Ethnicity & Health Migration, social mobility and ...

This article was downloaded by: [King's College London] On: 14 July 2014, At: 05:24 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Ethnicity & Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ceth20

Migration, social mobility and common mental disorders: critical review of the literature and meta-analysis a

J. Das-Munshi , G. Leavey

b e

, S.A. Stansfeld

c d

& M.J. Prince

a

Institute of Psychiatry , King's College London , London , UK

b

University College London , London , UK

a

c

Centre for Psychiatry , Queen Mary University of London , London , UK d

Barts and the London School of Medicine , London , UK

e

Northern Ireland Association for Mental Health , Belfast , UK Published online: 10 Nov 2011.

To cite this article: J. Das-Munshi , G. Leavey , S.A. Stansfeld & M.J. Prince (2012) Migration, social mobility and common mental disorders: critical review of the literature and meta-analysis, Ethnicity & Health, 17:1-2, 17-53, DOI: 10.1080/13557858.2011.632816 To link to this article: http://dx.doi.org/10.1080/13557858.2011.632816

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

Downloaded by [King's College London] at 05:24 14 July 2014

Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions

Ethnicity & Health Vol. 17, Nos. 12, FebruaryApril 2012, 1753

Migration, social mobility and common mental disorders: critical review of the literature and meta-analysis J. Das-Munshia*, G. Leaveyb, S.A. Stansfeldc,d and M.J. Princea


a Institute of Psychiatry, King’s College London, London, UK; bUniversity College London, London, UK; Northern Ireland Association for Mental Health, Belfast, UK; cCentre for Psychiatry, Queen Mary University of London, London, UK; dBarts and the London School of Medicine, London, UK

(Received 2 March 2011; final version received 5 October 2011) Objective. Changes in socio-economic position in people who migrate may have adverse associations with mental health. The main objective of this review was to assess the association of social mobility with common mental disorders in migrant and second-generation groups, to inform future research. Design. Systematic review and meta-analysis of English-language studies assessing the association of social mobility in migrant or second-generation groups with common mental disorders. Approaches to operationalise ‘social mobility’ were reviewed. Results. Twelve studies (n 18,548) met criteria for retrieval. Very few included second-generation groups, and most studies were cross-sectional in design. Approaches to operationalise ‘social mobility’ varied between studies. Downward intragenerational social mobility was associated with migration in the majority of studies. Random effects meta-analysis (n5179) suggested that migrants to higher income countries who experienced downward mobility or underemployment were more likely to screen positive for common mental disorders, relative to migrants who were upwardly mobile or experienced no changes to socio-economic position. Conclusions on second-generation groups were limited by the lack of research highlighted for these groups. Downward intragenerational mobility associated with migration may be associated with vulnerability to common mental disorders in some migrant groups. Conclusion. Given the increasing scale of global migration, further research is needed to clarify how changes to socio-economic position associated with international migration may impact on the mental health of migrants, and in their children. Keywords: migration; social mobility; common mental disorders; refugees; socioeconomic status; ethnic minorities

Introduction Over the past 30 years global migration has increased rapidly as a consequence of industrialisation, changing labour markets, post-colonial factors and displacement related to conflict (Held et al. 1999, Castles 2000). In 2005, global migration was estimated to stand at 195 million (United Nations, 2009), with 10.4 million people recorded as international refugees by the end of 2011 (UN HCR, 2011). *Corresponding author. Email: [email protected] ISSN 1355-7858 print/ISSN 1465-3419 online # 2012 Taylor & Francis http://dx.doi.org/10.1080/13557858.2011.632816 http://www.tandfonline.com


18 J. Das-Munshi et al. Studies of social mobility in people migrating to Britain (Heath and MacMahon 2005, Platt 2005), other parts of Europe (Abadan-Unat 1995, Kogan and Cohen 2008) and North America (Jong and Madamba 2001, Slack and Jensen 2007) suggest that many migrants pay an ‘occupational penalty’ to move and work internationally (Heath and MacMahon 2005, Platt 2005). ‘Occupational penalty’ encompasses working below one’s levels of skills and qualifications (‘underemployment’), or working in occupations below one’s pre-migratory socio-economic position (‘downward mobility’). This may also be thought of as a form of intragenerational social mobility, in that it captures changes to socio-economic position within the same lifecourse. The occupational disadvantage experienced by migrant groups has been attributed to a number of factors. These include racism and discrimination, lack of cultural knowledge and language fluency, or in overseas qualifications not being recognised (Heath and MacMahon 2005). Although some studies have examined the association of social mobility in migrant and second-generation groups with selfreported limiting long-term illness (Harding 2003), general health (Smith et al. 2009) and mortality (Harding and Balarajan 2001), to our knowledge there is scant evidence on the association of social mobility with mental health in migrant groups. The role of employment and social inclusion in promoting well-being and mental health has been widely acknowledged in a number of recent policy statements (Black 2008, World Health Organization 2008). If social inclusion is to be an important element of mental health strategy (Wahlbeck and Taipale 2006), then the factors associated with mental health and the inclusion of labour migrant and refugee groups in the workforce needs to be better understood. In addition, the growth in minority ethnic populations is mostly explicable by the relatively high fertility rates among migrant groups. In this article, we use the term ‘second-generation migrant’ to indicate children born in the country of settlement, to parents born elsewhere. Rates of mental disorder among second-generation groups in many countries are consistently high (King et al. 1994, Leao et al. 2005). The factors which relate to social mobility and mental disorders for these groups are poorly understood (Davey Smith et al. 2000). A British study suggested that second-generation groups were more likely to experience greater upward (intergenerational) social mobility compared to their White British peers and were more likely to end up in a higher social class than their parents (Platt 2005). This may partly be accounted for through the downward mobility which their parents experienced in migrating to Britain (Platt 2005). Other factors promoting intergenerational social mobility may relate to education (Platt 2005), or as suggested by Beiser and colleagues using data from Canada, whereas ‘poverty may represent a transient and inevitable part of the resettlement process for new immigrant families . . . for long-stay immigrant and receiving-society families . . . poverty . . . is not part of an unfolding process; instead, it is the nadir of a cycle of disadvantage’ (Beiser, 2002). Thus the life-course socio-economic trajectories of second-generation groups may not be the same as children without a parental history of migration and may also have differing mental health consequences. There is also a growing body of evidence linking migration as a potential risk for mental disorders, with associations for schizophrenia (Cantor-Graae and Selten 2005) and common mental disorders (Lindert et al. 2009) suggested. Economic circumstances in the host country, alongside reasons for migration, may play an important role in accounting for the differing prevalence of common mental


Ethnicity & Health

19

disorders in migrant groups (Porter and Haslam 2005, Lindert et al. 2009). The wider literature suggests an association between lower socio-economic position and both onset and persistence of depression (Lorant et al. 2003). The way in which this may pattern the mental health of migrant groups is less certain, particularly as departure and settlement may expose people to at least two different economic systems (both in the country of origin and in the receiving country) (Davey Smith et al. 2000). Using data from observational studies (casecontrol, cohort and cross-sectional surveys) we sought to assess the association of downward social mobility (relative to upward/stable social mobility) with common mental disorders in international migrants. We hypothesised that international migrants experiencing downward social mobility would be more likely to screen positive for common mental disorders, relative to those experiencing stable/upward social mobility. The review will also examine how this impacts on the children of migrants, who may have socio-economic trajectories which differ from the non-migrant population in the receiving country (Heath and MacMahon 2005, Platt 2005). Finally, we critically review the methods taken to operationalise ‘social mobility’ in studies of migration and mental health. Methods The guidelines for Meta-analysis of Observational Studies in Epidemiology (MOOSE) were followed in the planning of the review and reporting of findings (Stroup et al. 2000). Systematic searches were conducted in the following databases: Medline, Ovid, Psychinfo, Embase, Social Policy and Practice, British Nursing Index and Archive, by the first author. Searches were conducted from November 2009 to February 2010. Due to limited resources only English-language papers and dissertations were considered. Keywords and mesh headings used for searches were ‘mental disorders’; ‘mental health’, ‘depress*’, ‘anxiety’, ‘psych*’; Intrageneration* mobility, Intergeneration* mobility, underemployment, *mobility, Migra*, immigra*, emigra*, refugee*, asylum*. Searches were supplemented by contacting first authors of each included study, and experts in the field, to ensure that unpublished work could also be reviewed. Bibliography sections of papers as well as one review (Porter and Haslam 2005) were hand-searched for additional references. Where it was unclear if the paper met inclusion criteria the paper was retrieved and examined. If results from a study had been published in more than one source, then all papers were retrieved and presented as findings from one study. Exclusion criteria were: Studies not using a structured instrument or clinical interview to assess psychological well-being and common mental disorders, ecological or qualitative studies, studies not reporting original data. Searches were performed by the first author, discussion between first and last author resolved issues relating to inclusion/exclusion. A ‘log’ of studies with reasons for inclusion/exclusion was also kept (available from first author on request). The quality of studies was assessed using the criteria shown in Figure 1. In acknowledgement of the diversity of ways in which social mobility is defined in the literature (Loury et al. 2005), inclusion criteria for definitions of ‘social mobility’ were kept purposefully broad and simply had to reflect change in socio-economic position across two time points, either within the same generation (‘intragenerational social

20 J. Das-Munshi et al. Quality criteria 1. Cross-sectional, cohort, case-control 2. Response rates (60%) 3. Sampling methods (convenience vs. random) 4. Sample size (50) 5. Attrition (if prospective) 6. Methods to define migrant group or ethnicity 7. Assessment of confounding and interactions 8. Instruments used to assess common mental disorders 9. Methods used to define social mobility, underemployment Source of heterogeneity


1. Reason for migration (labour migrantversus refugee or asylum-seeker) 2. Geographical destination 3. Country of origin

Figure 1. Quality criteria and sources heterogeneity.

mobility’) or between generations (‘intergenerational social mobility’). Post hoc, we decided to critically examine the variety of approaches used to assess social mobility. The main outcome variable in meta-analysis and narrative synthesis was the presence of ‘common mental disorders’. This term refers to neurotic disorders commonly encountered at the level of primary care, associated with disability and impaired functioning. The use of screening tools or diagnostic assessments results in either ‘counts’ of symptoms whereby a higher score indicates greater distress and disability or utilises cut-points usually validated against suitable ‘gold standards’. For the purposes of this review, we retained studies employing both approaches. Ethical approval was not required for this review and meta-analysis.

Statistical analyses Analyses were performed in STATA/IC 10.1 (StataCorp 2007). Meta-analysis was used to assess the association of downward social mobility (relative to upward/stable social mobility) in migrants with common mental disorders. The dependent variable in meta-analysis was common mental disorders as assessed through instruments employing validated cut-points. Using this approach we derived the overall pooled odds ratio (OR) from all of the studies on migrant groups and then investigated this according to the sub-groups/criteria given in Figure 1. Pooled estimates for the odds of downward social mobility and common mental disorders (relative to migrants who experienced stable/upward social mobility) were estimated using DerSimonian and Laird’s (1986) method. In calculating odds ratios, wherever a ‘0’ occurred in any cell of the 22 table, a correction factor of 0.5 was added (Bradburn et al. 1998). Where meta-analysis was not possible (i.e., those studies which employed continuous measures for common mental disorders (CMD) and where the underlying constructs differed) studies were qualitatively appraised against extracted quality criteria (Figure 1), and a narrative synthesis of results was performed.

Ethnicity & Health

21

A priori, we assumed that a random effects meta-analysis would best capture the variability between studies, (e.g., due to differing migratory contexts, geographical regions and study designs).


Assessment of study quality Sources of heterogeneity (Figure 1) were explored by visually examining forest plots and through the I2 statistic (Higgins et al. 2003). Tentative I2 statistic cut-offs to assess heterogeneity are: B25% (low), 50% (moderate) and 75% (high) (Higgins et al. 2003). Meta-regression was then used to assess whether there was formal statistical evidence of a difference in ORs (for the association of downward social mobility with common mental disorders) by a priori sub-groups (Egger et al. 1998, Higgins et al. 2003). Meta-regression assesses the association of covariates (in this case these were the sub-groups shown in Figure 1) in accounting for the association of downward social mobility with common mental disorders. For ease of interpretation, metaregression coefficients were exponentiated, giving a ratio (or a ‘ratio of odds ratios’) (Higgins and Green 2011). Results Twelve papers met inclusion criteria (Figure 2). Two included second-generation groups (Table 1). Studies covered the years 19782010. All studies were conducted in higher income countries and included two studies from North America, three from Europe and Israel and one from New Zealand (Table 1).1 Table 1 summarises characteristics of retrieved studies. Operationalisation of ‘social mobility’ Table 2 summarises the methods used in the studies to assess ‘social mobility’. As with the wider literature, social mobility is not a ‘single homogenous phenomenon, but rather . . . a cluster of interdependent social processes’ (Loury et al. 2005). The studies reviewed here attested to this complexity. ‘Socioeconomic position’ has been preferred over the more commonly used term ‘socio-economic status’ by a number of commentators (Krieger et al. 1997, Muntaner et al. 2004, Galobardes et al. 2007) as ‘status’ has been judged to ‘blur distinctions between two different aspects of socio-economic position . . . actual resources . . . and prestige or rank-related resources’ (Krieger et al. 1997). Corresponding to this, we detected a wide array of terms to characterise changes to socio-economic position. Several studies utilised prestige-based measures to assess changes in socio-economic position. This included one study (Inclan 1983) which used the ‘Hollingshead Index of Social Position’, a measure that combines education with occupational ‘rank’ (cited in Krieger et al. 1997). As discussed by Krieger et al. (1997), ‘occupational rank’ for this scale was determined by the scale’s founder, based on a population in New Hampshire, USA, in the 1960s and so may not have adequately captured ‘rank’ or status in these migrant groups at a later time in different countries. Three Canadian studies used the Blishen Occupational Index to assess socioeconomic position (Eaton and Lasry 1978, Beiser et al. 1993, Aycan and Berry 1996).

22 J. Das-Munshi et al.

90 papers identified

62 excluded as no information on social mobility


2 excluded as used inadequate measure for common mental disorders

10 excluded as no information on migrationonly data on ethnicity or race provided.

2 excluded as data from same study

2 excluded as no analysis of social mobility and common mental disorders 12 papers retained for review

Estimates from 4 papers analysed with metaanalysis

Estimates from 8 papers analysed using narrative synthesis

Figure 2. Flowchart of retrieved studies.

This scale is based on 1970s data from Canada and takes into account years of education and income in the ranking of occupations (Blishen and Roberts 1976). Like the Hollingshead scale, assumptions that occupation ‘ranks’ are the same irrespective of country may be misleading. However, in one of the studies the correlation between education and ‘employment prestige’ was assessed in Vietnamese refugees when comparing their education with their occupation in Vietnam, prior to migration (Beiser et al. 1993). A similar correlation was found for non-migrant

Ethnicity & Health

23

Table 1. Study characteristics. Number of studies Total participants N (%)


Total Included second-generation group

12 2

18,548 359 (2%)

Region or country of origin* ‘Scandinavian’ regions Finland Europe Turkey Former Soviet Union Middle East North Africa Iran Iraq Southeast Asia Central/South America Puerto Rico ‘Outside Scandinavian or European regions’

1 1 1 2 1 1 1 1 2 3 1 1 1

295 (2%) 211 (1%) 335 (2%) 327 (2%) 587 (3%) 124 (1%) 166 (1%) 250 (1%) 553 (3%) 3157 (17%) 1518 (8%) 142 (1%) 479 (3%)

Destination country USA Canada New Zealand Sweden Netherlands Israel

2 3 1 3 2 1

3198 (17%) 1943 (10%) 271 (1%) 11,285 (61%) 1264 (7%) 587 (3%)

Reason for migration Refugee Labour migrant

5 9

2482 (13%) 16,422 (89%)

Key *as described in the original study.

Canadians living in Canada, suggesting cross-cultural comparability (Beiser et al. 1993). Two studies assessed subjective appraisal of changes to status as a result of migration (Lindencrona et al. 2007, Nicklett and Burgard 2009). Both studies were cross-sectional; therefore, findings may have been affected by recall bias. Operationalisation of ‘underemployment’ ‘Downward mobility’ may be thought of as a decline in socio-economic position over time. Several studies also assessed the related concept of ‘underemployment’, whereby people are employed below their level of skill, prior expertise, or training. Although notions of ‘downward mobility’ and ‘underemployment’ clearly overlap, there are also differences, in so far as people may be ‘upwardly mobile’ with respect to their income and the living standards that this affords them, but at the same time may have accepted employment at a level below their skill-set or education. For example, in one cross-sectional analysis which assessed changes to income and occupational status in migrants from Turkey to Canada (Aycan and Berry 1996)


24 J. Das-Munshi et al. whereas ‘income’ in absolute and relative terms was reported as having improved with time, respondents reported ongoing occupational decline, associated with adverse psychological health (Aycan and Berry 1996). This study examined the ‘financial ease’ afforded by monthly incomes, by asking respondents to compare the ease in which they felt their monthly incomes allowed them to participate in certain activities in Turkey compared to Canada (Aycan and Berry 1996). This measure taps into notions of ‘wealth’ as well as income (Galobardes et al. 2007), highlighting the distinction between the two concepts, where income represents material resources at a fixed time-point, whereas ‘wealth’ represents an accumulation of assets (Galobardes et al. 2007). The findings suggested that while participants experienced high levels of downward mobility and underemployment (associated with adverse mental health) in migrating to Canada, income in terms of ‘financial ease’ improved (Aycan and Berry 1996). Conclusions are limited by this study’s cross-sectional design; ‘time’ in this study was assessed according to participant recall of experiences since arrival in Canada. In two studies from Sweden (Tingho¨g et al. 2007, Tingho¨g et al. 2010) ‘status incongruence’ was considered present in blue collar, unemployed (Tingho¨g et al. 2007) or ‘currently studying’ (Tingho¨g et al. 2010) migrants who had education 12 years (Tingho¨g et al. 2007) or to university level (Tingho¨g et al. 2010). Other investigators assessed ‘underemployment’ by directly asking respondents if they thought they were employed in occupations below their level of expertise (Aycan and Berry 1996). Assessment of quality Table 2 summarises design and quality of included studies. Response rates ranged from 45% to 82% (Table 2). Of the types of study design retrieved, only two utilised a longitudinal study design, while the rest employed a cross-sectional design. No casecontrol studies were identified. Three of the studies used convenience sampling (Table 2). Sample sizes ranged from 110 people up to 10,441 people (Table 2). Few studies considered a priori confounders or interactions in the association of social mobility with common mental disorders. Indeed, in the majority of studies, an examination of the association of social mobility with common mental disorders in migrant groups was not the primary objective. None of the studies assessed mental health prior to migration. A variety of instruments were employed across studies to assess common mental disorders. Two studies utilised screening instruments which had been validated in the cultural context of the migrant group (Beiser et al. 1993, Abbott et al. 1999). Differential attrition due to return migration whereby migrants with health problems return to their country of origin may have been an additional source of bias in some studies (Razum 2006). For example, one longitudinal study did not find an association between underemployment and common mental disorders in Vietnamese refugees to Canada, although people lost to follow-up were more likely to be male, depressed and unmarried (Beiser et al. 1993). In the other longitudinal study there was a higher attrition rate amongst Turkish respondents compared to non-migrant Dutch respondents (VanOort et al. 2007); although as this was a study of second-generation Turkish children, attrition through return migration should not have been an issue. In most of the studies the method to define ethnicity or nativity

Table 2. Design features of studies. Sampling strategy, response rate

Method of defining ethnicity/nativity

RS

Country of birth

Finnish, Iraqi and Iranian Total: 720 migrants living in Sweden, aged 2075 and resident for 3 years. 53% female Finnish: 211, Iranian: 250, Iraqi: 259

Not given in paper

Migrants from Puerto Rico, Cuba, Mexico and ‘other Latinos’; people from China, Vietnam, Philippines and ‘other Asian’ countries to USA 1411 males, 1645 females

Demographic details

No. in sample

Labour migrants Tingho¨g et al. CS: 2005 (2010)*

Response rate: 47.9% Nicklett and Burgard, (2009)

CS: 20022003

RS

Weighted response rates: 75.5% (Latino sample), 65.6% (Asian sample) Tingho¨g et al. CS: 19982000 (2007)

RS

Country of birth

1538 Asian respondents

Total: 10,441, Swedenborn: 9314, Scandinavian-born: 295, Europe-born: 335, born outside Europe/ Scandinavian regions: 479

25

Migrants to Sweden: Swedish-born, ‘Scandinavian-born’ (Finland, Norway, Iceland, Denmark), European-born, and ‘born outside of Europe’; Age: 2065. Percentage of male: Swedish-44.5%; Scandinavian-born-34.2%; Europe-born-44.5%; Born outside Europe-49.5%

1518 Latino respondents

Ethnicity & Health


Study design and Author (year) time period



Sampling strategy, response rate


Demographic details

No. in sample

Response rate: 53% VanOort et al. CO: 10-year follow (2007) up

217 Turkish respondents, 753 Dutch respondents Migrants:1993 2003* Dutch: 19871997*

RS

Turkish children who Migrant Turkish children had to have at least one (74%) and secondparent born in Turkey generation Turkish children (26%) growing up in the Netherlands. Percentage of boys: 50% Baseline response: Turkish (Turkish), 46% (Dutch) children 71%, Dutch children 80%

Response at follow-up: Turkish adults: 51%*, Dutch adults: 58% *Baseline assessment *Larger proportion of Turkish children lost through at 1118 years; follow-up assessment out-migration at 2128 years Abbott et al. (1999)

CS 1995

Conv Response rate: 45%

Not given

Migrants to New Zealand. 271 90% from Hong Kong & Taiwan, smaller proportion from China, Malaysia, Singapore & Macau. 1%asylum seekers or refugees. Mean age: 39, Percentage of female: 67%


Table 2 (Continued )


Aycan and Berry (1996)

CS, year not given



Conv RS: 100 Individuals Not given selected from Turkish cafes, restaurants and clubs; 150 individuals randomly selected from membership lists of Turkish organisations. Response rate: 46.8%

Demographic details

No. in sample

Turkish migrants to Canada, unclear generational status. 53.2% held Canadian citizenship

110

Mean age 38.2 (SD: 10.6). Percentage of male: 88%

Inclan (1983)

CS, Year not given

Sampling methodology and response rates not given.

Married second-generation 142 Women with Puerto Puerto Rican women in Rican parents who USA migrated from Puerto Rico and primarily raised in USA or who migrated to USA before completing first grade of school Age range: 1859

Eaton and Lasry (1978)

CS, year not given

RS

Not given

Response rates-not given.

Jewish migrants from North Africa to Canada. Only employed men interviewed

166

Ethnicity & Health



27





Conv

Not given

New arrivals from Middle 124 East at a refugee resettlement programme, in Sweden Percentage born in Iraq: 83%, mean age: 34, percentage of male: 67%

Not given

Migrants from Former Soviet Union to Israel

Demographic details

No. in sample

Refugee or asylum seekers Lindencrona CS: 20022004 et al. (2007)

Response rates not given.

Lerner et al. (2005)

Two CS (different participants at each wave): 1990, 1994/ 1995

Laban et al. (2005)

CS: 1999 and 2000

RS

Response rates-not given

(In 1995 sample), percentage of male: 44%, age: 24% above pension age

Not given RS: Two groups: New arrivalsasylum seekers living in the Netherlands on one day in 1999 Response rate: (New arrivals) 82%; Response rate: (Living in Netherlands for at least 2 years) 79%.

Iraqi asylum seekers to Netherlands

587

New arrivals: 143

Age: ‘New arrivals’- 1824 Living in Holland for (22%), 2524 (42%), 3544 2 years: 151 (15%), 4564 (14%), 64 (8%).



Table 2 (Continued ) Sampling strategy, response rate


Demographic details

No. in sample

‘Lived in the Netherlands (Total: 294) 2 years’: 1824 (9%), 25 34 (49%), 3544 (26%), 45 64 (13%), 64 (3%) Gender (percentage of male): ‘New arrivals’: 49.7%, ‘Living in Holland 2 years’: 78.8% Beiser et al. (1993)

RS

CO: Baseline: 1981, with follow-up at 2 years

Response rates: Refugees Baseline (T1)-92%

Follow-up (T2): 87% of baseline sample

Not given

Chinese and non-Chinese (from Laos, Vietnam, Cambodia) refugees (numbers in sub-groups not given) to Canada Age: 5% of sample 55 years, most (75%) B35 years. Percentage of male: 56% Non-migrant Canadians interviewed at follow-up, age and gender-matched to refugees

At baseline 1348 refugees.

319 non-migrant Canadians included at T2 follow-up for comparison

Canadians (assessed at T2 only)-81% Note: Key and abbreviations: *These studies may have included refugees or asylum seekers; however, it was not possible to differentiate this from the data given; CS, crosssectional; Co, cohort; RS, random sampling; Conv, convenience sample.

Ethnicity & Health



29

30 J. Das-Munshi et al. was conspicuously absent. Where this was present this was always determined through country of birth (Inclan 1983, Tingho¨g et al. 2007, VanOort et al. 2007, Tingho¨g et al. 2010). Key findings Table 3 summarises the research question and reported findings for each of the reviewed studies.


Downward mobility, common mental disorders and migration Random effects meta-analysis suggested that the (pooled) association of downward social mobility with common mental disorders, relative to stable/or upward mobility was 1.56 (95% CI: 1.04, 2.33), in people who had migrated (Figure 3). An I2 of 52% (X72 14.6; p0.042) suggested moderate heterogeneity, such that the association of downward social mobility with common mental disorders varied between groups. This was visually confirmed in the forest plot (Figure 3). There was moderate heterogeneity between studies assessing refugee/asylumseeker groups (p 0.05; I2 67%) which was negligible for studies assessing labour migrants (p 0.94; I2 0.0%). There was a high level of heterogeneity between studies which had assessed social mobility by self-report measures (pB0.001; I2 90.1%), whereas negligible heterogeneity between studies assessing social mobility using ‘objective’ measures such as changes to employment (p0.86; I2 0.0%). There was moderate heterogeneity between studies of people who had migrated from outside Europe/Scandinavian regions (p 0.008; I2 71.2%) which was negligible in studies of migrants from these regions (p 0.70; I2 0.0%). Results of meta-regression are shown in Table 4. 95% CIs spanned the null for each potential source of heterogeneity (Table 4). The residual variance due to heterogeneity was reduced to 0% (with 100% of between-study variance accounted for), when all meta-regression covariates (as displayed in Table 4) were entered into the model. ‘Reason for migration’ (refugee/asylum seeker versus labour migrant) accounted for the largest proportion of the variance. Studies assessing common mental disorders as a continuous measure A narrative synthesis of findings from studies not subjected to meta-analysis is presented here. In one study of migrants to Canada there was an association of subjective status loss and downward mobility with a latent construct of poorer ‘psychological health’ (Aycan and Berry 1996). In another study of refugees from the Middle East to Sweden, subjectively perceived ‘loss of status’ loaded with ‘reported discrimination’ on to an underlying construct which was associated with elevated mean GHQ scores (Lindencrona et al. 2007). A cross-sectional study assessing mental health in Chinese migrants to New Zealand (Abbott et al. 1999) suggested that underemployment was associated with greater psychological morbidity in recent migrants (B2 years), compared to people settled for longer (Abbott et al. 1999). Conversely, two studies did not find associations between downward mobility and psychological morbidity, despite downward mobility being a relatively frequent

Measure for social Author (year) mobility

Instrument to assess common mental disorders

Study question/reference group

Proportion in sample experiencing social mobility and direction of mobility (up/ Findings, confounders down/stable) and potential interactions

Labour migrants Tingho¨g et al. ‘Status incongruence’: Hopkins symptom checklist (2010) People of university education, unemployed or studying at time of assessment

What is the association of N (%) with ‘high status status incongruence and incongruence’. CMD in migrants?

Finnish: 8 (3.8%); Iranian: 52 (20.9%); Iraqi: 30 (12%)

Nicklett and Burgard (2009)

Net movement on 10point scale of ‘subjective social status’ (MacArthur ladder) comparing perceived status in the USA with status in country of origin.

CIDI for major depression

What is the association of social mobility in Latino or Asian migrants to USA with depression, when compared to migrants who report no change to social status?

(Unweighted N, Weighted%): Stable: 549 (17%), 1 step down: 464 (15%); 2 steps down: 420 (14%); 3 steps down: 749 (23%)

1 step up: 261 (10%); 2 steps up: 267 (11%); 3 steps up: 346 (10%)

31

Percentage CMD in ‘high’ status incongruence vs. low/ intermediate status incongruence: Finnishborn: 0 vs. 12.1; Iranianborn: 63.3 vs. 46.5; Iraqiborn: 62.7 vs. 60.0 Adjusted OR (95% CI) of high status incongruence with CMD (full sample) 1.27 (0.60, 2.72) Association* of changes to social status and major depression: OR (95% CI)], no change: 1.00 [ref], 1 step down: 1.79 (0.83, 3.86), 2 steps down: 1.94 (0.95, 3.94), 3 steps down: 2.97 (1.33, 6.61) 1 step up: 1.51 (0.53, 4.26); 2 steps up: 1.10 (0.33, 3.73); 3 steps up: 1.50 (0.49, 4.61)

Ethnicity & Health


Table 3. Summary of findings.



Van Oort et al. (2007)

Comparison of parental occupational class with cohort member’s occupational class in adulthood


Dutch Youth Self Report questionnaire at age 1118 for internalising/ externalising behaviours (childhood). Young Adult Self-Report & Adult Self-Report questionnaires (adulthood)



*Taking into account survey weights, adjusted for: social status in country of origin, ethnicity, gender, age, education, duration of residence in USA, citizenship, English fluency, importance of finding employment in USA 1. Upwardly mobile 56% of second-generation Does upward individuals experienced intergenerational mobility Turkish young adults had improvements in mental moved into a higher lead to improvements in socioeconomic position than disorders over time. 2. mental health and does The association of this differ by ethnic group? their parents, compared to 43% of Dutch young adults intergenerational mobility with mental (p 0.0007) health did not differ in Turkish compared to Dutch cohort members



Measure for social Author (year) mobility Tinghog et al. ‘Status incongruence’: (2007) People who were unemployed or blue collar with 12 years education

Instrument to assess common mental disorders 10-item major depression inventory



What is the association of Percentage of blue collar or ‘status incongruence’ with unemployed people with ‘status incongruence’: depression in migrants? Swedish (not migrant): 78 (4%), Scandinavian migrant: 5 (7%), Other European migrant: 10 (13%) Migrants from outside of Europe: 25 (13%)

Proportions screening positive for depression (in people with ‘status incongruence’ vs. no status incongruence): 1. Swedish non-migrant: 5/ 79 (6%) vs. 193/1515 (11%); 2. Scandinavian migrant: 0/5 (0%) vs. 5/60 (7%); 3. Other European migrant: 3/10 (30%) vs. 12/63 (19%); 4. Migrants from outside Europe: 7/ 25 (28%) vs. 38/154 (25%) No adjustment for confounders.

Ethnicity & Health



33




12-item Chinese Abbott et al. The proportion of Health Questionnaire (1999) people employed in ‘business or professions’ premigration, but no longer employed at this level post-migration.

Study question/reference group In a migrant sample what is the association of underemployment with common mental disorders, compared to people reaching occupational parity?

Proportion in sample experiencing social mobility and direction of mobility (up/ Findings, confounders down/stable) and potential interactions 45% (271) reported being employed in business or professions prior to migration, 18% employed at this level post-migration.

In recent migrants ( B2 years), underemployment postmigration was associated with mean CHQ 12.8; SD 1.7 compared to those reaching occupational parity (mean CHQ 4.6; SD 2.7) p B0.001. No difference in mean CHQ scores in people resident for 2 years. No adjustment for confounders



Measure for social Author (year) mobility Aycan and Berry (1996)


‘Status’ assessed using SEM used to assess income, education and ‘psychological health’ latent variable. occupation. Manifest variables were: ‘acculturative stress’ (20-item scale for somatic and psychological items), 9-item ‘self concept’ scale, 13-item ‘alienation’ scale



What is the association of Significant decline in mean SEP, comparing Turkey to status loss, mobility and Canada (pB0.001) underemployment on common mental disorders in Turkish migrants to Canada?

SEM to assess latent variables ‘employmentrelated experiences’ (manifest variables: duration of unemployment, status mobility, status loss, employment status) and its associations with latent variable ‘poor psychological health’. Coefficient for association of employment-related experiences and poor psychological health was 0.60, p B0.01. Fit of model: GFI: 0.93, RMSEA: 0.06; CFI: 0.95

Ethnicity & Health



35


Measure for social Author (year) mobility Inclan (1983) Hollingshead social status scale to determine SEP.



SCL-90 (revised) scale. What is the association of intergenerational mobility with common mental disorders in secondgeneration Puerto Rican women living in America?

Blishen scale to assess occupation rank.

Intergenerational mobility: Socioeconomic position (grouped as either ‘low’ SEP and ‘middle’ SEP) between two time points (i.e., birth & adulthood)

Lowlow SEP group: 49 (34%); Low-middle SEP: 61 (42%); middlemiddle SEP: 32 (23%) No women moved from middle to low SEP

Current SEP compared to (reported) parental SEP assessed.

Eaton and Lasry (1978)


‘Mental health’ determined by the Langner scale

What is the association of social mobility with common mental disorders in employed migrants from North Africa to Canada?

Variance for the mobility score was: 94.9 (range: 37.1 to 25.2) suggesting extreme upward downward mobility within the sample

Women who were ‘low’ SEP at both time points had higher mean psychological scores than women who remained middle SEP at both time points or women who were upwardly mobile (p B0.05) No adjustment for confounders

1. Association of (upward) occupational mobility with mental health score, after adjusting for present job and education was r 0.12 p ns; after adjusting for present job only was r0.20 p B0.01






Proportion in sample experiencing social mobility and direction of mobility (up/ Findings, confounders down/stable) and potential interactions 2. Partial correlation of (upward) occupational mobility with mental health score in those employed 2 years was r 0.30 p 0.02; in those employed B1 year: r 0.55 p 0.01.

Asylum seekers and refugees GHQ-12 Lindencrona Respondents asked et al. (2007) ‘Have you experienced lost respect and status?’ assessed on 5-point scale

Is lost respect and status Mean (SD): Lost roles in associated with CMD in a society: 2.3 (1.6) group of refugees from the Middle East to Sweden?

Lost respect and status: 0.54 (0.95) Lost roles at home: 1.4 (1.6)

‘Lost respect and status’ loaded on to an underlying ‘discrimination and status loss’ construct. Zero-order correlations between discrimination and status loss and GHQ scores was 0.29, p B0.01

Ethnicity & Health



37


Measure for social Author (year) mobility Lerner et al. (2005)

10-point hierarchical scale of occupation. Difference in employment status comparing pre- to post-migration.

Instrument to assess common mental disorders PERI-D: measure for common mental disorders in populations



What is the association of Percentage of downwardly mobile: males: 42.9%; ‘change in employmentfemales 58.7% status’ as a result of migration with common mental disorders?

Downward mobility not associated with a decrease in PERI-D scores.

No adjustment for confounders. Analyses stratified by gender Laban et al. (2005)

CIDI ‘Work below level’ method of determining this not given in paper

What is the association of ‘working below level’ with common mental disorders in asylum seekers to the Netherlands?

‘Work below level’ in 22.7% of asylum seekers who had migrated in last year; and 44.9% of asylum seekers who had lived in the Netherlands for at least 2 years

1. On all mental disorder outcomes percentage who were ‘working below level’ higher than those who did not have a disorder 2. Association of ‘work below level’ and ‘one or more psychiatric disorders exc. PTSD’: OR 1.37 (95% CI: 1.11, 1.69)* *Adjusted for gender, family issues, uncertainty over asylum status



Measure for social Author (year) mobility Beiser et al. (1993)

‘Status discrepancy’ determined by using Blishen scale to assess occupation, taking into account previous education (both migrants and refugees) or prior job (refugees only).

Instrument to assess common mental disorders Symptom inventory from CES-D, Senegal Health Scales, SRQ, DIS. Items from Vietnamese Depression Scale.



What is the association of underemployment with depression in refugees to Canada and in nonmigrant Canadians?

Full sample of refugees were downwardly mobile. Greatest downward mobility was experienced by the most educated refugees.

1. No association of under-employment with depression in refugees; regression coefficient, r 0.03, pnon significant at first time point, r 0.00 at second time point 2. In Canadians: correlation between status discrepancy scores and depression was r 0.32 (p B0.001)

Ethnicity & Health



39


ratio (95% CI)

origin

destination

Tinghog 2007

Scandinavian regions

Sweden

1.18 (0.06, 24.57)

Tinghog 2010

Finland

Sweden

0.45 (0.03, 8.16)

Tinghog 2007

Europe

Sweden

1.82 (0.41, 8.09)

Tinghog 2007

Outside of Europe & Scandinavian regions*

Sweden

1.19 (0.46, 3.06)

Burgard 2009

Central & S. America; Asia

USA

1.13 (0.85, 1.51)

author

Labour migrant

1.15 (0.87, 1.50)

Pooled OR (Labour migrant)

Refugee or asylum seeker Tinghog 2010

Iran

Sweden

1.98 (0.90, 4.36)

Tinghog 2010

Iraq

Sweden

1.16 (0.62, 2.18)

Laban 2005

Iraq

Netherlands

3.15 (1.94, 5.13)

Pooled OR (Refugee or asylum seeker)

1.98 (1.06, 3.73)

Pooled OR (Labour migrant & refugee/asylum seeker)

1.56 (1.04, 2.33)

.25 .5 1 1.5 2 2.5 3.5 Odds ratio for association of downward social mobility with common mental disorders; (reference group is 'stable/ upward social mobility')

Figure 3. Association of downward social mobility with common mental disorders in international migrants; random effects meta-analysis.


Odds

Ethnicity & Health

41

Table 4. Meta-regression of study characteristics in predicting common mental disorders. Study characteristics Response rates

60% B60%

Method for determining social mobility

Pre-migration SEP compared to post-migration SEP by investigator Participant self-reported changes to social status following migration Labour migrant Refugee or asylum seeker

Reason for migration


Ratio (95% CI)

Destination

Europe/Scandinavian regions USA

Region of origin

Europe or Scandinavian regions Outside Europe/Scandinavian regions

REF 1.32 (0.44, 3.92) REF 1.32 (0.44, 3.91) REF 1.73 (0.72, 4.16) REF 0.65 (0.21, 1.98) REF 1.24 (0.22, 7.07)

occurrence amongst the migrant groups in the studies. This included a study of depression in Vietnamese refugees to Canada (Beiser et al. 1993), and a study of migrants from the Former Soviet Union to Israel (Lerner et al. 2005). Only one study suggested that being upwardly mobile was associated with greater psychological morbidity (Eaton and Lasry 1978). This was the oldest study (from 1978) and assessed a group of employed men who had migrated from North Africa to Canada.

Studies in second-generation groups Only two studies assessed intergenerational social mobility in second-generation groups (Inclan 1983, VanOort et al. 2007). In both studies upward intergenerational social mobility was associated with improvements in mental health (Inclan 1983, VanOort et al. 2007). The study of second-generation Turkish adults living in the Netherlands was a good quality study (Table 2) which employed a longitudinal design. This study did not find a differential association of social mobility with common mental disorders in the Turkish group compared to Dutch adults, although Turkish adults were more likely to experience upward intergenerational social mobility across their life-course than Dutch adults (VanOort et al. 2007).

Discussion Main findings In keeping with the broader literature (Abadan-Unat 1995, Jong and Madamba 2001, Heath and MacMahon 2005, Platt 2005, Slack and Jensen 2007), migrants tend to experience downward mobility and underemployment. Our review revealed a paucity of research examining this as a possible factor in understanding the mental health and well-being of migrants, with a more striking absence of research examining this linkage among second-generation groups, for whom employment


42 J. Das-Munshi et al. and social mobility may be different to the majority population (Heath and MacMahon 2005, Platt 2005). Meta-analysis suggested that international migrants who experienced downward social mobility as a result of migration were more likely to screen positive for common mental disorders than migrants who retained a stable socio-economic position or who were upwardly mobile (Crude OR: 1.56; 95% CI: 1.04, 2.33). This was despite the inclusion in this estimate of groups migrating from within Europe which would have included skilled migrants moving to relatively well-paid jobs. Meta-regression suggested a larger association in refugee/asylum seeker groups than in labour migrants, although the strength of the evidence for a difference was weak. Conclusions were limited by the small number of studies retrieved. There were conflicting results from three studies which did not support an association of downward mobility or underemployment with psychological morbidity (Eaton and Lasry 1978, Beiser et al. 1993, Lerner et al. 2005). The widely divergent international migratory contexts, as well as the design and quality of the studies, accounted for the observed heterogeneity. In addition, the variety of ways in which social mobility has come to be operationalised within the literature added a further tier of complexity in attempting to synthesise results.

Relationship with previous research Although this review focused on the association of social mobility with mental disorders in migrant groups, there are other factors associated with social mobility which might also have independent associations with mental health. For example, discrimination is associated with poorer health in migrant and ethnic minority groups (Davey Smith et al. 2000, Nazroo 2003). In the present review, several studies assessed discrimination alongside social mobility (Laban et al. 2005, Lindencrona et al. 2007, Tingho¨g et al. 2010). ‘Status loss’ in one study loaded on to the same underlying construct as ‘discrimination’ (Lindencrona et al. 2007), and in other studies, were associated with mental disorders(Laban et al. 2005, Tingho¨g et al. 2010). Therefore attempting to examine migrant social mobility out of the context of discriminatory practices may present only part of the picture (Nazroo 2003). The paucity of studies retrieved meant that we were unable to assess this further. Future research might consider the interaction of discrimination with social mobility in patterning adverse mental health in migrant groups, as the two processes may not be mutually exclusive (Nazroo 2003). Many of the studies did not specify how they assessed migrant or secondgeneration status, although where this was mentioned, asking respondents their ‘country of birth’ tended to be the preferred approach. Migration is one facet in the construction of ethnic identities. Whereas this might be partially instructive in understanding the role of social mobility across the life-course for first-generation migrants, for second-generation groups ethnic identity as a social construct is considerably more complex and contentious (Modood 1996, Modood et al. 1997). The paucity of research examining intergenerational social mobility in secondgeneration groups with mental health limited how far this could be assessed here, although this should be examined in closer detail in future research.

Ethnicity & Health

43


Selection or causation? In a previous review, lower socio-economic position was associated with an increased risk of depression (Lorant et al. 2003), while a review of literature from developing countries showed an association with lower education (Patel and Kleinman 2003). Whether common mental disorders are a function of adversity and material disadvantage (Dohrenwend et al. 1992) or conversely, cause people to ‘drift’ into a lower socio-economic position remains contested (Muntaner et al. 2004). This poses challenges in attempting to understand migration-related mobility processes and associations with mental health. A limitation of all of the studies reviewed was that the mental health of migrants prior to migration was not assessed. The finding that asylum seeker/refugee groups had a larger risk of common mental disorders (compared to labour migrant groups in the meta-analysis) might be accounted for by the pre-migration trauma which these groups were more likely to have experienced. Conversely, evidence suggests that people who have the economic facilities and capital to migrate will tend to be selectively healthier and therefore are likely to have a reduced risk of morbidity (Marmot et al. 1984). Given that much of the evidence appears to support a social causation model for depression (Dohrenwend et al. 1992, Muntaner 1999, Lorant et al. 2003), we would contend that the likely direction of association is consistent with a social causation model.

Psychosocial versus neo-material explanations for common mental disorders The variety of approaches used to assess social mobility and underemployment highlights the challenges in attempting to operationalise this in migrant groups, who may experience absolute improvements in income yet also experience a relative decline in social status, or work in occupations for which they are over-qualified. This experience of social mobility in migrant groups is a paradox which bears a relationship to wider debates in health inequalities. On the one hand, adversity related to material deprivation has been suggested to exert a direct effect on health, through ‘the differential accumulation of exposures and experiences that have their sources in the material world’ (Lynch et al. 2000). It has also been suggested that the relative effects of perceived lower status are fundamental causes of poorer health (Marmot and Wilkinson 2001). It may be that both are important, particularly if ‘psychosocial’ pathways tap into the status loss and discrimination frequently experienced by migrant groups (Nazroo 2003). Both material and psychosocial pathways (alongside the direct effects of racial discrimination) may operate simultaneously to account for adverse health outcomes in ethnic minority and migrant groups (Nazroo 2003).

Limitations It was not possible to assess gender differences across studies, as frequently studies did not present results which allowed analysis by gender. The reasons for migration and patterns of work taken may have marked differences for men and women; the ‘feminisation of migration’ has been noted as a more recent shift (Castles 2000). The effects on downstream adult common mental disorders also vary by gender and lifecourse social mobility (Tiffin et al. 2005). This should be investigated in future work.


44 J. Das-Munshi et al. Despite our best efforts we were unable to locate literature examining migration from all geographical regions. For example, there was a conspicuous absence of studies examining migration within Africa or from South Asia. We located only one study examining migration from the Former Soviet Union and no studies examining migration from other parts of Eastern Europe. This might impact on the generalisability of the findings. More research is urgently needed to address this gap in the literature. The paucity of studies retrieved meant that it was not possible to assess for publication biases using funnel plots. A concern is that smaller studies showing positive associations may have been more likely to be published and therefore included in this review. We attempted to address this by locating ‘grey literature’; however, contact with known experts in the field and first authors did not reveal additional papers. A final issue relates to the assessment of ‘common mental disorders’ across cultures. An assumption underpinning our review is that this was assessed in a consistent manner across all studies. We minimised problems with low reliability by including only studies which employed structured assessments. However, it may be that the heterogeneity observed across some of the studies was due to the way in which ‘common mental disorders’ was conceptualised in each of the studies.

Implications for future policy and research Given the current scale and likely expansion of global migration, the lack of research exposed by this review suggests an important area of future enquiry. The findings suggest that future investigation should contextualise the socio-economic position of migrants and second-generation groups within the life-course. There may be premigratory time points in this life-course, and disadvantage may continue to be felt into the next generation. There may be a ‘penalty’ associated with international migration for work which could be associated with independent mental health consequences for first-generation migrants and their children. It has been suggested that simply ‘adjusting’ for socio-economic position in analyses of ethnic minority groups and their health to ‘reveal’ underlying associations may obscure the importance of socio-economic deprivation as a fundamental explanation for health in ethnic minority groups (Nazroo and Williams 2006). Further, for ethnic minority groups who have experienced migration, this review suggests that ‘snapshot’ measures of socio-economic position may fail to take into account complex migration-related changes to social class and status and associations with mental health. In some countries ‘managed migration’ utilising a points-based system to favour skilled migrants is the norm. Such programmes may build in selection for ‘healthier’ migrants into the screening process. In many of the countries reviewed here this has been the practice for many years, but paradoxically, has not necessarily led to the inclusion of migrant workers in the workforce at occupational parity. Although limited by the small number of studies retrieved, there was some evidence to suggest that this may have detrimental consequences for mental health. Understanding the implications of such policy initiatives should be an important public health priority.

Ethnicity & Health

45

Key messages What is already known: (1) Downward occupational social mobility may be a frequent occurrence in people who migrate internationally. (2) Few studies have examined the association of social mobility with common mental disorders in migrant and second-generation groups.


What this study adds: (1) Downward social mobility was a frequent consequence of international migration in reviewed studies. (2) Using random effects meta-analysis, migrants who were downwardly mobile were also more likely to have common mental disorders (Pooled OR: 1.56 95% CI: 1.04, 2.33). Very few studies included second-generation groups. The paucity and quality of studies retrieved constrained further interpretation. (3) Given the scale of global migration, the lack of research exposed by this review suggests an important area of future enquiry. Acknowledgements We are grateful to all of the investigators who took the time to respond to queries. In particular we are grateful to Dr P. Tingho¨g, Dr K. Laban, Dr R. Pernice, Professor F. Van Oort, Dr S. Burgard, Dr C. Abdou and Dr E.J. Nicklett, Professor M. Abbott and Professor M. Beiser for additional comments, references and in some cases, data. In addition we are grateful to Girmay Medhin and Professor Michael E. Dewey for statistical support. JD is funded by a Medical Research Council fellowship. This funding body played no part in the review.

Notes 1. Changes to socioeconomic position across generations has been termed intergenerational social mobility. 2. Two unpublished studies were also identified through searches; however, the results of these were not available at the time of compiling this review. This included one study from the USA (CM Abdou, personal communication) and one study from New Zealand (A. SobrunMaharaj, personal communication).

References Abadan-Unat, N., 1995. Turkish migration to Europe.R. Cohen. The Cambridge Survey of World Migration (p. 279). Cambridge: Press Syndicate of The University of Cambridge, 279. Abbott, M.W., et al., 1999. Chinese migrants’ mental health and adjustment to life in New Zealand. Australian and New Zealand Journal of Psychiatry, 33 (1), 1321. Aycan, Z. and Berry, J.W., 1996. Impact of employment-related experiences on immigrants’ psychological well-being and adaptation to Canada. Canadian Journal of Behavioural Science, 28 (3), 240251. Beiser, M., et al., 2002. Poverty, family process, and the mental health of immigrant children in Canada. Americal Journal of Public Health, 92 (2), 220227. Beiser, M., Johnson, P.J., and Turner, R.J., 1993. Unemployment, underemployment and depressive affect among Southeast Asian refugees. Psychological Medicine, 23 (3), 731743.


46 J. Das-Munshi et al. Black, C., 2008. Working for a healthier tomorrow: Dame Carol Black’s Review of the health of Britain’s working age population. London: TSO. Blishen, B.R. and Roberts, H.A.M., 1976. A revised socioeconomic index for occupation in Canada. Canadian Review of Sociology/Revue canadienne de sociologie, 13 (1), 7179. Bradburn, M.J., Deeks, J.J., and Altman, D.G., 1998. sbe24: Metan: an alternative metaanalysis command. Stata Technical Bulletin., 44, 415. Cantor-Graae, E. and Selten, J.-P., 2005. Schizophrenia and migration: a meta-analysis and review. American Journal of Psychiatry, 162 (1), 1224. Castles, S., 2000. International migration at the beginning of the twenty-first century: global trends and issues. International Social Science Journal, 52 (165), 269281. Davey Smith, G., et al., 2000. Ethnic inequalities in health: a review of UK epidemiological evidence. Critical Public Health, 10 (4), 375408. DerSimonian, R. and Laird, N., 1986. Meta-analysis in clinical trials. Controlled Clinical Trials, 7 (3), 177188. Dohrenwend, B., et al., 1992. Socioeconomic status and psychiatric disorders: the causationselection issue. Science, 255 (5047), 946952. Eaton, W.W. and Lasry, J.-C., 1978. Mental health and occupational mobility in a group of immigrants. Social Science & Medicine, 12, 5358. Egger, M., Schneider, M., and Smith, G.D., 1998. Meta-analysis spurious precision? Metaanalysis of observational studies. British Medical Journal, 316 (7125), 140144. Galobardes, B., Lynch, J., and Smith, G.D., 2007. Measuring socioeconomic position in health research. British Medical Bulletin, 8182 (1), 2137. Harding, S., 2003. Social mobility and self-reported limiting long-term illness among West Indian and South Asian migrants living in England and Wales. Social Science & Medicine, 56 (2), 355361. Harding, S. and Balarajan, R., 2001. Longitudinal study of socio-economic differences in mortality among South Asian and West Indian migrants. Ethnicity & Health, 6 (2), 121 128. Heath, A. and MacMahon, D., 2005. Social mobility of ethnic minorities. In: G.C. Loury, T. Modood, and S.M. Teles, eds. Ethnicity, social mobility and public policy. Cambridge: Cambridge University Press, 393413. Held, D., et al., 1999. Global transformations: Politics, economics and culture. Cambridge: Polity. Higgins, J.P.T. and Green, S., eds., 2011. Cochrane handbook for systematic reviews of interventions version 5.1.0. Available from: www.cochrane-handbook.org, The Cochrane Collaboration. Higgins, J.P.T., et al., 2003. Measuring inconsistency in meta-analyses. British Medical Journal, 327 (7414), 557560. Inclan, J., 1983. Psychological symptomatology in second generation puertorican women of three socioeconomic groups. Journal of Community Psychology, 11 (4), 334345. Jong, G.F.D. and Madamba, A.B., 2001. A double disadvantage? Minority group, immigrant status, and underemployment in the United States. Social Science Quarterly, 82 (1), 117 130. King, M., et al., 1994. Incidence of psychotic illness in London: comparison of ethnic groups. British Medical Journal, 309 (6962), 11151119. Kogan, I. and Cohen, Y., 2008. Educational selectivity and labour market attainment of Jewish immigrants from the Former Soviet Union in Israel and Germany in the 1990s. In: H. Kolb and H. Egbert, eds. Migrants and markets: perspectives from economic and the other social sciences. Amsterdam: Amsterdam University Press, 104119. Krieger, N., Williams, D.R., and Moss, N.E., 1997. Measuring social class in US public health research: concepts, methodologies, and guidelines. Annual Review of Public Health, 18 (1), 341378. Laban, C.J., et al., 2005. Postmigration living problems and common psychiatric disorders in Iraqi asylum seekers in the Netherlands. The Journal of Nervous and Mental Disease, 193 (12), 825832. Leao, T.S., et al., 2005. Incidence of mental disorders in second-generation immigrants in Sweden: a four-year cohort study. Ethnicity & Health, 10 (3), 243256.


Ethnicity & Health

47

Lerner, Y., Kertes, J., and Zilber, N., 2005. Immigrants from the former Soviet Union, 5 years post-immigration to Israel: adaptation and risk factors for psychological distress. Psychological Medicine, 35 (12), 18051814. Lindencrona, F., Ekblad, S., and Hauff, E., 2007. Mental health of recently resettled refugees from the Middle East in Sweden: the impact of pre-resettlement trauma, resettlement stress and capacity to handle stress. Social Psychiatry and Psychiatric Epidemiology, 43 (2), 121 131. Lindert, J., et al., 2009. Depression and anxiety in labor migrants and refugees: a systematic review and meta-analysis. Social Science & Medicine, 69 (2), 246257. Lorant, V., et al., 2003. Socioeconomic inequalities in depression: a meta-Analysis. American Journal of Epidemiology, 157 (2), 98112. Loury, G.C., Modood, T., and Teles, S.M., 2005. Introduction. In: T.M.G.C. Loury and S.M. Teles, eds. Ethnicity, social mobility and public policy: comparing the US and UK. Cambridge: Cambridge University Press, 117. Lynch, J.W., et al., 2000. Income inequality and mortality: importance to health of individual income, psychosocial environment, or material conditions. British Medical Journal, 320 (7243), 12001204. Marmot, M., Adelstein, A.M., and Bulusu, L., 1984. Lessons from the study of immigrant mortality. The Lancet, 323 (8392), 14551457. Marmot, M. and Wilkinson, R., 2001. Psychosocial and material pathways in the relation between income and health: a response to Lynch et al. British Medical Journal, 322 (7296), 12331236. Modood, T., 1996. If races do not exist, then what does? Racial categorisation and ethnic realities. The racism Problematic: contemporary sociological debates on race and ethnicity. Lewiston: The Edwin Mellen Press, 89105. Modood, T., et al., 1997. Ethnic minorities in Britain: diversity and disadvantage. London: Policy Studies Institute. Muntaner, C., 1999. Invited commentary: social mechanisms, race, and social epidemiology: seeking causal explanations in social epidemiology. American Journal of Epidemiology, 150 (2), 121126. Muntaner, C., et al., 2004. Socioeconomic position and major mental disorders. Epidemiological Review, 26 (1), 5362. Nazroo, J., 2003. The structuring of ethnic inequalities in health: economic position, racial discrimination, and racism. American Journal of Public Health, 93 (2), 277284. Nazroo, J. and Williams, D.R., 2006. The social determination of ethnic/racial inequalities in health. In: M. Marmot and R.G. Wilkinson, eds. Social determinants of health (p. 238). Oxford: Oxford University Press, 238. Nicklett, E.J. and Burgard, S.A., 2009. Downward social mobility and major depressive episodes among Latino and Asian-American immigrants to the United States. American Journal of Epidemiology, 170 (6), 793801. Patel, V. and Kleinman, A., 2003. Poverty and common mental disorders in developing countries. Bulletin of the World Health Organization, 81, 609615. Platt, L., 2005. Migration and social mobility: the life chances of Britain’s ethnic minority communities. Bristol: Joseph Rowntree Foundation. Porter, M. and Haslam, N., 2005. Predisplacement and postdisplacement factors associated with mental health of refugees and internally displaced persons: a meta-analysis. Journal of American Medical Association, 294 (5), 602612. Razum, O., 2006. Commentary: of salmon and time travellers-musing on the mystery of migrant mortality. International Journal of Epidemiology, 35 (4), 919921. Slack, T. and Jensen, L., 2007. Underemployment across immigrant generations. Social Science Research, 36 (4), 14151430. Smith, N.R., Kelly, Y.J., and Nazroo, J.Y., 2009. Intergenerational continuities of ethnic inequalities in general health in England. Journal of Epidemiology & Community Health, 63 (3), 253258. StataCorp 2007. Stata statistical software: release 10. College Station, TX: StataCorp L.P. Stroup, D.F., et al., 2000. Meta-analysis of observational studies in epidemiology. JAMA: The Journal of the American Medical Association, 283 (15), 20082012.


48 J. Das-Munshi et al. Tiffin, P.A., Pearce, M.S., and Parker, L., 2005. Social mobility over the lifecourse and self reported mental health at age 50: prospective cohort study. Journal of Epidemiology and Community Health, 59 (10), 870872. Tingho¨g, P., Hemmingsson, T., and Lundberg, I., 2007. To what extent may the association between immigrant status and mental illness be explained by socioeconomic factors? Social Psychiatry and Psychiatric Epidemiology, 42 (12), 990996. Tingho¨g, P., et al., 2010. The association of immigrant- and non-immigrant-specific factors with mental ill health among immigrants in Sweden. International Journal of Social Psychiatry, 56 (1), 7493. UN HCR 2011. The UN Refugee Agency [online]. http://www.unhcr.org/pages/49c3646c1d. html [Accessed 31 October 2011]. United Nations: Department of Economic and Social Affairs: Population Division 2009. International Migration Report 2006: A global assessment. UN: New York. VanOort, F.V.A., et al., 2007. Development of ethnic disparities in internalizing and externalizing problems from adolescence into young adulthood. Journal of Child Psychology and Psychiatry, 48 (2), 176184. Wahlbeck, K. and Taipale, V., 2006. Europe’s mental health strategy. British Medical Journal, 333 (7561), 210211. World Health Organization 2008. Policies and practices for mental health in Europe- meeting the challenges. Available from: http://www.euro.who.int/__data/assets/pdf_file/0006/96450/ E91732.pdf [Accessed 31 October 2011].

Supplementary Material: MOOSE Checklist

Reporting of background should include Problem definition

Hypothesis statement

Description of study outcomes

Type of exposure or intervention used Type of study designs used Study population Reporting of search strategy should include Qualifications of searchers

Brief description of how the criteria were handled in the meta-analysis The literature suggests that downward social mobility is a frequent consequence of international migration. We sought to establish if an association existed between downward mobility and common mental disorders in migrant groups relative to migrants who had experienced no changes to social mobility or who experienced upward social mobility We hypothesised that international migrants experiencing downward social mobility would be more likely to screen positive for common mental disorders (relative to those experiencing stable/ upward social mobility). The main outcome/ dependent variable in meta-analysis and narrative synthesis was the presence of ‘common mental disorders’. This term refers to neurotic disorders commonly encountered at the level of primary care, associated with disability and impaired functioning. The use of screening tools or diagnostic assessments result in either ‘counts’ of symptoms whereby a higher score indicates greater distress and disability, or utilise cutpoints usually validated against suitable ‘gold standards’. For the purposes of this review we retained studies employing both approaches. The main exposure was downward social mobility versus stable/ upward social mobility in migrant groups. Searches for case-control, cross-sectional or cohort study designs were undertaken International migrants from any geographical region. The first author is a qualified Psychiatrist and has an academic background in Epidemiology and Sociology.

Ethnicity & Health


Criteria

49


Criteria Search strategy, including time period included in the synthesis and keywords

Brief description of how the criteria were handled in the meta-analysis

Keywords and mesh headings used for searches were: ‘‘mental disorders’’; ‘‘mental health’’, ‘‘depress*’’, ‘‘anxiety’’, ‘‘psych*’’; Intrageneration* mobility, Intergeneration* mobility, underemployment, *mobility, Migra*, immigra*, emigra*, refugee*, asylum*. Where it was unclear if the paper met inclusion criteria the paper was retrieved and examined. If results from a study had been published in more than one source then all papers were retrieved and presented as findings from one study. Databases and registries searched Systematic searches were conducted in the following databases: Medline, Ovid, Psychinfo, Embase, Social Policy and Practice, British Nursing Index and Archive, by the first author. Searches were conducted from November 2009-February 2010. Due to limited resources only English-language papers and dissertations were considered. Search software used, name and version, including special features Software relevant to each of the databases was used. References were stored in Endnote X2. Use of hand searching Bibliography sections of papers as well as one review(Porter & Haslam, 2005) were hand-searched for additional references. Searches were supplemented by contacting first authors of each included study, and experts in the field, to ensure that unpublished work could also be reviewed. List of citations located and those excluded, including justifications A ‘reject log’ of excluded citations is available from the first author upon request. The flow chart in Figure 2 outlines the reasons for excluding studies. Method of addressing articles published in languages other than English Due to limited resources we were unable to search for non-English studies. This is listed as a limitation in the discussion section of the review. Method of handling abstracts and unpublished studies These were retrieved and assessed for inclusion by contacting main authors. Description of any contact with authors All first authors were contacted as well as a list of further experts in the field were contacted.




Reporting of methods should include Description of relevance or appropriateness of studies assembled for assessing the hypothesis to be tested

Rationale for the selection and coding of data

Assessment of confounding

Assessment of study quality, including blinding of quality assessors; stratification or regression on possible predictors of study results

Assessment of heterogeneity Description of statistical methods in sufficient detail to be replicated


51

Very few cohort studies were identified, and most studies came from a restricted geographical area. These limitations are highlighted. The results of meta-analysis should be interpreted with caution given the concerns around potential unmeasured confounding. In order to perform meta-analysis we derived the odds of common mental disorders in downwardly mobile migrants relative to migrant experiencing stable or upward social mobility. Each study was assessed for assessment of confounders and potential interactions in associations. Very few studies considered this in the association of social mobility with common mental disorders. Meta-regression was used to assess if stratification by specific study design features impacted on the overall association between social mobility and common mental disorders. In addition the studies were assessed according to a priori quality criteria shown in Figure 1. This was assessed using I2 tests for heterogeneity and further assessed using Meta-regression. Analyses were performed in STATA/IC 10.1 (StataCorp, 2007). Metaanalysis was used to assess the association of downward social mobility (relative to upward/ stable social mobility) in migrants with common mental disorders. The dependent variable in meta-analysis was common mental disorders as assessed through instruments employing validated cut-points. Using this approach we derived the overall pooled OR from all of the studies on migrant groups, and then investigated this according to the subgroups/ criteria given in Figure 1. Pooled estimates for the odds of downward social mobility and common mental disorders (relative to migrants who experienced stable/ upward social mobility) were estimated using DerSimonian and Laird’s method (1986).

Ethnicity & Health


Criteria


Criteria


A priori, we assumed that a random effects meta-analysis would best capture the variability between studies, (eg. due to differing migratory contexts, geographical regions and study designs). Provision of appropriate tables and graphics A flowchart showing the process by which final studies were obtained is shown (Figure 2). A figure showing the results of meta-analysis both pooled and by migration status is also shown (Figure 3). Reporting of results should include Graph summarizing individual study estimates and overall estimate Table giving descriptive information for each study included

Figure 3 displays this. Table 1 gives overall study characteristics, whereas table 2 and 3 show characteristics of individual studies.

Results of sensitivity testing

Indication of statistical uncertainty of findings Reporting of discussion should include Quantitative assessment of bias

Justification for exclusion Assessment of quality of included studies

We assessed associations by sub-groups which were determined a priori (Fig. 1). ‘Reason for migration’ (labour migrant versus refugee or asylum seeker) appeared to account for the largest proportion of the variance. This finding informed the way in which we chose to display the final metaanalysis figure (Fig. 3). This is indicated by the confidence intervals for estimates derived through meta-analysis and meta-regression. This was assessed using meta-regression, however issues relating to sample size and potential response bias did not appear to impact on the overall estimate of association. This was documented in the ‘reject log’ which is available from the main author, and the overall reasons for excluding studies are shown in Figure 2. This was assessed qualitatively and through using meta-regression where possible




Reporting of conclusions should include Consideration of alternative explanations for observed results

Generalization of the conclusions Guidelines for future research

Disclosure of funding source


We discuss the possibility that asylum seekers and refugee groups were more likely to have experienced trauma which may account for the larger association of downward social mobility with common mental disorders in this group. We suggest that we were unable to separately assess the role of gender in possibly also accounting for findings as this data was not readily available. We also consider the role of publication biases in accounting for the findings, although we did make attempts to identify ‘grey’ literature (as well as all first authors and experts in the field) to minimise this. We have noted that the review was limited by a lack of studies from Africa and South Asia which may affect the generalisability of the findings. We suggest that future research should consider the role of gender in accounting for the observed associations. We also suggest that ‘snapshot’ views of socioeconomic position in migrant groups may mask the role of lifecourse socioeconomic position (which may have a distinct pattern in international migrants) in accounting for mental health differences. We suggest that the lack of research particularly concerning second generation groups is a concern and should be addressed in future research. The first author was funded by the MRC. The finding body did not play any part in the design of the protocol, analysis of results or preparation of the manuscript.

Ethnicity & Health


Criteria

53