.59 to 0.86

0.85

0.56 to 1.14

0.456

0.46

0.36 to 0.55

0.29

0.10 to 0.49

0.163

SDQ—total difficulties

    Crude delta

−0.32

−0.61 to −0.02

−0.20

−0.83 to 0.42

0.751

0.23

−0.04 to 0.51

0.38

−0.16 to 0.93

0.635

    Adjusted delta

−0.33

−0.64 to −0.03

−0.13

−0.79 to 0.53

0.588

0.24

−0.04 to 0.52

0.36

−0.22 to 0.94

0.721

HSCL-10

    Crude delta

0.05

0.02 to 0.08

0.12

0.06 to 0.18

0.063

0.10

0.07 to 0.14

0.16

0.09 to 0.24

0.117

    Adjusted delta

0.06

0.03 to 0.09

0.10

0.04 to 0.17

0.193

0.10

0.07 to 0.14

0.17

0.04 to 0.09

0.120

* General linear model repeated measure showed a statistical interaction between change in time and gender (p < 0,001)

** Only participants who answered every question included in the adjusted analysis are included in the crude analysis, thus number of participants (n) are slightly lower than in Table 2

In Table 3 we present the mean change in mental health scores (SDQ and HSCL-10)—both crude and adjusted for covariates in ethnic Norwegians and ethnic minorities separately. Even though some changes in mental health mean scores were more pronounced in ethnic Norwegians than ethnic minorities, there was no statistically significant different change between the two ethnic groups in crude or adjusted analysis (Table 3, p ethnic). The crude and adjusted estimates were very similar, indicating that the covariates had little overall effect on the change scores in both gender and ethnic groups. We found no statistically significant interaction between ethnic group and the covariates (perceived family economy status, parent’s maternal status, socioeconomic region of residence and invitation group) in relation to change in mental health.

Discussion

Ethnic minority boys and girls reported poorer mental health overall than ethnic Norwegians of the same sex, both at baseline and follow-up. Exceptions were hyperactivity–inattention problems and prosocial behaviours where no differences were found between the two groups. No ethnic differences were found for changes in mental health from age 15 to 18 years. Perceived family economy, socioeconomic region of residence, parents’ marital status and invitation group were not differently associated with changes in mental health between ethnic Norwegians and ethnic minorities.

In our sample, ethnic minorities reported more problems than ethnic Norwegians. Other European population studies comparing host and immigrant adolescents mental health by self-report are inconclusive [19, 24, 41, 48]. The differences in mental health between immigrants and host adolescents may be explained by several factors associated with both pre- and post-migration [6], but also factors independent of migration [19, 47]. Pre-migration experienced stress may vary, i.e. depending on whether the migrant is a refugee or a child of a labour-immigrant. Adolescent immigrants may also be second-generation immigrants, and have therefore never had any direct experience with migration themselves. Post-migration factors include psychological processes taking place when immigrants meet the host culture. Berry [4] describes an acculturation model based on two (basic) underlying mental issues: (1) the extent to which the heritage culture and identity are maintained; and (2) the extent to which the relationship among different cultural groups is formed. Based on how the two basic issues are handled by the individual, four acculturation strategies are derived (assimilation; integration; separation; marginalization). The strategies are differently associated with mental health, with marginalization being most strongly related to problems [4].

As mentioned in the introduction, immigrant adolescents have specific challenges (i.e. acculturation processes) that are common to all immigrant group justifying our analysis of all ethnic minorities together. However, various minority groups also experience different challenges. Studying them together might conceal differences in mental health between ethnic groups.

Changes in self reported mental health in late teens among ethnic Norwegians and ethnic minorities

Generally longitudinal studies focusing on changes in mental health among adolescents, find similar patterns of change as we do in both boys and girls: an increase in emotional problems and mental distress [10, 34], and a decline in conduct problems [44]. A Norwegian cross-sectional study comparing self-reported mental health (by SDQ) in different age cohorts [33], found similar pattern in several problem subscales according to age as we found in our longitudinal study. They also reported prosocial behaviour to be higher in late- than mid-teens in both boys and girls [33]. However, there are several methodological differences between the two studies (e.g. design, drop-outs, study population) making a direct comparison of results difficult.

We found that girls had an increase in hyperactivity–inattention problems, but have not found any comparable studies reporting the same trend.

Little has been reported on changes in mental health among immigrant adolescents in comparison with the host population. To our knowledge the only European population based longitudinal study doing such comparisons followed adolescents (11–18 years) into young adulthood (21–28 years) [27]. In this study van Oort et al. [27] found differences in internalizing and externalizing problems between Turkish immigrants and Dutch hosts to decrease as adolescents entered young adulthood. However the Turkish reported more problems than the Dutch at both time points, most pronounced for internalizing problems. Harris et al. [18] reported findings in line with van Oort et al. [27] in a study from the US, even though the ethnic groups/immigrants in the US have a somewhat different history than immigrants in Europe.

Van Oort et al. [27] discussed developmental processes from adolescence into young adulthood as possible explanations for their findings. Ethnic differences in parenting style (conformity versus autonomy), social network (within family versus outside family) and identity development (migrants experience distinct worldviews) are some of the issues discussed. During transition into adulthood social networks outside of the family are growing and the migrant may adapt the behaviour of the host society [5], possibly resulting in a mental health profile similar to the host population. Growing up with different cultures may cause considerable confusion with more problems in adolescence [3, 6]. When identity development is more settled in young adulthood, this could explain the smaller difference in mental health problems [27].

We found a stable difference in mental health between ethnic Norwegians and ethnic minorities from age 15 to 18 years.

Both van Oort et al. [27] and Harris et al. [18] followed adolescents into young adulthood (21–28 and 18–26 years), which may be one explanation why the findings differ from ours. In our study most participants attended 13th grade (final year secondary high school) and are probably living at home with their families. It could be that mechanisms discussed by van Oort et al. [27] have not yet been sufficiently manifested to result in reduction in mental health differences between ethnic minorities compared with ethnic Norwegians. Van Oort et al. [27] reported stronger decrease in problems for older adolescents than younger adolescents.

Our study indicates that the differences in mental health detected at age 15 between ethnic Norwegians and ethnic minorities remain the same through teenage years. This early inequality could be important for future development as differences in mental health problems between host and immigrant adolescents are shown to influence later life: In another study by van Oort and colleagues [28], the same cohorts from the Netherlands are used to study behaviour problem in adolescence and ethnic disparities in social class in adulthood. They concluded that ethnic disparities in occupation level in adulthood could partly be attributed to disparities in mental health between Turkish migrants and Dutch adolescents [28].

The influence of the adolescents’ mental health on their later life course calls attention to the importance of preventing differences in mental health between hosts and immigrants at an early age.

We found no statistical difference between ethnic Norwegians and ethnic minorities in the effects of perceived family economy, socioeconomic region of residence and parents’ marital status on change in mental health. In the longitudinal study from the Netherlands adjustments were made for age, gender and socioeconomic position measured by parents’ occupation [27]. They concluded that associations with mental health were similar for host and immigrants, and that the ethnic groups did not change differently after adjustments. In the study by Harris et al. [18], adjustments were done for socioeconomic status measured by parents education and income, and little influence on the estimates and no changes in race/ethnic disparities in mental health over time were found. Based on limited research, demographic factors seem to have the same impact on changes in mental health over time among both host and immigrant adolescents. It is possible that other background factors, not measured in our study, could have different impact on changes in mental health among hosts and immigrants.

Strengths and limitations

Our study is one of few population based longitudinal studies comparing mental health among host and immigrant adolescents. Another contribution by this study is the use of a multidimensional measure studying mental health (SDQ). However, a limitation with SDQ in our study is the age of the participants at follow-up (18–19 years). Even though we found the internal validity (Cronbach’s alpha) to be about the same at both baseline and follow-up for all subscores, the validity of the SDQ at age 18–19 years is not yet established.

Another concern is the attrition. Of those participating in the baseline study, 70% of the ethnic Norwegians and 54% of the ethnic minorities participated in the follow-up. Ethnic Norwegian participants at follow-up reported fewer symptoms of mental health problems/mental distress at baseline than the ethnic Norwegians lost to follow-up. In ethnic minorities there was no statistical difference at baseline. Ethnic Norwegians participating after reminders had a slightly poorer development in mental health than ethnic Norwegians participating after the primary invitation, indicating that the selection bias might hide a slightly poorer development among ethnic Norwegians [40]. The response rate among ethnic minorities (54%) might also have influenced the results even though there was no difference in mental health scores at baseline. Normally, those with the least successful life trajectory [50] and poorest mental health [17] drop out from prospective studies.

Conclusions

Ethnic minority boys and girls reported more mental health problems than ethnic Norwegians at age 15–16, and the difference remained the same through the teenage years.

There was no different effect of perceived family economy, parents’ marital status and socioeconomic region of residence in Oslo on change in mental health between ethnic Norwegian and ethnic minority boys and girls from age 15 to 18 years.

Based on our findings further research focusing on ethnic disparities in mental health in a longitudinal perspective should try to overcome our limitation of differential losses to follow-up. The specific ethnic groups and the length of stay in the host country should also be taken into consideration.

Acknowledgements

This project has been financed with aid of EXTRA funds from the Norwegian Foundation for Health and Rehabilitation. The data collection was carried out and funded by the Norwegian Institute of Public Health, the University of Oslo, the City of Oslo (baseline). The Centre for Child and Adolescent Mental Health, Eastern and Southern Norway, Oslo contributed to the funding of the follow-up study.

Copyright information

© Springer-Verlag 2008