Main results
Among 50 years old or older persons living in Norway, our study reveals that a higher proportion of Norwegians use PHC services compared to immigrants. However, important differences in utilization between immigrants depending on their countries of origin are observed. While HIC immigrants used the PHC system less than Norwegians, they had similar number of diagnoses when in contact with primary care, especially the oldest ones. Among OIC immigrants, 50 to 65 years old used both the GP and the EPC more often and had higher comorbidity levels, but this pattern was reversed for older adults. For all immigrants, utilization of PHC services increased with longer stay in the host country and was higher for refugees but lower for labour immigrants compared to immigrants for family reunification. Adjustment for socioeconomic factors reduced, however, the differences between immigrants and Norwegians.
Strengths and limitations
Our study has several strengths. Firstly, it covers the whole population in Norway aged 50 or more, eliminating self-selection bias. Secondly, the rich information on socioeconomic position and migration status allowed us to categorise immigrants into relevant groups both based on country of origin, reason for migration and length of stay in Norway. Thirdly, the data on utilization of PHC services is objective and complete as opposed to other studies relying on self-reported utilization of services.
However, our study has some limitations. Firstly, our information is only about PHC. Even though the main rule in Norway is that the GP is the only gatekeeper in the system, we do not have information about patients already using specialised care, including home consultations from secondary care to older patients. Nevertheless, PHC is by large the first point of contact, and a necessary step to secondary care also for emergencies and the share of patients attended by secondary care has earlier been reported to be similar for Norwegians and immigrants in Norway [
29], whose comparison is the main aim of this study. Secondly, the diagnoses were based on ICPC-2 codes reported for administrative claims and not, for example, extracted from electronic records. This was the only possibility to obtain data on morbidity for the whole population in Norway, but it reduces the accuracy of the diagnoses, and therefore cannot be used to calculate real prevalences of diseases. They are, however, widely used and validated for research at the chapter level for comparison of populations [
27],[
28]. Thirdly, the HELFO-database does not include consultations for individuals living in nursing homes. This will explain part of the reduction in utilization of PHC services for the oldest patients regardless of country of origin. The proportion of people living at home decreases from 96% among those 67-79 years old, to 82% and 56% for those 80 to 89 and older than 90 years old respectively [
2]. Unfortunately, we have no data on the proportion of immigrants that live in nursing homes, but there are indications that immigrants might be reluctant compared to Norwegians to live away from their own homes, which would increase the differences we find among the oldest patients. Lastly, our study only takes into account consultations in Norway. Patients from other countries, especially those located near to Norway, may have travelled abroad to see a physician, and this might contribute to the differences we found between immigrants and Norwegians. Similarly, remigration to the country of origin, if not registered, might give lower utilization rates among immigrants. Recent analyses for immigrants from Asia, Africa and Latin America in Norway, however, indicate that a low proportion among the elderly move back to their countries of origin [
2].
Interpretation of results
Because of its cross-sectional nature, our study cannot provide explanations for the different utilization of PHC services among immigrants. Lower utilization would be appropriate if it reflects better health, but it could also be attributed to barriers for utilization of health services, posing a public health challenge. Identified potential barriers in the literature at patient, provider and system level include (i) patient characteristics represented by demographic and social variables, health beliefs and attitudes, personal and community enabling resources, perceived illness and personal health practices; (ii) provider characteristics such as skills and attitudes, and (iii) barriers related to the system characteristics, mainly due to the organisation of the health care system [
30],[
31]. Older age has more recently been identified as an additional potential barrier [
32]. In accordance to the largest study among elderly immigrants in Europe [
8], socioeconomic variables did not completely explain disparities in utilization of services in our study, but there were substantial differences between immigrants depending on the wealth of their country of origin. Also, the inclusion of income and education levels in our analyses contributed to decrease the differences in utilization between immigrants and Norwegians. Thus, our results point to socioeconomic barriers at the patient level including both pre-migration aspects and factors at the host country that should be further studied. However, there are probably still other reasons explaining differences in utilization between immigrants and Norwegians and among the different immigrant groups.
As explained above, lower utilization could be appropriate if it reflects better health among those not attending to PHC. Some researchers would expect lower health care utilization among immigrant populations because of the initial selection of relatively healthy persons as immigrants, known as “healthy immigrant effect” [
33]. This effect is thought to be strongest among recent immigrants and health is likely to converge toward that of the native-born population with longer time, adoption of new ethnocultural habits of the host country and increasing age [
8]. Our results support both theories. Although utilization of PHC services increased with longer stay in Norway for all immigrants, refugees used the system more often and labour immigrants, to whom the “healthy immigrant effect” would apply the most, more seldom compared to immigrants for family reunification. It is noteworthy that the effect of the reason for migration was still significant for older immigrants who had stayed in Norway an average of more than 20 years.
Several studies also suggest that morbidity burden is the most important variable at the patient level to predict utilization of services [
8],[
34], though it has not been studied if this affects all age groups in the same way. We cannot determine if non-users were healthy or not, but our study can give some indications regarding the morbidity of PHC users across groups. Overall, HIC immigrants had lower and OIC immigrants higher number of diagnoses (ICPC-2 chapters) compared to Norwegians. An earlier health survey among selected groups of immigrants in Norway, most of whom migrated from other than HIC, disclosed lower self rated health and higher disruption of daily live because of illness among immigrants 55 to 70 years old compared to Norwegians [
29]. However, in the same study the number of self-reported consultations to the GP during the last 12 months was 7.2 for immigrants and 2.6 for Norwegians, much higher, especially for immigrants, than those reflected by objective measures in this study. This gap could be due to selection bias or recall bias in the health survey, and should be further investigated.
Similarly to a previous study among elderly patients in PHC [
35], the two most common diagnoses reported in our study were cardiovascular and musculoskeletal ones. Compared to Norwegians, OIC immigrants more often had diagnoses related to endocrine, digestive and musculoskeletal problems, but less often had cardiovascular diagnoses. These results concur with a recent study of use of cardiovascular and antidiabetic drugs among immigrants in Norway (Diaz, personal communication), and with two other studies describing high prevalence of diabetes among immigrants, but underdiagnose of cardiovascular disease in this group [
36],[
37]. However, the higher comorbidity levels for younger OIC immigrants and the lower comorbidity levels among the oldest ones in our study are hard to explain. As all the subjects in Figure
2 are PHC users, it only includes persons that have surely been in Norway at least at one point in 2008, so that the “salmon bias” effect reflecting back migration is truly not the main explanation. The concept of “health survivor” has been mainly used for workers [
38], but it can also apply to the elderly among immigrants, who are well above the mean life expectancy age for some of the countries they migrate from [
21], especially taking into account the tendency of immigrants to become diagnosed at a younger age than Norwegians [
22],[
23]. Further research on health need among non-PHC users is thus necessary to determine the appropriateness of PHC utilization among immigrants, especially the eldest OIC immigrants.
Potential barriers at the professional and system levels can be partially explored through characteristics of PHC consultations. As explained in the methodological section, fees regarding consultations lasting more than 20 minutes (long consultations) and interpreter are mutually exclusive, but the GP gets a higher economical incentive when using an interpreter compared to using long time. Having this in mind, our results reflect that interpreters are used in one out of five GP consultations with OIC immigrants, and more seldom in consultations with HIC immigrants, who often speak either English or an Scandinavian language, both usually understandable for the GP. The lack of available interpreters at the EPC is thus probably the reason for a higher proportion of long consultations at the EPC for OIC immigrants, and can represent a barrier for quality of care that the physician tries to compensate by taking several laboratory tests, as previously described [
39]. In contrast, in consultations with immigrants, GPs tended to order similar or less ECG and blood test, except for glucose and CRP test. According to Tran et al, patients from minority groups have worse glycaemic control than Norwegians [
22], which could explain the higher use of glucose related tests. However, given that ECG should be regularly taken in patients with diabetes and cardiovascular diseases and that the prevalence of these two are probably high among several immigrant groups [
37], ECGs might be underused by GPs for this population. Last, and according to the previous literature [
17], the main relative differences between Norwegians and immigrants were regarding the proportion of home consultations, which were more than twice as often paid to Norwegians compared to other immigrants. This might constitute a system barrier for older immigrants that should be further investigated.