Background

The number of international migrants continues to grow each year. According to the United Nations Migration Report, the number of migrants has reached 244 million in 2015 up from 191 million in 2005, representing an increase of 28% over the decade in comparison with an increase of 13% during the period 1990–2000 [1, 2].

Between 2000 and 2015, Europe has absorbed the second largest number of international migrants following Asia [1, 3]. Despite the global economic crisis which started in 2007–2008, Europe and Northern America have recorded an annual growth rate in the international migrant stock of 2% per year [1].

These transformations have both quantitative (i.e. an increasing number of migrants) and qualitative (i.e. evolving reasons for migration) aspects. There is a trend towards permanent migration and reunification of families with immigrant setting in the host country in a more definitive way [4]. And most recently, we have seen an increasing number of asylum seekers and refugees, which is reaching the highest levels seen since World War II [1].

This situation has generated various responses in the host countries, as immigration is acquiring a significant social and political dimension. Immigration is influencing public opinion and triggering a debate, often improperly informed, regarding the pressure on public services—including health services [3]. This has even led to the adoption of new legislation [5–7] limiting access to health care for migrants, that may pose, as a result, a risk to public health.

The dramatic changes in demographics, socio-economics and politics require an update of the analysis of health service utilization by immigrants in order to properly determine the breadth and scope of the current situation. Consequently, research on migrant access and utilization of health services has proliferated in recent decades [8, 9]. Results from a previous review point to a lower utilization rate of general and specialist medical services by immigrants compared to native-born populations [10]. However, and since patterns of healthcare utilization depend on factors that may have evolved in recent years, such as age, sex, socio-economic level, time of stay in the host country or origin of the immigrants, and the specific features of healthcare services of the host countries, it seems necessary to revisit the state of knowledge on this subject.

The objective of this study is to describe the available scientific evidence that has investigated the differences in healthcare service utilization between immigrant and native populations in the last 3 years (June 2013 through February 2016), and to explore the possible effect on the differential use of variables associated with health needs, socio-economic status or other factors.

Methods

A systematic literature review was performed to identity the available empirical evidence comparing immigrant’s healthcare utilization with native populations using a predefined protocol [10]. Inclusion criteria for articles to be considered were original studies with quantitative data that compared the use of healthcare services between native and immigrant populations. Service use was defined as the interaction between health professionals and patients [11]. Only studies with both population groups properly defined, i.e. immigrant and native, were included. For the purposes of this review, we used the European Union definition of immigrant status based on foreign country of birth including up to the second generation [12].

Papers that considered undocumented immigrants, asylum seekers and/or refugees were also included. The indigenous majority population served as the native reference group. No limitation in gender or ethnic characteristics was stipulated.

Articles were excluded if they (1) exclusively evaluated healthcare utilization for children or adolescents younger than 18 years of age, (2) were editorials, letters or reviews and (3) were qualitative studies.

Search strategy and study selection

Two strategies were utilized in the search for relevant articles on this review.

Firstly, in February 2016, a librarian conducted a systematic review of the electronic database MEDLINE (PubMed) in search of the literature published between June 2013 and February 2016. No language restrictions were applied; no authors were contacted for additional information. MeSH terms and key words used, as well as search strategies performed, are shown in Table 1.

Table 1 Search strategy for healthcare service utilization’s comparative studies
Full size table

The initial screening of the articles was based on abstracts. Two researchers reviewed all abstracts independently. Selection of relevant articles was based on the information obtained from the abstracts and was agreed upon in discussion. If the abstract was not available, the full text was examined. In the case of discrepancies between the two researchers, the original paper was obtained and an agreement was achieved after it was read.

Secondly, a researcher (AIHG) conducted a manual search of grey literature through Google Scholar, including published papers from 2013 through February 2016 taking into account the terms (Health care use; Comparison; Immigrants; Natives) and (Needs, demands and barriers; Coverage; Primary care; Emergency services; Utilization patterns; Native; Foreign; Autochthonous; Immigrant). Both English and Spanish web pages were included in the search results. Appropriateness for inclusion was based on titles; in the event of doubt, abstracts were retrieved. Studies without electronic abstracts were not included.

Subsequently, two researchers examined the full text of all papers that satisfied the inclusion criteria (AIHG, ASS).

Data extraction

The following information were extracted from each publication: context of the study (country and year), characteristics of the included population (definition of native and immigrants groups, sample size for each group), methodological components (design of the study, statistical analysis, source of information), area of healthcare services assessed, confounders affecting healthcare utilization (individual determinants, measures of need, socio-economic indicators, cultural factors), objective of the study and main results.

Results

Characteristics of the studies

Thirty-six papers met the inclusion criteria in this study. The process followed to include those papers is shown in Fig. 1. Table 2 shows the information extracted from the included publications. Of the 36 studies included, 8 were duplicated in both the manual and electronic search [13–20], 12 were included after the manual search [21–32] and 16 through the electronic search [33–48]. Among them, at least 9 partly describe the same dataset [13–16, 19, 20, 25, 47, 48]. Nevertheless, as these articles focused on different aspects of healthcare use or outcome measures, all were included in this review.

Fig. 1
figure 1

Study flowchart for the selection process of the final included studies

Full size image
Table 2 Descriptive summary of the studies included in the review
Full size table

Distribution of studies regarding publication year was as follows: 8 studies published in 2013 [17, 22–24, 27, 28, 41, 42], 15 in 2014 [14–16, 19, 21, 30, 32, 33, 35, 36, 38, 40, 43, 44, 47], 10 in 2015 [13, 18, 25, 26, 29, 31, 34, 45, 46, 48] and 3 in 2016 [20, 37, 39]. The majority of the publications analysed data from European countries (28; 78%), both North and Central (12) (Norway [13–15, 19, 20], Denmark [45], Sweden [35], the Netherlands [17, 32, 34, 40] and Austria [41]) and South Europe (15) (France [22, 36], Italy [18, 24, 29, 37, 43], Spain [23, 27, 28, 31, 38, 39, 46] and Portugal [33]) and 1 from the UK [26]. Seven papers (19%) explored this issue in North America (2 from USA [30, 34] and 5 from Canada [16, 21, 25, 47, 48]); and 1 (3%) in Asia (Singapore) [42] (see Fig. 2).

Fig. 2
figure 2

Distribution of studies according to country of destination

Full size image

Geographical coverage of the studies has some variation: 21 performed at the national level [13–15, 17, 19–22, 28, 30, 32, 34–36, 38, 40, 41, 45–48], 10 at a regional level [16, 18, 23, 25–27, 29, 31, 37, 44], 3 at a local level [28, 33, 42] and 1 multi-country study [39] with data from a regional level of 1 country and the national level of the other. There were only 4 longitudinal studies (2 prospective [18, 42] and 2 retrospective [27, 43]) and 1 case-control study [35]. Sample sizes ranged from 74 [35] to 7,856,348 [43]. Multivariable regression (Poisson or logistic) was the most frequent analysis. Only 9 studies conducted univariate analysis [29, 32, 33, 35, 38, 43, 48].

Sources of information

Service utilization could be assessed from two perspectives: the physician’s perspective, based on recorded databases and volume of medical services, and the patient’s perspective, based on patient-reported use of services through healthcare surveys [49].

The largest number of papers (18) used information from administrative [13–16, 18–20, 23, 25, 29, 33, 35, 37, 39, 43] or insurance system databases [32, 34] and specific hospital registries [28] as source of information. Among the 16 papers (44.4%) that analysed healthcare surveys, where people report their individual healthcare use, 14 studies used population-based surveys which were elaborated for other purposes [17, 21, 22, 24, 26, 30, 36, 38, 40, 44, 46–48] while 3 of the surveys were specifically designed to explore immigrants healthcare use [31, 41, 42]. Only 2 studies [33, 45] (5.6%) combined health survey and administrative information and 1 study also used a national survey for general practitioners (GPs) [17].

Subjects

There were diverse definitions of immigrants. Country of birth was the most common criteria used to define immigrants (18), or country of birth of the subject and their parents (10). In addition, name recognition (2) [32, 34], citizenship (3) [18, 24, 28] or a combination of citizenship and country of birth (3) [30, 42, 45] were also used.

The majority of papers classified the immigrant population in sub-groups usually based on country of birth (13). However, some studies considered geographic area of origin (8) or World Bank categories of income level (5). Other less frequent categories considered were legal status (3), reason of migration (1), length of stay in the country (3) and being first of second generation (1). Only 2 studies (5.6%) [18, 22] compared the use of services considering the immigrant populations as a whole, without defining specific sub-groups in those populations.

Findings

The outcome “healthcare service utilization” could be organized in seven focus areas: primary care, specialist’s services, hospitalizations, emergency services, mental health, dental care and medication prescription. Some studies reported on more than one outcome. In total, 8 papers analysed the use of primary care (including GP visits, dental care and physiotherapy) [13–15, 21, 27, 36, 44, 48], 6 evaluated the use of specialist services (including hospitalizations or emergency care) [23, 28, 30, 33, 35, 42], 5 assessed mental health services [17, 18, 20, 29, 45], 10 evaluated the use of both primary care and specialists [22, 24, 31, 32, 34, 37, 38, 43, 46, 47], 2 evaluated primary care and mental health [19, 40], 4 evaluated both primary care, mental health and hospitalizations [16, 25, 26, 41] and 1 studied pharmaceutical use and prescriptions [39]. In addition, 6 studies also reported medication consumption [20, 31, 32, 37, 42, 43].

The measurement of healthcare utilization was either continuous (number of contacts) or dichotomic (having had any contact). The period of time used to determine utilization ranged from 4 weeks through 1 year.

The more frequent outcome was that immigrants have lower [17–20, 22, 25, 27, 28, 30, 33, 35, 40, 43, 44, 48] or similar [13, 21, 34, 36, 41, 42] healthcare utilization. However, studies that included analysis by sub-groups of immigrants identified some differences across groups [14–16, 23, 26, 31, 37, 39, 40, 45, 46] as well as with the type of service assessed [14, 24, 29, 31, 32, 38, 40, 46, 47].

The immigrant population showed a similar [23, 24, 29, 31, 32, 34, 36–40, 46] or lower [17, 18, 22, 27, 28, 33, 43] use of primary care and specialized care in countries with universal access to health care—even for undocumented migrants [50]. This finding was consistent regardless of the source of information used. In other countries, some differences were identified associated with the source of information: immigrants showed higher use of health services when estimates were based on surveys [26, 41, 45], while their rates were lower [19, 20, 35] or similar [13–15] when registries or administrative data were used.

Discussion

The main result of this review is that migrant populations appear to have a lower use of health services than native populations, with a similar level of use of primary care services. This result appears to be independent from differences in need of access. Nevertheless, the great heterogeneity of the studies included in this review, considering both the sources of information, as well as factors used for controlling health need and to classify immigrants in sub-groups, requires caution when making an overall estimation valid for all immigrants.

Different sources of heterogeneity should be mentioned. First, and probably the factor with the highest relevance, was the definition of immigrant and their characterization. This review has identified several factors that could be involved with differences in healthcare utilization among immigrants: income of the original native countries [13–15, 28, 38], the specific reasons motivating migration [15, 16, 19, 25, 26], fluency in the host country language [16, 17, 21, 25, 44, 45, 47] and length of time of stay [13, 15, 19–21, 26, 38, 45, 47, 48].

There were also differences in how medical need was determined and how to estimate factors that predispose to healthcare use. The majority of studies assessed health needs from the point of view of self-perceived health, and through commonly used socio-demographic variables, such as education, income or working status, following the model of Aday and Anderson [51, 52]. Multivariable models were adjusted by these variables to eliminate the effect they could have on utilization, but whether they had a differential influence on immigrants or native populations remains inconclusive.

Variables which could have a significant effect on healthcare service use and in particular for mental health care [53], such as health beliefs and cultural concepts on the part of the immigrants, fear of stigmatization, taboos, perceived efficacy of health interventions or use of alternative services, were usually not considered. The effect of these variables is most commonly explored through qualitative techniques, and papers that used those methods were not included in this report.

Variation in countries’ healthcare systems limits direct cross-country comparisons, although immigrants showed similar patterns of utilization in countries with significant differences in their healthcare services. Nevertheless, studies reviewed pay little attention to the structural and organizational dimensions of healthcare systems, other than reporting the specific conditions for accessing health services. One paper explored the influence of attitudes of professionals regarding immigrants [54], 2 studies assessed the reasons for unmet healthcare need [31, 38] while 2 underscored the patient workload of healthcare professionals [22, 23]. In addition, the effect that new legislation enacted in different countries could have had on access to healthcare services by immigrants has not yet been evaluated and published and therefore cannot be assessed in this review.

Attempting to expanding the scope of previous reviews, we tried not to constrain the inclusion criteria regarding areas of healthcare services assessed [10, 55, 56], context of the study (country) [54, 55], or characteristics of immigrants [54, 55].

This work adds also new information regarding the use of mental health services, both in terms of primary [19, 26] and specialized mental services [16–18, 20, 25, 29, 41, 45]. Nevertheless, and although immigrants have shown a higher susceptibility to emotional and mental health problems that could be linked to the stressors of adapting to the host country [57], those studies reported similar findings as for other health services: an overall lower use by migrants, also with differences across sub-groups and with an occasional higher use of emergency care.

This review also provides the opportunity to have an insight of the healthcare use of certain vulnerable sub-groups, as the handicapped [13], the elderly [13, 15, 32, 37] or patients with chronic conditions [21, 34, 36], but the pattern of use of those sub-groups is similar to that of the general population, even when immigrants seem to have less health problems than natives [13, 34], or a poorer health status [36]. Immigrants also showed a higher use associated with longer periods of stay in the host countries [15, 21] as well as significant differences of use among migrant sub-groups [32, 37].

The effect of gender differences was assessed most notably in papers evaluating the use of mental health services [16, 19, 20, 25, 41, 45]. Nevertheless, no conclusive evidence could be established: compared to their native counterparts, Straiton et al [19, 20] and Durbin et al [16, 25] found a lower use of mental health services for immigrant women, while Kerkenaar et al [41] and Smith-Nielsen et al. [45] found a higher use.

The possibility to analyse the use of different levels of care may help to determine the existence of gaps in utilization (less use in one area could explain an increased use in another area) or highlight the existence of different referral criteria (primary care specialists) [23]. De Luca et al. found [24] an over-utilization of emergency services associated with an under-utilization of preventive care services among the immigrant population. Tormo et al. [31] and Díaz et al. [14] obtained similar results, although they concluded that the higher use of emergency services did not compensate the lower use of GPs. The identification of differences in pharmaceutical consumption could also lead to identify particular health problems or economic barriers accentuated by the development of restrictive health policies.

Lastly, the large number of European studies, particularly from western and central Europe, has to be highlighted, probably depicting the interest about the migratory pressure these countries have faced in the last years—migration from Eastern Europe after the fall of the Iron Curtain; from Latin America, North and sub-Saharan Africa; from internal migration flows south-north after the economic crisis; or most recently, the refugee crisis.

Study limitations

The literature search was conducted only in one database (MEDLINE), although the electronic search was manually completed using Google Scholar. There were implied limitations in the manual search, since it was not systematized and was susceptible to errors as it relied on title appropriateness (particularly for articles with ambiguous titles). Furthermore, no backward citation of the papers included in the systematic review was performed. Additionally, the systematic search only identified 50% of the papers accepted for inclusion, which raises some doubts regarding the intrinsic limitations of the system to classify and assign terms to papers that compare the use of healthcare services between native and migrants.

Finally, qualitative papers that explored the use of healthcare services were not included, as it would be difficult to draw comparisons from these studies.

Conclusions

Overall, and regardless of the changes in the immigration process, data here analysed is coincident with results obtained in previous reviews [10, 54, 56], confirming that immigrants show a general tendency to a lower use of health services than native populations. But these data also indicate the existence of differences within the immigrant populations, reinforcing the conclusion that further studies intended to compare the rate of healthcare use between native and immigrant populations should incorporate information that allows for better identification and characterization of the immigrant population. The immigrant population cannot be considered as a uniform whole. Their diversity has to be taking into account when describing and analysing their healthcare utilization. This will also require improvement and standardization of the information collected [55, 58].

In this sense, the limitations of health surveys have to be emphasized. Surveys are not just subjected to memory bias, but they are less suited to be representative of all relevant sub-groups of the immigrant population, as their samples usually do not include enough participants to reflect the wide variability of the diverse immigrant population to estimate their differential use. For instance, only one paper includes immigrants in irregular status [44]. Therefore, the use of data that overcome these limitations has to be encouraged. Further studies should be based on other information, such as registers, administrative or insurance data, or data from non-governmental organizations [59].