Health services and the treatment of immigrants: data on service use, interpreting services and immigrant staff members in services across Europe

Abstract Background The number of immigrants using health services has increased across Europe. For assessing and improving the quality of care provided for immigrants, information is required on how many immigrants use services, what interpreting services are provided and whether staff members are from immigrant groups. Methods Structured interviews were conducted with 15 health services (9 primary care, 3 emergency departments, 3 mental health) located in areas with high immigrant populations in each of 16 European countries (n = 240). Responses were collected on the availability of data on service use by immigrant patients, the provision of interpreting services and immigrant staff members. Results Data on service use by immigrants were recorded by only 15% of services. More than 40% of services did not provide any form of interpreting service and 54% of the services reported having no immigrant staff. Mental health services were more likely to use direct interpreting services, and both mental health and emergency services were more likely to have immigrant staff members. Discussion For assessing and improving the quality of care provided for immigrants, there is a need to improve the availability of data on service use by immigrants in health services throughout Europe and to provide more consistent access to interpreting services.

that language differences can lead to negative judgements and stereotyping [39].
Health services have attempted to adapt to the linguistic needs of their patients with several solutions aimed at improving communication between practitioners and immigrant patients. One example has been to employ bilingual staff, however this solution favours large services with a high demand for bilingual staff and the resources to adopt this solution for all the main patient languages [21]. Another solution has been the use of interpreting services, which can offer assistance in a variety of languages [9]. The importance of language barriers might differ in different types of health service. For mental health services, language and communication are particularly salient as they represent the principle means for diagnosis and treatment, particularly in the application of psychotherapeutic treatments [20,21].

Staff members with immigrant backgrounds
To improve the quality of health services for immigrant patients, authors have argued for the use of 'culturally sensitive' or 'culturally competent' service provisions [6,37]. Some have supported the mainstreaming of cultural sensitivity in all health services, while others have favoured separate service provisions for immigrants [6]. One method for promoting cultural sensitivity has been the use of specifi c training programmes for health service providers and practitioners. Some experts have recommended actively recruiting staff from the same cultural and ethnic backgrounds as the patient population [11,13,14]. The latter could contribute in both direct and indirect ways to improving understanding between patients and health services. Exposing services to the experience of diverse cultural understandings of health and treatment [31] may increase awareness of cultural norms and values, while indirectly lessening the impact of stereotypical views [12,38]. However, services are primarily concerned with employing staff that are competent in their work. Matching the cultural or ethnic backgrounds of their patients with their staff is a diffi cult task, and of secondary concern to most services. Nevertheless, a workforce that is more representative of the population from which patients come from, may increase sensitivity to the dynamics of alienation and cultural differences [38].
In this paper, we present the availability of data on service use, immigrant staff, and the provision of interpreting services across 16 European countries and three types of health service. The participating countries were: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Hungary, Italy, Lithuania, Netherlands, Poland, Portugal, Spain, Sweden, and the UK. The range of countries was selected to obtain a cross-European view of data availability and allocation of provisions. In each country, emergency hospital departments, mental health care services and primary care services were included in the study to gather data on the extent to which variations in service use by immigrant patients are based on service type as noted in previous fi ndings (e.g. 4,9,25]. This study provides a descriptive analysis of data collected from interviews with representative staff from these services, across the participating countries, to identify the availability of routinely collected data across Europe on service use and service provision for immigrants.

Introduction
The number of immigrants using health services in Europe has increased in recent years [26], which has given rise to the challenge of maintaining service quality, whilst still meeting the needs of diverse populations of patients [7,8]. A refl ection of these changes can be seen in an increase in the research on health service provision for immigrants in Europe [4]. However, there remains a lack of data across several European countries on key service features including: data on service utilization [29], data on the availability of interpreting services, and staff members from immigrant groups. Data on factors such as the availability and use of interpreting services are essential as they represent some of the basic elements of good practice when implementing services for immigrants [22,32].

Availability of information about service use among immigrants
Composition and extent of service use by immigrants differ between health services and countries. Information is therefore required to capture the distribution of immigrants actually using services. This requires the recording of disaggregated data on immigrant use within individual health services. At present, information on service use at this level is scarce, limiting the opportunity for comparative analysis across services and countries within Europe [4,26]. Various studies have attempted to describe immigrant preferences for accessing health services. In the United Kingdom, for example, McCrone et al. [27] reported that Somali refugees with mental health issues preferred to seek the help of general practitioners and refugee services, rather than accessing community mental health services. In addition, several other countries have reported that immigrants have a tendency to overuse emergency departments for general health problems [3,10]. Findings such as these emphasise the importance of capturing health service use across different types of service, as well as between countries.

Interpreting services
Several challenges can arise when providing health care to meet the needs of immigrants. Notable diffi culties have included: the lack of knowledge about the health care system; mistrust of public institutions [15,26]; disparities in the explanatory models of illnesses used in different cultures [19,31]; and lack of complete entitlement to the utilisation of health services [36]. The most frequently mentioned diffi culty in providing suitable health care to meet the needs of immigrants has been the barriers in language and communication [2,16,18,35]. For this specifi c reason, we have focused on the provision of interpreting services in response to challenges associated with language barriers.
Roberts et al. [33] reviewed the evidence on language barriers and concluded that language barriers hamper access to health services [1,23]. Language barriers can also result in longer visit times, fewer visits, more misunderstandings, more emergency room visits, and a reduction in satisfaction with the treatment received [26,40]. Some authors also speculated Data collected included: levels of service utilisation; patient characteristics; existence of any immigrant specifi c departments, programmes and policies; availability of specialised staff for immigrant patients; staff diversity; interpreting services and evaluation issues. Information collected concerned the preceding 12 months, or the most recent 12 month period for which fi gures were available. All data were descriptively analysed. The present paper describes fi ndings concerning service use by immigrants, the existence of data registers, the provision of interpreting services, and the presence of staff members from immigrant backgrounds. Results concerning other aspects of the dataset have been published elsewhere [32].
Ethical approval was not required for this study in the participating countries, as no patient data were sought, and the study was regarded as service evaluation.

Availability of data registers
Only 114 of the 240 services (48%) kept data-based usability fi gures on use of services by all patients. In another 109 services (45%), reported rates were based on estimates. For the remaining 17 health services (7%) neither data on actual fi gures nor estimated fi gures were made available to the interviewers.
Data on immigrant status was available for regular immigrants in 8% of services, refugees 7% of services, and asylum seekers 5% of services. Those fi gures had been lower, than for victims of human traffi cking (10% of services) or irregular immigrants (14% of services). See Table 1, for the number of all services with available data on different groups of immigrant patients by status. These fi gures do not necessarily indicate actual rates of service usage by different immigrant groups based on status, but were more indicative of when information was recorded and for which groups based on immigration status.
There were marked differences on the availability of actual data on all patients and immigrant patients for the different types of services. Comparing the three different types of services, the highest level of data availability for all patients was found in emergency departments. 69% (n=33) of emergency departments had usage data for all patients, followed by mental health services, where 56% of services (n=27) had data for all patients, whereas primary care services had the lowest proportion with 38% of services (n=54). Variations were also noted between countries

Method
The data were gathered as part of the project: European Best Practices in Access, Quality and Appropriateness of Health Services for Immigrants in Europe (EUGATE). Questionnaires were conducted with representatives from emergency departments, mental health services and primary care services, across all 16 participating countries. In each country, services were selected based on their location to ensure they provided care to immigrant populations. For each country, three districts within a major city with the highest proportion of immigrants were selected. In most cases the selected major city was the capital, with the exception of Lithuania (Kaunas) and Spain (Barcelona), and four cities were selected instead of three districts for Finland (Malax, Oravais, Pietarsaari & Vaasa) and the Netherlands (Amsterdam, the Hague, Rotterdam & Utrecht). These deviations refl ected the different distributions of emergency departments and mental health services in different countries. For a full list of services by country, see previous publication [32].
Once cities and districts were identifi ed, representatives from three emergency departments, three mental health services, and nine primary care services were recruited in each of the 16 participating countries. Resulting in respresentatives being contacted in 48 emergency services, 48 mental health services, and 144 primary care services across Europe. In total 240 structured interviews were conducted on questions regarding service organisation, utilisation of services by immigrants, and monitoring systems within the services. An interview tool, developed using a Delphi process among partners within the participating countries, was used to collect the necessary data. This tool is publicly available from the EUGATE web page http: //www.eugate.org.uk/outcomes/index.html.
Immigrants were defi ned as persons born outside of the country of current residence. In line with EU directives, this included: regular immigrants who are non-EU nationals but legally residing in the country (e.g. labour immigrants); asylum seekers awaiting a decision on their application; refugees as defi ned in Article 1 of the 1951 Convention; victims of human traffi cking; and irregular immigrants as defi ned as persons who have not been granted permission to enter, or to stay, in a given country (e.g. undocumented immigrants). Immigrant populations were not differentiated on grounds of ethnicity, culture or country of origin for this study, as these data were too heterogeneous for the participating countries. All immigrants patients 15% (35) 10% (14) 25% (12) 19% (9) Regular immigrant patients 8% (18) 6% (9) 15% (7) 4% (2) Asylum seeker patients 5% (13) 4% (5) 13% (6) 4% (2) Refugees patients 7% (16) 5% (7) 13% (6) 6% (3) Victims of human traffi cking patients 10% (24) 10% (14) 13% (6) 8% (4) Irregular (undocumented) patients 14% (34) 14% (20) 21% (10) 8% (4) of interpreting services, such as bilingual staff and mediators. 53% of services never provided any direct interpreting service, and 59% of responding services never provided a telephone interpreting service. In addition, 24% of the responding services reported always using direct interpreting service for patients with language diffi culties, and only 17% of services always used a telephone interpreting service when language barriers were present. For fi gures on the frequency of availability of direct and telephone interpreting services see Table 2.
The availability of information on interpreting services within the three types of services studied varied across countries. Altogether 101 out of the 240 services (42%) did not provide any form of interpreting service. In Finland, Sweden and the UK, all interviewed services provided some kind of interpreting service. In Denmark, the Netherlands and Spain, 14 out of the 15 services had interpreting services available for immigrant patients. In Greece, no interpreting service use was reported in any of the services, whilst in Lithuania only two services offered some type of interpreting service. By contrast, in Spain, three services offered all the three types of interpreting service listed above (direct, telephone and mixed), followed by Belgium and Germany with two services providing all three types of interpreting service. All of the interviewed services in the UK had access to two different types of interpreting service. The same was the case for 14 of the 15 services in Sweden and 10 of the15 services in Germany. In Denmark, at least 14 of the 15 services provided one or more of the types of interpreting service.
In addition to these fi gures, differences were noted between the countries regarding access and use of telephone interpreting services, which were absent in some of the participating countries. For all the services in Austria, Greece and Poland there was no access to such service, and in Germany and Hungary it was only offered by one service of the 15 interviewed.
Except Greece, where there was no interpreting service at all, all other countries had at least one service with a direct interpreting service.
Mental health services most often provided direct interpreting services (35%), followed by emergency departments (23%) and primary care services (20%). Emergency services had a higher proportion of telephone interpreting service use, as time critical care cannot wait for an interpreter to arrive, whilst mental health services provided telephone interpreting services less often than either of the other two types of service.

Staff members with immigrant background
For the 240 services interviewed in this study, 147 (61%) stated that they employed individuals with immigrant backgrounds among their staff. A quarter of those that did (16% of the total) had only one member of staff with an immigrant background. Eighty on the availability of service usage data by all patients. In Germany and Austria, only in emergency departments usability fi gures had been based on data (for the other services, registers had been based on estimates), in Denmark, no data-based fi gures were available for any of the nine primary care services interviewed, and Finland only kept such data in the mental health services.
The order for the availability of data about service usage among immigrant patients differed from that for all patients. One in four of the mental health services (12 services) registered data on service usage by immigrant patients separately. This was the case for 19% of emergency departments (9 services) and only 10% of primary care services (14 services).
Only 15% of the fi gures on immigrant patients were based on actual data, while 69% were based on estimates and 16% of professionals interviewed did not provide any general fi gures on the number of immigrant patients. There was again considerable variation in the availability of data registers across countries. The highest numbers of services collecting data on immigrant patients were in Spain (9 out of 15 services interviewed) and Sweden (6 out of 15 services interviewed), whereas in Austria, Belgium, Denmark, France, Germany and the Netherlands, none of the services interviewed collected data on service use by immigrant patients.
From a country-level perspective: eight countries had at least one service interviewed that collected data on service use by refugees, while seven countries reported as least one service with data on asylum seekers, six countries for victims of human traffi cking and 11 countries for irregular immigrants.

Information on the use of services
Despite the relative sizes of these services, the highest proportion of immigrant patients was found in the mental health services (23%), followed by 16% for primary care services and 13% for emergency departments. There was no overall pattern across the participating countries. Taking all the services into account, the average reported number of immigrant patients was higher in Austria, the Netherlands and Sweden. For emergency departments, this was the case again for Austria and the Netherlands, as well as for Greece. In primary care services, the number of immigrant patients was especially high in France, Germany and the Netherlands. For mental health services immigrant patient numbers were high for Greece, the Netherlands, and Sweden compared with the other participating countries.

Interpreting services
The dataset distinguished between three different types of interpreting services: 1) direct/face-to-face interpreting services, 2) telephone interpreting services, and 3) a mix of other kinds Table 2 Responses given for how often direct and telephone interpreting services were used for patients with a language barrier.

Type of interpreting service
Never used Sometimes used Always used Direct interpreting service (n=239) 53% (128) 23% (54) 24% (57) Telephone interpreting service (n=211) 59% (124) 24% (50) 17% (37) ephone interpretation, it is surprising that telephone services are not more widely used. There was also considerable variation between countries in the use of the different types of interpreting services (direct, telephone, and mixed). Mental health services tended to provide more direct interpreting services, than did primary care services and emergency departments. These fi ndings refl ect the structure of the services and the methods they apply to diagnose and treat patients. Longer treatment periods with longer contact sessions with single patients, and a lower case-load affords greater benefi ts from using more direct interpreting services. Where communication timing is more critical, as is often the case in emergency departments, then telephone interpreting services offer more distinct advantages and are therefore more widely used. The choice of interpreting service needs to compliment the way services actually function, in order to provide a suitable and effective health care.
For the Scandinavian countries participating in the study, Spain, the UK, and the Netherlands, interpreting services tended to be more mainstreamed in their delivery for immigrant patients. This was quite different for Austria, Finland, Greece and Lithuania, where there was either no interpreting services available or only very few services. For Finland and Lithuania, this could be attributed to immigration and health service provision for immigrants being a relatively new issue for policy and practice, but this would not explain the case for Austria. While for Denmark, Finland, Sweden, Spain and the UK, there were regulations and policies concerning the costs of interpretation. This is not the case for Austria and Germany, where the costs are not covered by state funds or health insurance regulation [21]. In the Netherlands, interpreting services have been provided free by the government since 1985, but these subsidies are to be abolished from 2012 [28].
The data presented here indicate wide heterogeneity between the 16 countries and the three types of services studied in terms of the provision of services for immigrant patients. Some of the data seemed to suggest that countries with fewer individuals from immigrant backgrounds tended to have fewer immigrant staff members employed within their services. Also to test this hypothesis further, research is required to compare these data with data on the different backgrounds of immigrant groups served within the catchment areas of these services.

Comparison with literature
In keeping with previous studies, our data confi rms that very little data is available on the use of health services by immigrant patients. Unlike the study conducted by Nielsen et al. [29], which surveyed national statistics agencies and relevant national health authorities, the EUGATE study was based on a direct assessment of service provision. These two approaches for gathering information on the use of health services by immigrant patients yielded similar results. But Nielsen et al. found registry data on the utilization of health services at least in 11 European countries (Austria, Belgium, Denmark, Finland, Greece, Italy, Luxembourg, the Netherlands, Poland, Slovenia and Sweden). Austria, Belgium, Denmark and the Netherlands were included in the EUGATE study, but no registry data were available at the service provider level for these four countries. eight services (37%) reported having none, while data from fi ve services (2%) were not available to report. In summary, for the 126 (54%) out of the 235 providers for which data were available, either no staff with an immigrant background were employed in the service or only one member of staff with an immigrant background.
In Sweden, all services reported having some immigrant members of staff, followed by Belgium, the Netherlands and the UK (14 out of 15 services interviewed), whereas in Lithuania none of the services reported having immigrants among their staff members. These differences did not appear to be related to variations between countries in terms of service size. The percentage of services that employed immigrants among their staff members was considerably higher within mental health services and emergency departments, than in primary care services (see Table 3 for fi gures by service type).

Main fi ndings
This study highlights the diffi culty in investigating service use by immigrant patients, due to limited availability of data on service use for all patients in most of the participating countries, and across the three different types of health service investigated. In nearly half of the participating European countries, none of the services studied had any data available, and the availability of data registers for immigrant patients was much smaller. Only 15% of the services interviewed held any fi gures for service use by immigrant patients that were based on actual data. In most cases, the data were unavailable or based purely on estimates. These fi ndings could be attributed to lower immigration rates for some European counties, or the view that collecting such data on immigrant patients might be regarded as discriminating or is not considered as necessary for planning service provision. Alternatively, some services and countries may consider current service provision as suitable for meeting the needs of immigrant patients, or that the needs of immigrant patients do not differ greatly from those of nationals.
In terms of specifi c provisions, the availability and use of interpreting services seemed generally low [5], particularly in primary care where only a few services reported using either direct interpreting, or telephone interpreting services. More than half of the assessed services did not provide any direct interpreting service. Telephone interpreting services were even less often available than direct interpreting services. Considering the lower cost and the relative convenience of organising tel- European countries. An obvious next step would be to relate national data on the use of health services by immigrants, to percentages of immigrants located in areas served by these services.
Although interpreting services were regarded as one of the most important elements of appropriate health care delivery for many immigrant patients in the literature [7,8], several European countries have yet to take steps to implement such services. Provisions for covering the costs of interpreting services are rare; the absence of such provisions in countries like Germany is one factor contributing towards the low availability of interpreting services [21]. By contrast, in Sweden where the costs of interpretation are covered by the government [9], there is a high level of interpreting service availability. Such variations are likely to have an impact on both the utilization of services and their quality in terms of patient satisfaction. Due to recent cut backs in the fi nancing of interpreting services, as had been the case in the Netherlands, the development of such services across European countries might be on the decline.
The discussion on the employment of staff with immigrant backgrounds needs to be informed by the extent to which immigrants use certain health services, and whether the employment of staff from immigrant backgrounds actually increases integration, sharing of cultural practices, and acceptance of immigrants into mainstream services and society. Also, previous papers from the EUGATE project have suggested advantages for both the inclusion of qualifi ed staff from immigrant backgrounds, and the training of all health service staff on cultural sensitivity and the specifi c needs of immigrant patients [6].
Summarising the data presented in this paper, there is still a lack of general good quality data within health services on the use and service provisions for immigrant patients across European countries. Notwithstanding, some differences do exist between countries and different types of health service. The fi ndings presented here do increase the knowledge base somewhat for services providing health care in areas densely populated with immigrants, particularly on the availability and use of interpreting services and employment of staff with immigrant backgrounds.

Confl ict of interest statement
None.
Studies in some European countries have shown that immigrants tend to make more use of emergency departments than national born populations do [3,30,34]. Secondary or tertiary services, such as mental health services [24,26], tend to be used to a lesser extent by immigrant patients. For this study, data records were either too fragmented or totally missing in the majority of cases for previous fi ndings to be compared with our dataset.
Our study was able to suggest that telephone interpreting services were more frequently used in emergency departments, a fi nding confi rmed in previous studies. Leman [25] pointed out that telephone interpreting services were more appropriate for emergency departments for their immediate availability and 24 hour coverage. He stated that there has been an enormous effort in the UK to improve telephone interpreting services, especially for emergency departments [25]. Our data extends this fi nding somewhat to other European countries. However, several countries did not provide any form of telephone interpreting services in any of the three types of health services included in this study. The impact of the missing payment regulations in some of the participating countries might have been one of the reasons for the lack of interpreting services in some of the services assessed [5,16].

Strengths and Limitations
This study had a wide scope, including service providers from three different types of services in 16 European countries. The collaboration of partners from different European countries, and research centres, was regarded as an advantage when collecting data from local health services. However, the limited availability of suitable data in these services restricted the analyses that could be conducted on this dataset. Collecting data simultaneously in multiple countries also had its limitations, especially when local factors, such as the defi nition of an immigrant, differ from country to country [17]. Considerable design and planning was required to produce consistent assessment tools, agreement on defi nitions of the several immigrant groups, the types of services studied, as well as providing training of interviewers across countries to ensure consistent data collection for comparative analyses. Despite the efforts to ensure consistency of study procedures, a degree of variation should be taken into account in line with national variations on local policy and health service practice.
Furthermore, only a small fraction of the total number of service providers in each country was included in this study. By singling out services situated in the largest cities, in areas with the highest proportion of immigrants among the population, we encountered the services with the most challenges in the provision of health services to immigrants. The fi ndings presented here as a consequence, would most probably misrepresent the situation in rural services and services in smaller urban areas.

Implications
A better knowledge base and further scoping studies are required on the use of health services by immigrant patients for valid conclusions to be drawn at the national level in most