Excess use of coercive measures in psychiatry among migrants compared with native Danes

Abstract

Norredam M, Garcia-Lopez A, Keiding N, Krasnik A. Excess use of coercive measures in psychiatry among migrants compared with native Danes.

Objective:  To investigate differences in risk of compulsory admission and other coercive measures in psychiatric emergencies among refugees and immigrants compared with that among native Danes.

Method:  A register-based retrospective cohort design. All refugees (n = 29 174) and immigrants (n = 33 287) who received residence permission in Denmark from 1.1.1993 to 31.12.1999 were included and matched 1 : 4 on age and sex with native Danes. Civil registration numbers were cross-linked to the Danish Psychiatric Central Register and the Registry of Coercive Measures in Psychiatric Treatment.

Results:  Refugees (RR = 1.82; 95%CI: 1.45; 2.29) and immigrants (RR = 1.14; 95%CI: 0.83; 1.56) experienced higher rates of compulsory admissions than did native Danes. This was most striking for refugee men (RR = 2.00; 95%CI: 1.53; 2.61) and immigrant women (RR = 1.73; 95%CI: 1.45; 2.60). Moreover, refugees and immigrants experienced higher frequencies of other coercive measures during hospitalisation compared with native Danes.

Conclusion:  Coercive measures in psychiatry are more likely to be experienced by migrants than by native Danes.

Significant outcomes

• •
   Migrant psychiatric patients are more often admitted involuntarily than Danish-born psychiatric patients.
• •
   The results were most striking for refugee men and immigrant women.
• •
   Migrant patients more often experience compulsory detention, coerced treatment and use of physical force during psychiatric hospitalisation than Danish-born patients.

Limitations

• •
   Migration status (i.e. refugee status vs. family reunification) refers to only one dimension of being a migrant. The limited size of the cohort did not allow correction for other explanatory factors such as country of origin.
• •
   Socioeconomic factors were not considered because the data are inconsistent and of low quality for first generation migrants.
• •
   Data on use of coercive measures during hospitalisation were included for a 4-year period only. This allowed limited analysis only.

Introduction

Denmark, like other European countries (1), has a provision for the use of coercive measures in its mental health care law from 2006 (2) A coercive measure is defined as any act to which the patient has not given informed consent. Patients can be subjected to compulsory admission if i) the patient is psychotic or similar and iia) the patient is of danger to himself/herself or others (‘danger criterion’) or iib) the patient’s health is endangered (‘treatment criterion’). The conclusion of the medical assessment must have legal approval. Moreover, a patient can also be subjected to compulsory admission if he/she has a legal conviction to treatment (‘the other criterion’). During hospitalisation patients may be involuntarily detained and receive treatment under coercion (e.g. medicine, nutrition and treatment of physical disorders). In addition, medical staff are legally permitted to use physical force (e.g. to restrain a patient physically or with the use of a belt) and acute medication under certain circumstances.

Studies from the UK have found an increased risk of compulsory admissions among Blacks and Afro-Caribbeans compared with Whites (3, 4). In a systematic review, Singh et al. (5) included 19 UK studies and showed that Blacks and Asians were respectively 3.8 and 2.0 times more likely to be admitted to compulsory psychiatric care than Whites. Other coercive measures may be used during hospital admission, including compulsory detention, coerced treatment and use of physical force. However, only two UK studies address this area. Gudjonsson et al. (6) found that Black in-patients were more likely than their White counterparts to receive coerced treatment, but that they were not more likely to be physically restrained. Commander et al. (7) found that Blacks more often than Whites were involuntarily detained in hospital.

Until recently, it was unknown whether the higher use of compulsory admission and other coercive measures was specific to the UK or if it also occurred in other European countries and among other migrant groups. However, new studies from respectively Switzerland (8) and the Netherlands (9) also point toward higher rates of compulsory admissions to psychiatric wards among migrant populations compared with non-migrants.

The studies here all describe differences in risk of compulsory admission and coercive measures among migrants compared with non-migrants; however, these results are poorly understood. Several possible explanations have been offered including: higher rates of psychotic disorders among migrants due to migration processes; and misinterpretation of symptoms by health professionals resulting in misdiagnosis; as well as problems of timely access to mental healthcare services. Problems of access may lead to inequity in treatment and consequently to increased disease severity resulting in more frequent use of compulsory admission and other coercive measures.

Previous research on the use of compulsory admission and other coercive measures in psychiatric emergencies has several limitations: most studies are from the UK and few highlight the issue among migrant populations in other European countries; all studies use country of birth as a determinant rather than migration status (i.e. refugee status vs. family reunification etc), although this is increasingly considered an equally important determinant for migrants’ health problems and access to healthcare; studies tend to focus on patients with schizophrenic disorders and often exclude those with other psychiatric disorders (3, 4); analyses are often based on the evaluation of a single service and not on national data (5); and studies focus on compulsory admission and few include the use of other coercive measures during hospitalisation. To overcome some of these shortcomings, we conducted a national register-based retrospective cohort study to compare differences in risk of compulsory admission and other coercive measures in psychiatric emergencies among individuals with three major psychiatric disorders. In this study, we distinguish between refugees and family reunification immigrants as these groups are introduced differently to the Danish health care system upon arrival, which may affect their later health-seeking behaviour. Also, the two groups may have different morbidity patterns as refugees are more likely to have been exposed to traumatic life events resulting in mental illness.

Aims of the study

The study aimed at investigating the following two hypotheses: i) refugees and immigrants are at higher risk of experiencing compulsory admissions compared with Danish-born individuals and ii) refugees and immigrants are at higher risk of experiencing other coercive measures during hospitalisation compared with Danish-born individuals.

Material and methods

Study cohort

The study cohort was obtained through the Statistical Department at the Danish Immigration Service. Migrants who obtained residence permission as refugees or through family reunification in Denmark from 01.01.1993 to 31.12.1999 were included; 84 379 individuals were identified. Individuals less than 18 years (n = 18 861) when they obtained residence permission were excluded. Another 3042 individuals were excluded due to missing civil registration number or because their civil registration numbers appeared more than once in the sample. In addition, 15 migrants and their controls were excluded due to problems with their registration of nationality. A Danish-born reference population was identified through Statistics Denmark, which formed a 6 : 1 matching at population level on sex and age at commencement of residence permission in Denmark. Further, all controls were Danish-born residents with Danish-born parents to avoid including second-generation migrants. Controls were used only once. We then matched 4 : 1 on an individual level on age and sex through a random sampling procedure to identify to which case every single control belonged. We were able to make a 4 : 1 matching for all refugees resulting in 145 870 individuals: 29 174 refugees and 116 696 controls. Four of the family-reunited immigrants were missing a total of five controls due to difficulties with age matching because of outlying ages; accordingly, there were 33 287 family-reunited immigrants and 133 143 controls. We performed a two-stage hierarchical matching procedure because Statistics Denmark was only able to perform a crude population-based matching as a first step due to lack of resources. The individual matching therefore had to take place as an additional step by ourselves on the bases of the population matched study cohort. Our first step matching on 6 : 1 basis was carried out to ensure sufficient controls were available to ensure a later 4 : 1 matching on an individual basis for all cases.

Characteristics of the study population are shown in Table 1. Migrants were divided into the largest geographical areas of origin according to WHO guidelines: Asia, East Europe (excl. Former Yugoslavia), Former Yugoslavia, Middle East, North Africa, and Sub-Saharan Africa (10). We have separated Iraq as a country of its own due to the large number of Iraqis included in the study. Individuals born in Western countries were excluded from the analysis. Follow-up time was defined for both refugees and family-reunited immigrants and their corresponding controls from when residence permission took effect until one of the following events: i) start date of first psychiatric contact; ii) end of study (31.12.2003); iii) date of death and iv) date of emigration. The mean follow-up time was 8 years for both refugees and controls; and 6–7 years for both family reunited immigrants and their controls. From here on family-reunification immigrants are referred to as ‘immigrants’, and ‘migrants’ refers to all groups of newcomers.

Table 1.   Characteristics of the study cohort (total n = 312,300). Migrants are distributed by refugees and immigrants and their Danish-born controls

Population characteristics	Refugees (n = 29 174)	Danish-born (n = 116 696)	Immigrants (n = 33 287)	Danish-born (n = 133 143)
Age in years (mean and quartiles)	32.9 (26.4;41.7)	32.9 (26.4;41.7)	27.5 (23.3;33.7)	27.5 (23.3;33.7)
Follow-up in years (mean and quartiles)	8.0 (6.1;8.4)	8.1 (6.5;8.4)	6.1 (4.5;8.0)	6.6 (5.1;8.3)
Sex, % (n)
Female	44.4 (12 963)	44.4 (51 852)	64.0 (21 294)	64.0 (85 174)
Region of origin, % (n)
Eastern Europe	1.9 (555)		10.8 (3581)
Former Yugoslavia	52.7 (15 369)		5.7 (1894)
North Africa	15.0 (4366)		9.5 (3148)
Iraq	15.8 (4618)		4.1 (1365)
Sub-Saharan Africa	1.8 (530)		5.9 (1967)
Asia	2.6 (747)		17.9 (5951)
West	0.1 (35)		18.4 (6153)
Middle East	10.1 (2954)		27.7 (9228)
Events during follow-up, % (n)
Deaths	2.6% (761)	3.0% (3578)	0.8% (270)	1.5% (2019)
Emigrations	9.5% (2785)	3.4% (3933)	18.6% (6182)	4.4% (5893)
Population at study closure	87.9% (25 628)	93.6% (109 185)	93.6% (26 835)	94.1% (125 231)

Data collection and analysis of compulsory admission

Civil registration numbers of the study cohort were cross-linked to the Danish Psychiatric Central Register, which contains data on all admissions to Danish psychiatric in-patient facilities. The register has used the ICD-10 coding system from 01.01.1994, which is a change from the ICD-8 coding system used formerly. As the validity of translations from ICD-8 to ICD-10 is questionable, we decided to use only ICD-10 diagnoses starting from 01.01.1994 (although the cohort was established from 01.01.1993). By including people who got residence permit in 1993 a larger study group was ensured thereby offering more strength to the study. Accordingly, we identified all individuals who had been in-patients between 01.01.1994 and 31.12.2003. This amounted to 21 001 admissions: 2271 refugees (controls n = 9285) and 1014 immigrants (controls n = 8431). This corresponded to 876 refugee patients (controls n = 2526) and 508 immigrant patients (control n = 2566) in this period. Information about the use of coerce was available for all admissions except one.

First, we divided compulsory admissions into the three criteria stated in the Danish mental healthcare law: i) danger criterion; ii) treatment criterion or iii) other criterion. The ‘other criterion’ concerns individuals with a legal conviction to treatment. To avoid over-representing a few individuals who had many repeated admissions by coerce, we used the individual person rather than the individual admission as the observational unit as follows: if a person had ever been admitted compulsorily according to the danger criterion during the study period, the person was then categorised as ‘danger criterion’; if a person had ever been admitted compulsorily according to the treatment criterion (and never the danger criterion) during the study period, the person was then categorised as ‘treatment criterion’; if a person had ever been admitted compulsorily according to the other criterion (and never the danger or treatment criterion) during the study period, the person then was categorised as ‘other criterion’. The percentage distributions on these criteria were calculated (Table 2).

Table 2.   Frequencies of compulsory admissions among refugees and immigrants compared to their Danish-born controls from 1.1.1994–31.12.2003 (total n = 6,476). Compulsory admissions are subdivided according to the main criteria of the Danish mental health care law

	Refugees, % (n)	Danish-born controls, % (n)	Immigrants, % (n)	Danish-born controls, % (n)
*If a person had ever been compulsory admitted according to the danger criterion during the study period, the person was then categorised as ‘danger criterion’. †If a person had ever been compulsory admitted according to the treatment criterion (and never the danger criterion) during the study period, the person was then categorised as ‘treatment criterion’. ‡If a person had ever been compulsory admitted according to the other criterion (and never the danger or treatment criterion) during the study period, the person was then categorised as ‘other criterion’, which includes individuals with a legal conviction to treatment.
Compulsory admissions	22.5% (197)	13.9% (348)	15.3% (78)	13.3% (342)
Danger criterion*	9.7% (84)	5.5% (138)	6.9% (35)	4.9% (127)
Treatment criterion†	6.2% (54)	3.6% (92)	4.5% (23)	3.0% (77)
Other criterion‡	6.7% (59)	4.7% (118)	3.9% (20)	5.4% (138)
Voluntary admissions	77.5% (679)	86.1% (2,178)	84.7% (430)	86.7% (2,224)
Total	100.0% (876)	100.0% (2,526)	100.0% (508)	100.0% (2,566)

The main outcome considered was rate ratio, which represents the factor by which refugees and immigrants experience coercive measures compared with their Danish-born controls. We calculated rate ratios of compulsory admissions for the three most frequent diagnostic categories according to ICD-10: ‘Psychotic disorders’ (F20-29: Schizophrenia, schizotypal and delusional disorders), ‘Affective disorders’ (F30-39: Mood disorders) and ‘Neurotic disorders’ (F40-F48: Neurotic, stress-related and somatoform disorders) (Table 3). Also, we analysed ‘all psychiatric disorders’ combined of which the three above mentioned diagnostic categories accounted for 70% (14.646/21.001) of all first time admissions. Table 3 is based on calculations of first-time admissions because our initial analysis showed that 90% of the admission type (voluntary versus compulsory) of first admissions remained for the rest of the contacts. First admissions were dichotomised into: i) compulsory admission or ii) voluntary admission. For ‘all psychiatric disorders’ the first admission is chosen (whatever type of diagnosis). For selected diagnostic groups (psychotic, affective and neurotic) the diagnostic category was first selected and within this all first admissions were chosen. Thus, the category ‘all psychiatric disorders’ is not a total of the selected diagnostic categories, because a person may have had a first admission with an affective disorder and a later first admission for a psychotic disorder but still only count for one admission in the ‘all disorders’ category. The analyses were done separately for refugees and immigrants comparing with their corresponding controls native Danish-born controls. The analyses were stratified by sex. Rate ratios and 95% confidence intervals were evaluated by Poisson regression with log person years as off-set and with a log link. This was done using PROC GENMOD in sas version 9.1.

Table 3.   Rate ratios and 95% CI of compulsory first-time admissions among refugees and immigrants compared with their Danish-born controls distributed by diagnostic group and sex. Danish-born controls constitutes the reference group (RR = 1.00)

Psychiatric disorders according to ICD-10	Refugees		Immigrants
	RR (95% CI)	n* (Ref/Cont)	RR (95% CI)	n* (Imm/Cont)
*n refers to the total number of first time admissions (compulsory or not) for the diagnostic category in the studied time period. †The category all psychiatric disorders combines psychotic, affective, neurotic disorders and other types of psychiatric disorders.
Psychotic disorders
Men	1.56 (1.08;2.25)	248/459	1.00 (0.50;1.98)	63/352
Women	1.04 (0.57;1.91)	108/289	1.23 (0.68;2.20)	113/381
Total	1.44 (1.06;1.96)	356/748	1.07 (0.69;1.66)	176/733
Affective disorders
Men	1.11 (0.39;3.08)	78/290	0.36 (0.04;2.83)	37/181
Women	1.35 (0.46;3.96)	125/367	0.75 (0.17;3.32)	81/489
Total	1.16 (0.55;2.44)	203/657	0.58 (0.17;1.92)	118/670
Neurotic disorders
Men	2.64 (1.30;5.38)	207/302	0.47 (0.13;1.63)	68/224
Women	0.89 (0.23;3.43)	116/277	2.63 (1.06;6.69)	119/412
Total	2.23 (1.23;4.05)	323/579	1.22 (0.60;2.47)	187/636
†All psychiatric disorders
Men	2.00 (1.53;2.61)	561/1470	0.73 (0.44;1.19)	190/1026
Women	1.27 (0.81;2.01)	315/1055	1.73 (1.45;2.60)	317/1540
Total	1.82 (1.45;2.29)	876/2525	1.14 (0.83;1.56)	507/2566

Data collection and analyses of coercive measures during hospitalisation

The study of coercive measures during hospital admission was based on a subset of the total population because the Registry of Coercive Measures in Psychiatric Treatment was inaugurated on 01.01.1999. Therefore, this analysis includes only individuals in the cohort with psychiatric in-patient admissions between 01.01.1999 and 31.12.2003. This amounted to 1524 admissions for refugees (controls n = 6444) and 723 admissions for immigrants (controls n = 6034). This corresponded to 652 refugee patients (controls n = 1929) and 398 immigrant patients (control n = 2071). Data were stratified into the following variables during admission: i) compulsory detention; ii) coerced treatment and iii) use of physical force and/or acute medication. The percentage distributions on these criteria were calculated (Table 4).

Table 4.   Frequencies of different forms of coercive measures during hospitalisation among all admissions *distributed by refugees and immigrants and their corresponding Danish-born controls

All admissions 1999–2003*	Refugees (n = 1524)	Danish-born controls (n = 6444)	Immigrants (n = 723)	Danish-born controls (n = 6034)
*All admissions during the period 1st January 1999 to 31st December 2003.
Any kind of coercive measure	55.8% (851)	23.9% (1543)	47.9% (347)	30.3% (1829)
Compulsory detention	11.7% (178)	6.8% (441)	14.2% (103)	7.5% (451)
Coerced treatment	5.8% (89)	2.0% (128)	4.5% (33)	2.0% (121)
Use of physical force and/or acute medication	38.3% (584)	15.1% (974)	29.1% (211)	20.8% (1257)

Results

Table 1 shows characteristics of the study cohort. Migrants are distributed by refugees and immigrants and their corresponding Danish-born controls. The cohort had a mean age of about 33 years among refugees and 28 years among immigrants. Notably, refugees were followed for an average of 8 years and immigrants for an average of 6 years. Among refugees 44% were women and among immigrants 64% were women. Among refugees approximately half the cohort came from the Former Yugoslavia, whereas immigrants originated mainly from the Middle East, Western countries and Asia.

Table 2 shows the frequencies of all compulsory and voluntary admissions based on all individuals admitted during the study. Among refugees, 22.5% were admitted compulsorily, whereas this was the case for only 13.9% of their Danish-born controls. In contrast, 15.3% of immigrants were admitted compulsorily, which presented less of a difference compared with their Danish-born controls (13.3%). The most common criterion for compulsory admission for both refugees and immigrants was the danger criterion. Refugees and immigrants presented with percentagewise more coerced admissions according to all three criteria, except for immigrants admitted according to ‘other criterion’.

Table 3 shows rate ratios of compulsory first-time admissions among refugees and immigrants compared with their Danish-born controls distributed by diagnostic group and sex. Danish-born constitutes the reference group. For psychotic and neurotic disorders, refugee men were significantly more likely to have been admitted compulsorily; this was not the case for affective disorders. Among all psychiatric disorders combined, refugee men were twice as likely to have had a compulsory admission compared with their Danish-born controls (RR = 2.00; 95%CI: 1.53; 2.61). In contrast, refugee women did not differ significantly from their Danish-born controls in the likelihood of ever having had a compulsory admission in any of the diagnostic categories. However, in total, refugee men and women were significantly more likely than their Danish-born controls to have had compulsory admissions (RR = 1.82; 95%CI: 1.45; 2.29).

Table 3 also shows that immigrant men did not differ significantly from their Danish-born controls in their risk of compulsory admissions in any of the analysed diagnostic categories. In contrast, immigrant women showed a consistent trend towards more compulsory admissions for all diagnostic categories, although the results were significant only for neurotic disorders (RR = 2.63; 95%CI: 1.06; 6.69) and all disorders combined (RR = 1.73; 95%CI: 1.45; 2.60). In total, immigrants were slightly more likely to have been involuntarily admitted than native Danes (RR = 1.14; 95%CI: 0.83; 1.56), but the result was not significant.

Table 4 shows the frequencies of three different forms of coercive measures during admission in all hospital admissions during the study period. Some kind of coercive measure had been used in 55.8% of all admissions among refugees compared with 23.9% among their Danish-born counterparts. This striking difference was also true for admissions among immigrants in as much as some kind of coercive measure had been used in 47.9% of all admissions among immigrants compared with 30.3% among their Danish-born counterparts. The frequencies of: i) involuntary detention; ii) coerced treatment and iii) use of physical force and/or acute medication were about twice as high for all coercive measures for both refugees and immigrants compared with their Danish-born controls.

Discussion

We studied a sample of 6476 individuals with psychiatric in-patient contact from a total cohort of 312 300 persons. Our hypotheses were that refugees and immigrants were at higher risk of experiencing, respectively, compulsory admissions and other coercive measures during hospitalisation compared with Danish-born individuals. These hypotheses were confirmed. First, refugees and immigrants showed a tendency towards higher rates of compulsory admissions. The results were significant for refugee men with psychotic and neurotic disorders and in all diagnostic groups and for refugees in total, as well as for immigrant women with neurotic disorders and all immigrant women combined. Secondly, refugees and immigrants experienced use of i) compulsory detention; ii) coerced treatment and iii) use of physical force and/or acute medication about twice as often as did native Danes.

Weaknesses and strengths of the study

Before discussing the results the study’s methodological weaknesses and strengths should be noted. First, we did not control for sociodemographic variables other than age and sex because the available register data are inconsistent and of low validity for first generation migrants. However, studies have shown increased use of coercive measures independent of employment status and accommodation circumstances (3, 4, 8). Secondly, we did not stratify for country of birth as our numbers would then be insufficient for substantial analysis. Thus, we cannot rule out an effect of country of birth on our results. Thirdly, we have included the 1993 population to extend the population and thereby provide more statistical strengths to the study. We have no reason to believe that this introduces biased results. Fourthly, data on use of coercive measures were not collected before 1999 when the Registry of Coercive Measures in Psychiatric Treatment was started by the National Board of Health. Consequently, numbers were low even on contact level and therefore allowed only for frequency calculations, which renders conclusions less assertive. Thus, the data are descriptive, which warrants further analytical follow-up studies. Lastly, we used data from the Danish Psychiatric Central Register. Like other registers in psychiatric research the Danish Psychiatric Central Register is based on data on which only scarce systematic evidence exist regarding its validity for research purposes (11).

The study also has some strength. First, it is based on national register data. Thus, data do not depend on a single service; instead, they present a general picture on a national level during the study period. Secondly the population is larger than that in other studies; it is also robust as it constitutes a well-defined, large sample size based on an individual matching procedure on age and sex with native-born Danes with clear inclusion and exclusion criteria. Thirdly, data enabled us to divide migrants into refugees and immigrants based on information on migration status from the Danish Immigration Service.

Why do refugees and immigrants experience more compulsory admissions compared with native Danes?

Our results are supported by other studies on migrants that likewise find excessive rates of compulsory admissions independent of diagnosis and socioeconomic factors among migrants compared with non-migrants (3, 4, 8, 12, 13). To explain our findings, we have to investigate factors that are operating at or prior to first admission which may increase the risk of compulsory admission. Various hypotheses exist that attempt to explain excessive rates of compulsory admissions among migrants; however, most explanations are largely unsupported (5). Explanations are either patient-related or service-related; and below we attempt to relate our results to some of these explanations.

On the patient side, it has been proposed that more compulsory admissions are related to higher rates of psychosis among migrants than among non-migrants (14, 15). For example, it has been hypothesised that the process of migration is a risk factor for psychotic disorders among individuals who are already vulnerable (16–18). Our study allows us to divide migrants according to migration status; and our data supports this explanation in the sense that refugees – who are likely to have had a difficult migration process including premigration trauma – indeed have higher rates of compulsory admission in the study compared with native Danes. However, we do not have any plausible explanations for the gender differences that we find regarding that refugee men compared to women are more exposed to use of coerce. As such especially asylum seekers find themselves exposed to a number of stressors in exile including the uncertainty of prolonged stressful asylum procedures. Secondly, our findings may be explained by UK multicentre studies of ethnicity and psychosis which stress the importance of a supportive family structure in ensuring access to healthcare in ethnic groups. For ethnic minorities these studies show higher levels of social disadvantage over the life course compared to White British (19, 20). Indeed more refugees come to Denmark alone or with few relatives which makes them more vulnerable and which may result in a vulnerable network which is unable to help ill individuals in gaining access. This may contribute to excess compulsory admissions among refugees. Thirdly, it may be that the stigma of having a mental disorder is more pronounced among migrant populations, leading to social isolation and consequent delay in help-seeking. Fourthly, language and communication barriers may complicate access to mental healthcare; this is particular relevant for psychiatric patients where verbal communication and the anamnesis are crucial clinical tools. However, so far, the importance of stigma, language and communication has not been investigated in a Danish context.

On the side of the healthcare system, language also constitutes a problem that includes lack of interpreters and multilingual staff. Secondly, Denmark is traditionally an ethnic homogenous society; and unfortunately the rising diversity is not yet reflected among different groups of health professionals in the healthcare system. Therefore, health professionals should be encouraged to develop cultural competence, which implies ‘the ability of people of one culture to understand, communicate, operate and provide effective services to people of another culture’ (21). If not statutory services may be perceived by migrants as culturally insensitive and even discriminatory. As Singh & Burns (22) point out such perceptions are important to met because if not they end like a self-fulfilling prophecy contributing to more mistrust of services by migrants thereby leading to delayed help seeking with increased use of coercive measures for migrants. We cannot rule out that such views may have contributed to the excessive rates of coerce that we identified. Secondly, our results may be explained by lack of diagnosis of mental disorders among migrants resulting in later increased disease severity. Indeed, studies have shown that general practitioners tend not ask their foreign-born patients about trauma (23). Consequently, mental illness among migrants is less likely to be detected and less likely to receive active management (24). This may also in part explain why a UK study showed that compared with White British patients, general practitioner referral was less frequent for both Black patients and referral by a criminal justice agency was more common (25). Interestingly the study also showed that family members were more likely to seek access to help for a psychotic family member from the legal system rather than the medical system. In Denmark pathways of care for first time and recurrent psychosis have so far not been mapped. Fourthly, migrants may have complex needs that challenge existing service providers (26). In Denmark only one psychiatric hospital clinic have special skills in migrants’ mental health problems. These skills tend to be restricted to privately-funded rehabilitation centres for traumatised migrants; however, these services treat only out-patients and only tortured or otherwise traumatised individuals. Lastly, awareness of relevant service offers may be low among migrants. In Denmark, neither refugees nor immigrants receive a systematic introduction to the Danish healthcare system, and there is no official guidance in establishing initial contact with general practitioners. Asylum seekers are often introduced to the Danish healthcare system by the Danish Red Cross while residing in the asylum centres. In case of quota refugees – who come directly from refugee camps and are dispersed in municipalities all over Denmark – it is entirely up to the local municipality or individual general practitioners to make an introduction to health services. Also, family reunification immigrants have to rely on their general practitioners or relatives for an introduction. Lack of awareness about services may be even more pronounced among immigrant women if they find themselves in patterns of sex roles which in part may increase the stress of life in exile and in part prevent them from actively integrating. This could explain our finding of excess compulsory admissions among immigrant women.

Patient- and service-related factors are likely to interact and constitute barriers in migrants’ access to relevant and effective mental health services. This may lead to delay in help-seeking and increased disease severity and thus to excessive rates of compulsory admissions. This inequable contact with services between migrants and non-migrants may result in migrant patients perceiving services as unhelpful. Consequently, there is again a delay in help-seeking, resulting in a vicious circle.

Why do refugees and immigrants experience more coercive measures during hospitalisation compared with native Danes?

We also found that refugees and immigrants experienced more compulsory detention, coerced treatment, and use of physical force and/or acute medication. There are few data available to put these results into perspective (6, 12). The high rates of use of coerce may be related to some of the above mentioned explanations including higher rates of psychosis. But, to explain our findings we must also investigate factors that are operating after admission. In this context, problems of communication – both verbal and non-verbal may be of great importance. In acute situations, agitated patients are initially sought to be verbally calmed down before physical force and/or acute medication is used, however, language barriers may complicate this process for migrant patients. Psychotic patients may become more agitated if they do not feel understood and staff may misinterpret migrant patients’ agitation. Hence, when working with refugees and immigrants who are in-patients, cultural skills could be crucial in preventing the use of coercive measures. Again, these descriptive data clearly call for further investigation, preferably by qualitative analysis.

In conclusion, we found excessive rates of compulsory admission and coercive measures among migrants compared with that among native Danes. This was most striking for refugees. Our findings are a result of interplaying factors related to the individual patient as well as the healthcare system. Most important, healthcare providers, administrators and politicians need to assess how accessible mental healthcare services are to migrants and if the services meet the needs of these populations. Further research is needed to identify the underlying factors that can explain our findings.

Acknowledgements

We thank consultant Marianne Kastrup from the Center for Transcultural Psychiatry, Rigshospitalet Copenhagen, Denmark; senior researcher Karin Helweg-Larsen from the National Institute of Public Health; and chief physician Ebbe Munk Andersen from the Danish Red Cross for advice on the project and comments on the manuscript.

Declaration of interests

None.