The definition of immigrant status matters: impact of nationality, country of origin, and length of stay in host country on mortality estimates

Our study reveals three important findings. First, defining immigrant status based on nationality versus place of birth results in different estimations of mortality. Second, mortality estimates of the immigrant population in Spain are lower than those of the native population, particularly on arrival, regardless of the definition used. Last, mortality increases with the length of stay, reaching a level comparable to that of the native population after 10 years.

Our findings show that the definition of immigrant status (nationality versus place of birth) can lead to significant differences in mortality estimates. While place of birth is the criterion most frequently used [15, 19,20,21,22,23,24], immigrant status has been defined based on nationality in studies conducted in Germany [25, 26], France [27], Italy [28], Denmark [29], and Greece [30]. Several studies performed in the USA [31,32,33,34] have used the concept of ethnicity, which to date is less common in European studies. Finally, some researchers also take into account the immigration status of parents [16, 18, 35,36,37].

As our results indicate, the group for which the greatest disparities are observed consists of individuals whose nationality corresponds to their host country but who were born elsewhere. While multiple factors can contribute to this discordance, the following two scenarios are the most important:

Children born abroad with Spanish parents living abroad or mixed-nationality parents with at least one of them born in Spain. These individuals are usually not considered immigrants despite being born outside Spain, and would constitute “false positives” if considered as such.

Immigrants who acquire the Spanish nationality by naturalization. Failing to consider these individuals as immigrants could result in the introduction of “false negatives” into research studies. Naturalization policies are very diverse and influenced by historical, geographical, legal, and/or social factors. For example, while a non-national must, in general, reside in Spain for at least 10 years before applying for citizenship, those of Ibero-American, Andorran, Philippine, Equatorial Guinean, Portuguese, or Sephardic origin can apply after only two years.

As shown in Fig. 4, the lower death proportion observed for non-Spanish nationals born outside of Spain compared with Spanish nationals born outside of Spain appears to be due, to a large extent, to the longer duration of stay in the host country in the latter group. We therefore consider it essential that research studies take into account information relating to length of stay, in addition to nationality and place of birth, in order to avoid misinterpretation of results [12].

Our results showing lower mortality in the immigrant population are in line with those of previous studies conducted in Europe [3], Norway [16, 35], Belgium [18], Spain [22], Italy [28], England [23], Canada [15, 19, 20], USA [31, 33, 34], France [24], and Denmark [29], although some studies, including one conducted in Greece [30], have reported higher mortality rates in immigrant populations.

In general, the existence of a “healthy migration effect” is accepted as a justification for the better health outcomes seen in immigrant populations. However, some morbidity studies have highlighted the possibility of a registration bias during the first years of stay of immigrants. It is proposed that, even when ill, immigrants may not avail of healthcare services due to ignorance, work priorities, language problems, or other accessibility problems [1, 7, 38]. This hypothesis is not supported by our study, since the mortality registry in Spain is exhaustive and includes data even of individuals who do not avail of healthcare services. We cannot rule out the possible existence of “salmon bias”, whereby sick migrants return to their countries of origin awaiting the outcome of disease. However, recent studies on mortality in immigrant populations have not supported this theory [23, 39].

Our results also show that mortality estimates in immigrants from Western Europe and North America are similar to those of the native Spanish population. While most immigrants from countries outside the European Union or North America come to Spain for economic reasons and end up working in unskilled jobs, those who emigrate from countries with a level of income equal to or greater than that of Spain probably work in more skilled jobs, in which health status is less important [40]. However, in the absence of socioeconomic data we cannot confirm this hypothesis.

The influence of length of stay in the host country on mortality has been less studied. Several studies conducted in Norway [16], Belgium [18], and Canada [15] have shown an increase in mortality with increasing length of stay. This finding has been related to the often poor socioeconomic conditions in which immigrants live [6, 17, 34], the process of acculturation [16, 18, 34], and allostatic overload [5, 38]. Our results suggest a clear temporal relationship, with length of stay in Spain constituting a “risk factor” for the health of immigrants. To some extent, immigrants “invest” their good health in improving the health of their families via the remittances they send to their countries of origin.

The strengths of this study include region-wide coverage, the inclusion of all registered immigrants and the absence of selection bias, as well as the grouping of immigrants according to area of origin, which to a certain degree recognizes the heterogeneity of the immigrant population. The use of administrative data allowed us to study the effects of important sociodemographic characteristics such as region of origin and length of stay in the host country. The selection of the outcome variable “mortality” minimizes bias associated with registration problems, since the data corresponding to this variable are obtained from a database with broad coverage and good reliability.

Some limitations to the approach used should be noted. Our study did not consider socioeconomic variables such as income, education level, legal status, or reason for migration, which could have helped identify some of the complex factors that influence mortality. This important personal information is not stored in Spanish healthcare databases and could not be obtained in any other way while preserving anonymity. Another weakness relates to our inability to take into account the “salmon bias” or “unhealthy remigration effect”.