Multimorbidity and chronic diseases among undocumented migrants: evidence to contradict the myths

This study, which compares the prevalence of chronic diseases and multimorbidity among Spanish nationals, DMs, and UMs using data from a health system in which all 3 groups were equally and universally covered, includes one of the largest sample sizes of UMs studied to date. The prevalence of chronic diseases was lower in UMs than in both DMs and Spanish nationals, and in DMs than in Spanish nationals. The prevalence of multimorbidity was lower in UMs than in DMs and in Spanish nationals, and increased with length of stay only in the DM group. These findings do not support claims of a higher morbidity burden in UMs and of health tourism as a driver of migration [6].

The main strength of this study is its large scale: all UMs for whom real-world data from EHR were available were included in the study. Other strengths include the region-wide coverage, the inclusion of all registered immigrants, and the lack of selection bias. The data analyzed are derived from the EpiChron Cohort, and have been analyzed in several similar studies in recent years [10, 17, 19, 22]. At the time of the study, UMs had unrestricted access to the public health system, regardless of their legal status. The measurement of individual-level morbidity burden using an internationally validated tool and data from EHR ensures a broad and reliable assessment of the health status of the migrant population [23]. Our study was based on diagnoses made by physicians, avoiding self-reporting bias. Importantly, the global patterns and the prevalence of the most common chronic diseases were similar to those reported in previous studies [22, 24, 25]. To reduce potential misclassification of diagnoses by physicians, we used the EDCs created by the ACG system and selected those chronic diseases included in the list of Salisbury et al. [20], in accordance with previous multimorbidity studies [25]. Finally, the use of administrative data allowed us to study the effect of important socio-demographic factors, including area of origin and length of stay in the host country.

Several limitations of our study should be noted. We did not consider socio-economic variables such as income or education level. This personal information is not recorded in Spanish health care databases, and could not be obtained in any other way while preserving anonymity. Inclusion of these parameters in our analysis could have helped to account for some of the complex factors that condition the use of health care services, such as income, educational level, unemployment, housing, social class or working conditions [7, 26, 27]. Another limitation of our study relates to our inability to refute the “salmon bias” hypothesis or “unhealthy remigration effect”. This hypothesis proposes that severely ill migrants tend to return to their country of origin to be cared for by their families. Consequently, multimorbidity rates among immigrants may be underestimated. Although we cannot reject this possibility based on our findings, a recently published study appears to rule out this hypothesis as the main explanation for better health outcomes in migrant versus native populations [28]. Finally, although all immigrants have the right to request a health card regardless of their legal status (provided that they are registered in the local population census), it is possible that some may not apply for a health card unless they actually become ill. This would result in an overestimation of the morbidity burden in UMs compared with DMs and the native population, since healthy UMs would not be included in the administrative databases. Regardless, this bias would likely result in underestimation of the differences in the prevalence of chronic diseases and multimorbidity reported in the present study. A final limitation is the underrepresentation in the data source used of some subgroups of the migrant population, such as those born in Asia.

To the best of our knowledge, only one study has previously assessed the prevalence of multimorbidity in UMs. This study of UMs treated in primary care in Switzerland [29] reported a prevalence of multimorbidity of 23% in women and 14% in men, as compared with corresponding values of 12% and 8% in the present study. One possible explanation for these contrasting findings is that our analysis included all individuals with a health card, regardless of whether they availed of public health care. By contrast, one of the inclusion criteria in the Swiss study was having attended a medical consultation during the year of the study. Other possible reasons for these divergent results are the older mean age of the population in the Swiss study (42 years versus 38 years in the present study), differences between the migrant groups studied, and differences in the diagnostic criteria applied between one country and another. However, the main limitation of the Swiss study was the lack of any comparison with Swiss nationals. A study of Spanish workers [30] reported a higher prevalence of poor health in UMs that had been in the country for more than 3 years than in Spanish nationals, but reported no differences between UMs according to length of stay. A Dutch study [31] of primary care consultations reported a much higher prevalence of diabetes in UMs than that found in our study (7.4% and 1.1%, respectively). The remaining articles published on the subject mainly consist of analyses of cases recorded in NGO clinics, and therefore suffer from significant selection biases. We have found no studies that have analyzed the health of UMs in the context of health systems that afford equal access to UMs, DMs, and nationals.

The lower prevalence of chronic diseases in UMs than DMs is an important finding. It is possible that among UMs, health capital is an even more important determinant of the ability to migrate. Without work or residence permits, UMs are obliged to work in the submerged economy, without access to many of the benefits offered by the welfare state. This type of migrant therefore fits well with the “healthy migrant” paradigm. Another possibility is that, even when sick, UMs do not engage with the public health system for fear of being identified and deported, even though professionals working in the Spanish health system are not obliged to provide police with any information about the legal status of their patients. Even in countries where UMs are fully entitled to care, informal barriers such as language and communication problems, transport problems, poor knowledge of the health care system, the lack of a social network, and fear of deportation can undermine accessibility [13]. This is in line with the Tudor Hart inverse care law: “the availability of good medical care tends to vary inversely with the need for it in the population served” [26].

We found that the prevalence of multimorbidity among UMs was not influenced by length of stay in the host country. This is not the case within the general migrant population [19]. The irregular situation of this population may prevent them from engaging with the public health system, potentially resulting in under-registration and inadequate follow-up of medical conditions. However, a 2002 Spanish study of 380 Ecuadorian immigrants reported no differences between UMs and DMs in terms of health system access [32]. The fact that length of stay does affect the burden of morbidity among DMs is consistent with the results of a recent study that also analyzed data from the EpiChron cohort, and found that mortality (a variable for which there is no under-registration in Spain) was very low in immigrants with a length of stay < 5 years but tended to increase with a longer length of stay [10].