Introduction
As defined by the World Health Organization (WHO), a migrant is any person moving from one area to another for varying periods of time [1]. According to the United Nations Department of Economic and Social Affairs, the number of international migrants increased from 153 to 272 million during the period from 1990 to 2019 [2]. Nowadays, migrants represent 3.5% of the world’s population, and it is estimated that 10.6% of them are refugees [2]. Moreover, hundreds of millions of people move within a country [Reference Orcutt, Spiegel, Kumar, Abubakar, Clark and Hortan3]. Migration should be considered as a complex process along which migrants may be exposed to several stressors, including stressful premigratory experiences, a traumatic act of migration, and postmigratory difficulties. The complex interplay between these factors, along with biological and psychological determinants, can lead to poor mental health [Reference Bhugra and Becker4]. For instance, in order to explain the higher prevalence of psychotic and schizophrenia spectrum disorders among migrants and ethnic minorities compared to native-born individuals, Morgan et al. [Reference Morgan, Charalambides, Hutchinson and Murray5,Reference Morgan, Knowles and Hutchinson6] have proposed the existence of a socio-developmental pathway to psychosis. According to this model, early exposure to adversities which may occur prior, during, or after migration may interact with biological predisposition (e.g., genetic risk) and affect the neurobiological development creating an enduring proneness to psychosis. The onset of psychosis may then be detonated by the cumulative effect of further stressors, such as psychosocial adversities or substance abuse. Recent evidence from the field of neuroimaging supports the hypothesis that adverse social factors may increase the risk of psychosis in migrants via dysregulation of the dopamine neurotransmission system [Reference Egerton, Howes, Houle, McKenzie, Valmaggia and Bagby7]. Acquisition of good language proficiency (LP) in the host country language is a key factor for the postmigration adaptation process and seems to be crucial for both psychosocial and economic wellness [Reference Isphording and Otten8]. Specifically, according to a recent meta-analysis about mental health among minorities in the United States, LP was found to be negatively related to depression, anxiety, psychological distress, and negative affect [Reference Yoon, Chang, Kim, Clawson, Cleary and Hansen9].
LP relies on the ability of the interpretation of the linguistic code (words, sentences, etc.) as well as of the contextual knowledge of background and metalinguistic information—known as pragmatics, helping the recipient to understand the specific meaning of the speaker/sender [Reference Spence10]. Likewise, according to the socio-cognitive approach to second language acquisition, LP is more than the simple product of new knowledge acquisition [Reference Atkinson11]. Second language acquisition requires the interplay of cognitive, behavioral, and social skills, and occurs “in, for, and by virtue of integrated mind–body–world ecologies” [Reference Atkinson, Churchill, Nishino and Okada12]. Moreover, it has long been known that self-confidence, motivation, empathic skills, and several other affective factors may influence the second language acquisition process and act as an affective filter [Reference Schumann13].
Many studies have explored the social determinants of LP, including, among others, age at migration, educational attainment, host language exposure, reason for migration, and social interactions [Reference Chiswick and Miller14–Reference Atkinson16].
In terms of mental disorders, LP may be considered a postmigration mediating factor [Reference Bhugra and Becker4,Reference Morgan, Charalambides, Hutchinson and Murray5]. In fact, poor LP may prevent interpersonal communication and transmission of emotional states, leading to self-isolation and alienation. An explanation could be that the neural processes underlying second language acquisition are involved in social cognition processes and, above all, theory of mind (ToM) [Reference Atkinson16].
Interestingly, severe psychiatric illnesses are often related to deficits in one or more aspects of social cognition. Specifically, emotion processing, ToM, and social relationship perception seem to be altered in psychotic spectrum disorders, major depressive disorder, and bipolar disorder [Reference Hoertnagl and Hofer17],
In this light, migrants with a personal or family history of mental disorders may present impairment in social cognition and subsequent difficulties in tasks such as language learning. Shortcomings in LP could then be considered as the expression of a premigratory vulnerability to psychosis and other psychiatric disorders.
We aimed to review the studies analyzing the relationship between LP and (a) psychiatric disorder prevalence and/or severity among migrants and (b) psychiatric symptoms prevalence and/or severity among migrants. Secondarily, we aimed to consider whether sociodemographic and phase-specific migration factors may influence the relationship between LP and mental disorders.
Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [Reference Moher, Liberati, Tetzlaff and Altman18,Reference Liberati, Altman, Tetzlaff, Mulrow, Gotzsche and Aloannidis19] and the recommendations of the Cochrane Collaboration [Reference Higgins, Thomas, Chandler, Cumpston, Li and Page20] were followed for the systematic literature search.
Study types
Eligible articles included all English language papers published in peer-reviewed journals from January 1990 to June 2021, reporting primary studies and data on mental disorders in first- or further-generation immigrants and investigating proficiency in the host language. References from each study were screened and reviewed. We limited the search to studies published after 1990 to ensure that all included studies rely on the same definition of “migrants,” as first defined by the United Nations General Assembly [21]. The last literature search was conducted on June 16, 2021.
Search methods
MEDLINE, PsycArticles, EMBASE, and PsycInfo were systematically searched in April 2020 using controlled vocabulary and keywords to identify original studies reporting mental disorder prevalence among migrants and taking into account LP. Complete search algorithms are reported in Table S1 (see Supplementary Materials).
Population and selection of studies
After performing the initial search, duplicates were identified and discarded. Titles and abstracts were screened and full texts were checked for eligibility in order to find potentially relevant reports. Studies were excluded if the full texts were unavailable even after contacting the corresponding authors. Three of the authors (G.D., C.M., and F.C.) performed the search and the initial data extraction independently, removing duplicates and all irrelevant articles after reading the specific abstract. The full texts of the remaining studies were independently assessed for eligibility by all authors. A consensus conference with the entire team took place in case of ambiguity.
In order to be included, the screened studies had to:
(1) Examine a population of first- or further-generation adult migrants (18 years or older) as defined by the WHO [1]. Further definitions (e.g., on the basis of reasons for migration, legal and economic status) are beyond the scope of this paper. People born in the same country of residence from parents of different origin have been considered as further-generation immigrants and identified based on self-declaration.
(2) Assess mental disorders using validated questionnaires, subscales, semi-structured interviews, or Diagnostical and Statistical Manual of Mental Disorders (DSM)/ International Classification of Diseases (ICD) diagnostic criteria.
(3) Assess migrants’ LP.
(4) Examine the relation between LP and mental disorders.
(5) Have been published after 1990.
Data extraction and analysis
The following information was extracted from the studies and reported in Table 1: (a) authors and year of publication; (b) country in which the study was conducted; (c) population studied; (d) migration history (reason for migration, length of stay, country of origin, destination country, and migrant generation); (e) study design; (f) psychometric measures; (g) language assessment instruments; (h) prevalence and/or severity of mental disorders; and (i) LP and correlation with mental disorders. A narrative synthesis was performed to analyze relevant literature.
Included studies.
Psychometric Measures are extensively reported in Table S2.
Abbreviations: AD, anxiety disorders; ADJ, adjustment disorder; ASH, Acculturation Scale for Hispanics; ASS, Acculturative Stress Scale; BD, bipolar disorder; BDI, Beck Depression Inventory; CCM, collaborative care management; CES-D, Center for Epidemiologic Studies Depression Scale; CI, confidence interval; CIDI, Composite International Diagnostic Interview; CPSS, Cohen’s Perceived Stress Scale; DAM, Depressive Affect Measure Scale; DES, Depression Anxiety and Stress Scale; DSM, Diagnostical and Statistical Manual of Mental Disorders; FEP, first-episode psychosis; FG, first generation; FU, follow-up; GDS-SF, Geriatric Depression Scale Short Form; HR-SMI, high risk of severe mental illness; HTQ, Harvard Trauma Questionnaire; ICD, International Classification of Diseases; ICG, Inventory of Complicated Grief; IES, Impact of Events Scale; ISLPRS, International Second Language Proficiency Rating; KLLS, Korean Language Literacy Scale; LIB, Language, Identity, Behavioral Acculturation Scale; LP, language proficiency; MDD, major depressive disorder; N, number; NLAAS, National Latino and Asian American Study; OR, odds ratio; PCL-C, Post-Traumatic Stress Disorder Checklist-Civilian Version; PD, personality disorders; PGD, prolonged grief disorder; PHQ, Patient Health Questionnaire; PPV, permanente protection visas; PTE, Premigration Traumatic Experiences; PTSD, post-traumatic stress disorder; PV, political violence; RRP, Refugee Resettlment Project; RS, resilience; SAFE, Social, Attitudinal, Familial, and Environmental Acculturative Stress Scale; SCZ, schizophrenia spectrum disorders; SE, standard error; SECV, Survey of Exposure to Community Violence; SES, Socio-Economic Status; SG, second generation; SHS, Senegal Health Scale; SUD, substance use disorder; TPV, temporary protection visas; WIC, The Special Supplemental Nutrition Program for Women, Infants, and Children.
Quality rating
The quality of the studies was gauged considering the seven items of the quality assessment checklist for observational studies (QATSO Score) [Reference Fincher, Thornhill, Murray and Schaller59] adapted to our search. In particular, the checklist evaluates: (a) sample size and source; (b) use of validated tools for psychiatric measures; (c) use of validated tools for LP assessment; (d) reports of the response rate; (e) reports of migration history data (reason for migration, length of stay, country of origin, destination country, and migrant generation); (f) checking for confounding factors (e.g., stratification/matching/restriction/adjustment); and (g) privacy and ethical aspects considered. The QATSO score is reported in Table 1.
Results
Searching of electronic databases initially yielded 1,944 citations, as reported in the PRISMA flowchart (Figure 1). After removing duplicates, 1,283 citations remained. Of these, 1,086 citations were excluded after reading the abstract, as they were reviews, meta-analyses, commentaries, letters to the editor, dissertations, books or book chapters, or non-English language works, or because they did not meet the inclusion criteria. Of the 197 full texts assessed for eligibility, 41 studies were selected for inclusion in the systematic review (Figure 1). The studies included were published between 1993 and 2021 and were conducted in the following geographic areas: 4 in Asia (Korea), 6 in Australia, 9 in Europe (1 in Finland, 3 in Germany, 1 in Italy, 1 in Sweden, 1 in the United Kingdom, 1 in Austria, 1 in Portugal), and 22 in North America (4 in Canada, 18 in the United States). These 41 papers reported the results of 34 cross-sectional analyses and 7 longitudinal analyses (6 prospective and 1 retrospective cohort studies). Sample sizes ranged from 48 to 7,561.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.
Interplay between LP and mental disorders
Lower LP predicting higher prevalence of mental disorders or symptoms severity
Thirty-five of the papers reviewed [22–25, 27–36, 38–40, Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42,Reference Nicholson, Miller, Schwertz and Sorokin44,Reference Ngo, Tran, Gibbons and Oliver45, 47–52, 54–58, 60–63] found a significant association between low LP and mental disorders, with lower LP predicting higher prevalence of the disorder or worse severity of psychiatric symptoms. Of these papers, 13 studies focused only on the relation between depressive symptom severity and LP [Reference Wong, Cheung and Hart22, 24–26, Reference Kim, Yang, Kwon and Kim28,Reference Birman and Tran32,Reference Haasen, Lambert, Mass and Krausz36, 38–40, Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42,Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland47,Reference Schweitzer, Vromans, Brough, Asic-Kobe, Correa-Velez and Murray61]; 4 studies investigated the correlation between LP and depression prevalence [Reference Beiser and Hou35,Reference Wu, Renzaho, Hall, Shi, Ling and Chen51,Reference Furnham and Li55,Reference Beiser and Hou56]; 2 focused on post-traumatic stress disorder (PTSD) symptom severity [Reference Miller and Chandler58,Reference Alegría, Mulvaney-Day, Torres, Polo, Cao and Canino62], 1 on PTSD prevalence [Reference Carl, Frias, Kurtevski, Copo, Mustafa and Font54], 1 on both PTSD and anxiety severity [Reference Halcón, Robertson, Savik, Johnson, Spring and Butcher31], 1 on psychosis prevalence [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50], 1 on somatization prevalence [Reference Fenta, Hyman, Rourke, Moon and Noh33], and 1 on somatic symptom severity [Reference Ngo, Tran, Gibbons and Oliver45]. Of the remaining 10 papers, four on the symptoms severity of more than one condition [Reference Nicholson, Miller, Schwertz and Sorokin44,Reference Hamwi, Lorthe and Barros49,Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52,Reference Kim and Chen63], whereas 7 focused on the prevalence of more than one mental disorder [Reference Finch, Kolody and Vega23,Reference Kim and Kim27,Reference Arévalo, Tucker and Falcón29,Reference Bernstein, Park, Shin, Cho and Park30,Reference Morawa, Dragano, Jöckel, Moebus, Brand and Erim34,Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland48,Reference Kang, Domanski and Moon57]. Finally, one study [Reference Kang, Domanski and Moon57] retrospectively evaluated the charts of 243 migrants who needed to be hospitalized in a Psychiatry Unit between 2004 and 2018. Interestingly, subjects treated for their first-episode of any mental illness were found to be less proficient than those patients with a previous history of mental disorders.
Depression
Low LP was correlated with high prevalence of depression in five studies [Reference Bernstein, Park, Shin, Cho and Park30,Reference Morawa, Dragano, Jöckel, Moebus, Brand and Erim34,Reference Beiser and Hou35,Reference Wu, Renzaho, Hall, Shi, Ling and Chen51]. At the same time, depressive symptom severity assessed through different psychometric measures was found to correlate negatively with LP in 12 studies [Reference Wong, Cheung and Hart22,Reference Kang, Basham and Kim26,Reference Kim, Yang, Kwon and Kim28,Reference Birman and Tran32,Reference Haasen, Lambert, Mass and Krausz36,Reference Leong, Park and Kalibatseva38,Reference Marshall, Schell, Elliott, Berthold and Chun40,Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42,Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52,Reference Furnham and Li55,Reference Schweitzer, Vromans, Brough, Asic-Kobe, Correa-Velez and Murray61,Reference Kim, Kim, Moon, Park and Cho64].
Among Afghan refugees resettled in Australia, participants presenting with more depressive symptoms reported higher levels of communication difficulties and lower LP compared to nondepressed subjects from the same community [Reference Furnham and Li55]. Asian and Korean immigrants with good LP showed a lower prevalence of lifetime psychiatric disorders and depression [Reference Bernstein, Park, Shin, Cho and Park30,Reference Morawa, Dragano, Jöckel, Moebus, Brand and Erim34,Reference Beiser and Hou35]. Similar results were found among women resettled in Portugal from different countries: the prevalence of postpartum depression observed in migrant women was higher than in native speakers, and the risk was higher for limited or intermediate proficiency groups compared to fully proficient migrant women [Reference Beiser and Hou56]. Moreover, the effect of LP was stronger when the analysis was restricted to people with a recent history of migration [Reference Beiser and Hou56]. Among Asians migrating to the United States or Canada, LP correlated with depressive symptom severity assessed by different measures [Reference Wong, Cheung and Hart22,Reference Kim, Yang, Kwon and Kim28,Reference Birman and Tran32,Reference Haasen, Lambert, Mass and Krausz36,Reference Leong, Park and Kalibatseva38,Reference Marshall, Schell, Elliott, Berthold and Chun40,Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42]. In one case, the correlation was significant only among second-generation migrants [Reference Wong, Cheung and Hart22] or among subjects with low ethnic identity [Reference Kim, Yang, Kwon and Kim28]. Moreover, depression among first-generation migrants correlated with higher mother tongue usage (e.g., using only primary language at home) [Reference Wong, Cheung and Hart22]. Among Korean migrants living in the United States, the correlation between English LP and depression severity assessed through the GDS-SF was found to be significant even after adjusting for sociodemographic factors, years in the United States, acculturative stress and self-reported discrimination [Reference Birman and Tran32,Reference Leong, Park and Kalibatseva38,Reference Marshall, Schell, Elliott, Berthold and Chun40]. Also, LP was a significant predictor for improvement of depressive symptom severity over time in asylum seekers living in Germany [Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52]. Among Latinos, higher LP correlated with lower CES-D scores [Reference Miller, Sorokin, Wang, Feetham, Choi and Wilbur24,Reference Schweitzer, Vromans, Brough, Asic-Kobe, Correa-Velez and Murray61], but this correlation was inverted in high ethnic density neighborhoods [Reference Schweitzer, Vromans, Brough, Asic-Kobe, Correa-Velez and Murray61]. At the same time, LP was found to be a positive moderator of the detrimental effect of perceived discrimination on depression [Reference Miller, Sorokin, Wang, Feetham, Choi and Wilbur24]. Also, gender differences impacted on the interplay between LP and mental health: LP did not affect the mental health of Asian women, when they were analyzed separately from men [Reference Bernstein, Park, Shin, Cho and Park30]. On the other hand, in a group of 200 women from F.S.U. living in the United States for fewer than 6 years, both LP and English usage had a significant inverse correlation with depressive symptom severity assessed by CES-D [Reference Kang, Basham and Kim26]. In the same sample, English usage but not LP explained 33.0% of CES-D score variance [Reference Kang, Basham and Kim26]. Interestingly, LP was positively correlated with employment status [Reference Wong, Cheung and Hart22,Reference Wu, Renzaho, Hall, Shi, Ling and Chen51] and social interactions [Reference Wong, Cheung and Hart22]. Furthermore, age, education, and LP at baseline predicted LP at follow-up [Reference Wu, Renzaho, Hall, Shi, Ling and Chen51]. Finally, gender differences impacted on language skills. In fact, LP was found to be lower in Asian female participants than in males at baseline, even if female participants’ skills improved over time [Reference Wu, Renzaho, Hall, Shi, Ling and Chen51].
Anxiety
Low LP was associated with higher prevalence of anxiety symptoms among migrants [Reference Arévalo, Tucker and Falcón29,Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland48]. Among Russian and Kurdish women, LP significantly correlated with anxiety severity as assessed by the Hopkins symptoms checklist (HSCL-25) [Reference Nicholson, Miller, Schwertz and Sorokin44]. LP again significantly correlated with reduced anxiety among Vietnamese refugees resettled in the United States [Reference Halcón, Robertson, Savik, Johnson, Spring and Butcher31]. Interestingly, both gender and ethnicity differences impacted on anxiety symptom interplay with LP: Somali migrants and males did not show the same association found for Russian or Kurdish women [Reference Nicholson, Miller, Schwertz and Sorokin44]. Moreover, LP did significantly mediate the positive relationship between lifetime exposure to traumatic events and the severity of anxiety [Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland48] and was a significant predictor for anxiety severity decrease over time in asylum seekers living in Germany [Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52].
Post-traumatic stress disorder
A negative association was shown between low LP and higher PTSD prevalence [Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland48,Reference Nickerson, Liddell, Maccallum, Steel, Silove and Bryant65] or PTSD symptoms severity [Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52,Reference Alegría, Mulvaney-Day, Torres, Polo, Cao and Canino62,Reference Kim and Chen63]. LP was a significant predictor of lower PTSD symptoms or their reduction over time among asylum seekers in Germany [Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52], Iraqi refugees [Reference Kim and Chen63], and Somali or Oromo young refugees living in the United States [Reference Alegría, Mulvaney-Day, Torres, Polo, Cao and Canino62]. The same findings were replicated in Cambodian migrants, with greater LP being protective against PTSD [Reference Kim and Kim27]. Moreover, among Afghan refugees resettled in Australia, communication difficulties were more frequently reported by participants presenting with PTSD symptoms compared to non-PTSD subjects from the same community, with low LP predicting higher risk of PTSD in different multivariate models [Reference Carl, Frias, Kurtevski, Copo, Mustafa and Font54]. Interestingly, among Somali and Oromo refugees, LP varied significantly between men and women, with men reporting better language skills, but no gender differences were observed for mental disorder prevalence or severity [Reference Alegría, Mulvaney-Day, Torres, Polo, Cao and Canino62]. Gender differences were highlighted also among humanitarian migrants in Australia with low LP (defined as higher linguistic barriers) predicting higher PTSD scores in male refugees but not in female ones [Reference Miller and Chandler58]. Interestingly, even though they reported lower rates of PTSD, the effect of LP on PTSD seemed to be higher among male migrants arrived in Australia between 2016 and 2017 [Reference Miller and Chandler58].
Somatic symptoms
Greater LP was associated with lower prevalence of somatization [Reference Fenta, Hyman, Rourke, Moon and Noh33] and lower symptom severity [Reference Ngo, Tran, Gibbons and Oliver45], even after adjusting for demographic characteristics and migration factors.
Psychotic disorders
In a multicentric study conducted by Jongsma et al. on individuals with first-episode psychosis and healthy controls, including almost 35% of migrants in both groups, the authors defined “Linguistic Distance” combining LP in the host country language with a language-tree measured distance between first language and host country language [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50]. The authors found an almost two-fold increase in the odds of psychosis among those reporting linguistic distance [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50]. Moreover, linguistic distance was significantly associated with an ethnic minority background [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50]. According to another study, individuals with psychosis presented greater odds of language-related problems than those diagnosed with major depressive disorder (MDD), bipolar disorder (BD), or substance use disorder (SUD) [Reference Finch, Kolody and Vega23].
Subclinical symptoms and psychological distress
Four of the papers reviewed did not find any significant association between LP and clinically relevant major mental disorders; nevertheless, they found a negative relationship between LP and subclinical symptoms and/or psychological distress [Reference Kang, Boyas and Salehin25,Reference Kartal, Alkemade and Kiropoulos39,Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland47,Reference Hamwi, Lorthe and Barros49]. Low LP was associated with higher stress response evaluated by salivary cortisol levels [Reference Kartal, Alkemade and Kiropoulos39], higher alienation [Reference Kartal, Alkemade and Kiropoulos39], more pronounced “migration stress” [Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland47] and lower health literacy [Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland47]. Moreover, the exposure to premigratory trauma correlated with the severity of depression, and this relationship was mediated by LP: subjects with lower LP seemed to be more prone to developing depressive sequelae following exposure to premigratory trauma [Reference Kang, Boyas and Salehin25]. Among Puerto Ricans resettled in the United States because of Hurricane Maria, English language preference protected from psychological distress, as evaluated by the K-6 scale [Reference Hamwi, Lorthe and Barros49].
Better LP predicting higher prevalence of mental disorders or symptom severity
Two of the papers reviewed [Reference Jongsma, Gayer-Anderson, Tarricone, Velthorst, van der Ven and Quattrone46,Reference Chen, Zhang, Liu and Wang66] found higher LP predicting higher prevalence [Reference Chen, Zhang, Liu and Wang66] or severity [Reference Jongsma, Gayer-Anderson, Tarricone, Velthorst, van der Ven and Quattrone46] of psychiatric symptoms. Both papers assessed depression, PTSD, anxiety and other mental disorders in the same population [Reference Jongsma, Gayer-Anderson, Tarricone, Velthorst, van der Ven and Quattrone46,Reference Chen, Zhang, Liu and Wang66]. Specifically, higher LP predicted a higher risk for both lifetime and past-year risk of psychiatric disorders (MDD, AD, and SUDs) among Latinos, with higher prevalence of MDD in males and AD in females [Reference Chen, Zhang, Liu and Wang66]. Moreover, higher LP was related to higher anxiety symptom severity among female refugees resettled in Australia [Reference Jongsma, Gayer-Anderson, Tarricone, Velthorst, van der Ven and Quattrone46].
No relationship between LP and mental disorders
Four of the papers reviewed [Reference Takeuchi, Zane, Hong, Chae, Gong and Gee37,Reference Cordero and Kurz41,Reference Söndergaard and Theorell43,Reference An, Moon and Cha53] did not find any significant association between LP and mental disorders. No significant relationships between LP and depressive symptoms severity were found among elderly Korean immigrants living in Canada [Reference Takeuchi, Zane, Hong, Chae, Gong and Gee37] or female Asian immigrants residing in Korea after an international marriage [Reference Cordero and Kurz41]. No direct relationship was found between LP and depressive symptoms in Chinese mothers living with their children in the United States, even though LP was found to exert an indirect effect (mediated by economic and social factors) on depressive symptoms [Reference An, Moon and Cha53]. One other study did not find any significant relationships between LP and the prevalence of PTSD and grief-related symptoms [Reference Söndergaard and Theorell43].
The influence of sociodemographic and migration factors on the relationship between LP and mental disorders
Among the studies collected, the evidence of a difference in mental disorder prevalence and/or severity between first- and second-generation migrants was insufficient. Seven studies included both first- and second-generation migrants [Reference Wong, Cheung and Hart22,Reference Arévalo, Tucker and Falcón29,Reference Bernstein, Park, Shin, Cho and Park30,Reference Ngo, Tran, Gibbons and Oliver45,Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50,Reference Kim, Kim, Moon, Park and Cho64,Reference Chen, Zhang, Liu and Wang66], but only two of them assessed the correlation of LP and mental disorder prevalence and/or severity by migrant generation [Reference Wong, Cheung and Hart22,Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50]. First-generation migrants showed stronger effects of language distance on risk of psychosis, whereas second-generation migrants seemed to be more affected by social disadvantages [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50]. On the other hand, LP negatively correlated with the severity of both depression and psychological distress in a mixed sample of both first- and second-generation Chinese migrants, but, in adjusted models, the correlation between depression severity and LP was still significant only in the second-generation sample [Reference Wong, Cheung and Hart22].
Four of the studies collected were based on mixed and balanced populations and found gender-related differences in the interplay between LP and mental disorders [Reference Bernstein, Park, Shin, Cho and Park30,Reference Kartal, Alkemade and Kiropoulos39,Reference Nicholson, Miller, Schwertz and Sorokin44,Reference Chen, Zhang, Liu and Wang66]. More specifically, among women but not men, higher LP was protective against alienation [Reference Kartal, Alkemade and Kiropoulos39] and somatization [Reference Nicholson, Miller, Schwertz and Sorokin44]. On the other hand, among Asiatic migrants LP was a protective factor against lifetime and 12-month psychiatric disorder prevalence only among male participants [Reference Bernstein, Park, Shin, Cho and Park30]. Among Latinos, highly proficient men showed a higher risk of lifetime MDD, while a higher risk of lifetime AD was found for more proficient women [Reference Chen, Zhang, Liu and Wang66]. Both men and women with better LP showed an increased risk of SUD [Reference Chen, Zhang, Liu and Wang66]. Eight more studies were based on women-only samples [Reference Kang, Basham and Kim26,Reference Arévalo, Tucker and Falcón29,Reference Haasen, Lambert, Mass and Krausz36,Reference Cordero and Kurz41,Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42,Reference Jongsma, Gayer-Anderson, Tarricone, Velthorst, van der Ven and Quattrone46,Reference An, Moon and Cha53,Reference Beiser and Hou56]: one of them found a negative correlation between LP and AD prevalence rates [Reference Arévalo, Tucker and Falcón29]; one of them found that LP was negatively related with postpartum depression prevalence [Reference Beiser and Hou56]; two of them found that LP negatively correlated with depression severity [Reference Kang, Basham and Kim26,Reference Haasen, Lambert, Mass and Krausz36]; three studies did not find any correlations between LP and depression severity [Reference Cordero and Kurz41,Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42,Reference An, Moon and Cha53]; and one found that greater LP predicted more pronounced anxiety symptoms [Reference Jongsma, Gayer-Anderson, Tarricone, Velthorst, van der Ven and Quattrone46]. Two more studies were unbalanced, with more than 75% female participants [Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland47,Reference Schweitzer, Vromans, Brough, Asic-Kobe, Correa-Velez and Murray61]. In the only study based on a sample of males alone, depressive symptom severity was found to correlate negatively with LP [Reference Wong, Cheung and Hart22].
Quality assessment and limitations
The QATSO Score for each study is reported in Table 1. The most frequent limitations were the lack of validated instruments for LP assessment and the absence of information about response rate among surveyed individuals in most studies. Furthermore, migration history data are missing in 12 papers, while about 15 did not provide information about privacy or ethical requirements or failed to meet them. Sampling methods representative of the population were employed in most studies, as well as validated psychiatric instruments and checking for confounding factors.
Discussion
Our work of review shows that the large majority of the studies reported a significant association between lower LP and higher prevalence and/or severity of psychiatric symptoms and mental disorder. Only two records found the opposite relationship while four papers reported no association. Low LP was consistently associated with several mental illnesses in migrants, including psychosis, mood, anxiety, and PTSDs. This result was replicated in both cross-sectional and longitudinal studies. Most of the studies included had a cross-sectional design, which do not allow to draw causal relationships. Notably, all the longitudinal studies included in this review reported a positive effect of LP acquisition over time on the prevalence of mental disorders [Reference Wu, Renzaho, Hall, Shi, Ling and Chen51] or symptom severity [Reference Kindermann, Zeyher, Nagy, Brandenburg-Ceynowa, Junne and Friederich52,Reference Schweitzer, Vromans, Brough, Asic-Kobe, Correa-Velez and Murray61,Reference Kim and Chen63], suggesting a protective effect of LP on mental health.
As is known, most studies showed that migrants present higher risk of severe mental disorders than natives. Several hypotheses have sought to explain the poor mental health in migrants. Both pre- and postmigratory factors such as lack of social support, social networks and social opportunities in the host country have been invoked [Reference Bhugra and Becker4,Reference Morgan, Charalambides, Hutchinson and Murray5,Reference Tarricone, D’Andrea, Jongsma, Tosato, Gayer-Anderson and Stilo67], and LP may play a role in this process [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50]. Deficits in social skills and executive functions have proved to be a key feature of several mental disorders and may prevent appropriate LP acquisition, exacerbate isolation and lead to psychiatric symptoms [Reference Healey, Bartholomeusz and Penn68]. In this light, poor LP may be considered as a phenotype of the vulnerability to psychiatric disorders. At the same time, low LP may impair social functioning in the host country, which is well known to be related to poor mental health. On the contrary, the results of this review suggest that adequate LP might promote migrants’ inclusion in all the main societal areas and thus positively shape migrants’ adaptation and mental health in both the short- and the long-term.
Interestingly, some findings support the hypothesis that LP may have differing degrees of relevance through the postmigratory phase. Upon arrival in the host country, LP acquisition has been shown to be associated with cultural conflict and a sense of alienation. In this early postmigration phase, a strong ethnic identity along with adequate social support may be helpful for adaptation and psychological wellbeing of migrants. During later stages of the postmigration phase, however, an adequate LP and familiarity with the host culture are decisive for both migration and mental health outcomes. For instance, higher depression prevalence rates have been found among migrants who reported insufficient LP about 10 years after the initial migration [Reference Wu, Renzaho, Hall, Shi, Ling and Chen51]. Again, there is evidence that good LP is associated with lower levels of postmigratory social disadvantage: good LP has been linked with better education [Reference Takeuchi, Zane, Hong, Chae, Gong and Gee37,Reference Rask, Suvisaari, Koskinen, Koponen, Molsa and Lehtisalo42,Reference Wu, Renzaho, Hall, Shi, Ling and Chen51,Reference Schulze, Hall, Mcdonald, Marshall, Walshe and Murray60], more affective relationships with the host population [Reference Wong, Cheung and Hart22], nonsingle marital status [Reference Takeuchi, Zane, Hong, Chae, Gong and Gee37], having health insurance [Reference Njeru, DeJesus, St Sauver, Rutten, Jacobson and Wilson69], and higher levels of postmigration occupation [Reference Wong, Cheung and Hart22,Reference Kim, Yang, Kwon and Kim28,Reference Wu, Renzaho, Hall, Shi, Ling and Chen51].
With regard to generation differences, our review shows that findings on the association between LP and poor mental health vary in relation to specific disorders: while less proficient first-generation migrants seemed more vulnerable to psychosis [Reference Ventriglio, Bellomo, Petito, Pascucci, Cuozzo and Vitrani50], second generations were more prone to depression [Reference Hamrah, Hoang, Mond, Pahlavanzade, Charkazi and Auckland47]. Further studies should clarify these associations in order to achieve better understanding of the etiopathogenic mechanisms underlying the differential effect of LP on mental health by generational status.
Migrant women are often considered more vulnerable to psychiatric disorders [Reference Porter and Haslam70,Reference Jarallah and Baxter71]. We found insufficient evidence of gender-related differences in the interplay between LP and mental disorders prevalence and/or severity. Interestingly, two studies found a difference in LP between male and female participants [Reference Kim, Yang, Kwon and Kim28,Reference Alegría, Mulvaney-Day, Torres, Polo, Cao and Canino62], with women showing lower LP at the time of migration but significantly greater improvement over time.
Strengths and limitations
The strength of our study is the global prospective based on the extensive literature review carried out including studies conducted in a variety of countries and over a 30-year timespan. Despite this strength, we have to recognize that the majority of studies included present several limitations: (a) the lack of validated LP assessment tools resulting in high heterogeneity in LP assessment within studies, potentially contributing to differences in LP estimates; (b) the heterogeneity in psychopathological assessment instruments (Table S2) which may have resulted in different estimates of mental disorder prevalence and severity across the review; (c) the cross-sectional design of most of the studies included, which does not allow for causal relationships; (d) the heterogeneity of study samples in age, gender, education level, and ethnicity, which may reflect differences in target populations as well as in sampling design. Furthermore, only a few studies investigated the link between LP and psychosis [Reference Atkinson16,Reference Kartal, Alkemade and Kiropoulos39]. The lack of sample, outcome, and methods homogeneity among the reviewed studies may be responsible for the different estimates of the interplay between LP and mental health. Such heterogeneity has not allowed to conduct meta-analysis and might limit the generalizability of results. Finally, low LP may be considered as one of the main barriers involved in reducing migrants’ access to health services and thus limiting the participation in the studies [Reference Wong, Cheung and Hart22]: the samples included may thus not be completely representative of the general population of migrants with mental disorders.
Conclusions
Low LP was generally found to negatively affect migrants’ mental health, being associated with increased prevalence and severity of psychiatric symptoms as well as of mental disorders. This finding seems to indicate that people who are not proficient in the majority language may encounter mental health issues and consequent poor social and job functioning. In the light of our results, we would expect that early and effective interventions targeting migrants’ LP could reduce both incidence and prevalence of mental disorders among migrants in the medium-long term. Given these evidences, clinicians treating migrants with mental disorders should consider integrating interventions aimed at supporting language literacy as part of their individualized care programs.
European countries requiring LP in order to obtain permanent residence increased from one in 1990 to 18 in 2014 [72]. In some countries, language acquisition is supported by language courses and, in some cases, language or integration courses are mandatory for migrants and asylum seekers during the process of nationality acquisition [72]. Several no-profit organizations provide free language classes for migrants in different countries. Anyway, only two countries worldwide actually provide freely available language courses for migrants upon arrival (Sweden and Canada) [72]. Recently, Morrice and colleagues published an interesting article entitled “You cannot have a good integration when you do not have a good communication” [Reference Morrice, Tip, Collyer and Brown73]. We would like to readopt this concept and say that “you cannot have good mental health if you do not have a good communication.” We strongly hope that the present findings will inspire authorities to provide support and courses to improve migrants’ LP upon arrival. At the same time, in order better to evaluate the relationship between LP and psychiatric disorders we need longitudinal studies on larger samples of migrants.
Acknowledgements
We thank Dr. James B Kirkbride (Division of Psychiatry, University College London, London, England) and Dr. Hannah E. Jongsma (Department of Psychiatry, University of Cambridge, Cambridge, England) for their encouragement and suggestions on our work.
Financial Support
The authors received no specific funding for this work.
Conflict of Interest
Authors declare no competing interests.
Authorship Contributions
I.T., C.M., and G.D. contributed to the conception and design of the work. C.M., G.D., and F.C. contributed to data acquisition and interpretation as well as to the first draft of the manuscript. I.T., M.P., R.M., G.M., and M.D.G. critically revised the article. All authors approved the final version of the work.
Open access
Comments
No Comments have been published for this article.