To our knowledge, this is the first study to report OS differences between immigrant and Danish-born patients with hematological cancers. In the present study, no overall differences in OS between immigrant patients, Western as well as non-Western, and Danish-born patients were observed after controlling for confounders. Immigrants were more often in the lowest income quartile compared to Danish-born patients, but within the lowest income quartile, non-Western immigrants had superior survival compared with Danish-born patients. An explanation for this finding may be that low-income Danish-born patients were a more selected group as low-income comprised only 23% of the Danish-born patients versus 49% of immigrant patients. In the low-income group, Danish-born patients were less educated (8% classified as high vs. 18% for non-Western), more likely to be living alone (53% vs. 38%), and more often retired (55% vs. 28%) (Supplementary Table 2). Thus, Danish-born patients in the lowest income quartile appeared to have more unfavorable socioeconomic factors possibly associated with OS. Similarly, non-Western immigrants with reduced work capacity also had superior survival compared to Danish-born patients with reduced work capacity. Reasons for reduced work capacity may also differ substantially between populations.

A significant survival advantage was observed for non-Western immigrant patients with ALL compared to Danish-born patients (Fig. 4). However, in a Cox regression analysis, comparing non-Western patients to Danish-born patients with ALL using the fully adjusted model (which included socioeconomics), the survival advantage was no longer statistically significant (adjusted HR 0.33 [0.10–1.06] for all-cause mortality compared to Danish-born patients) (Supplementary Table 1). This survival advantage should be further explored, but false random associations are a risk as the ALL analyses were based on only 18 non-Western immigrants. Non-Western immigrant patients with aggressive lymphomas had a worse survival compared to Danish-born patients, which was not found for indolent or Hodgkin lymphomas (Fig. 4). After full adjustments, the HR for non-Western patients did not remain significantly elevated (HR 1.20 [0.99–1.44]) (Supplementary Table 1).

Non-Western immigrants diagnosed in the period 2016–2020 had a reduced 5-years OS compared to Danish-born patients of around 5 percent points (Fig. 4). This was also found in the simple adjusted HR for all-cause mortality plotted over calendar year of diagnosis, where there was a trend towards a higher all-cause mortality for non-Western immigrant patients compared to Danish-born patients diagnosed in the period 2017–2020 (Fig. 3B). Both results should be interpreted with caution, as the analyses are subject to uncertainty due to progressively shorter follow-up for patients diagnosed in the later part of the inclusion period. Reasons for this trend are not possible to elucidate in this register-based study. In 2018, Danish authorities cancelled financial support for interpreters in the hospital, necessitating non-Danish speaking patients, who have lived in Denmark for more than three years, to pay interpreter bills out-of-pocket. This resulted in a 33% decrease for all interpreter services in the period 2017 to 2019 in the capital region of Denmark [32]. Without interpreter support, the fraction of non-Western immigrants with poorest language proficiency may face difficulties in navigating in the complex pathways associated with diagnosis and treatment of hematological cancers. Interestingly, when the model was fully adjusted (including socioeconomic factors), OS differences were not significant at any periods (Supplementary Fig. 2).

The present study found no clinical difference in the number of days of inpatient hospitalization in the year prior to diagnosis between immigrant patients and Danish-born patients after controlling for confounders. However, non-Western immigrants had an average of 1.3 more days of hospitalization in the year prior to diagnosis, suggesting that non-Western immigrant patients required more hospital visits before being diagnosed with a hematological cancer, which could be related to language barriers preventing precise symptom reporting and efficient diagnostic work-up. Krasnik et al. reported consistent results in a study from 2002, where durations of hospital admissions were similar between immigrants and Danish-born citizens. However, Krasnik et al. did not stratify on Western and non-Western immigrant patients, and differences in hospitalization between non-Western immigrants and Danish-born citizens may have been evident upon further stratification as seen in the present study [33]. Nielsen et al. also reported highly comparable use of free-of-charge healthcare services between immigrants and Danish-born citizens [13].

In the year following diagnosis, Western immigrant patients had relative fewer inpatient hospitalization days compared to Danish-born patients (adjusted IRR 0.95 [0.94–0.96]). Western immigrants may be healthier, as shown in previous studies [34], and more resilient to treatment complications. In contrast, non-Western patients had a clinically significantly higher rate of hospitalization (adjusted IRR 1.14 [1.13–1.16]), possibly suggesting more treatment-related complications and in line with previous studies showing a higher healthcare utilization among non-Western immigrants in Denmark [11,12,13].

These findings are relevant beyond a Danish context, as the study explores differences by immigrant status in a healthcare setting with free access to all citizens. Therefore, patients’ ability to directly finance or co-finance cancer therapy would not have had major impact on survival outcomes, which largely eliminates the risk of poorer outcomes for immigrant patients directly caused by financial hurdles and limited access to healthcare services. In the present study, possible disparities in survival would more likely be attributed to inability to navigate/use the healthcare services offered without costs, for example because of barriers related to language proficiency, culture, or systemic inequality in the quality of care provided to immigrant patients.

Research into cancer outcomes by race, ethnicity, and immigration have generally been based on American populations and with an emphasis on racial and ethnic differences. There is a distinct difference between ethnicity, race, and immigrant status, which may translate to differences in overall survival. Ethnicity refers to the cultural characteristics of a particular group, and race is based on physical attributes, while immigrant status indicates legal residency in a new country. Individuals of different races and ethnic minorities are often born in the host country and may speak the language fluently but are still often affected by systematic discrimination, lower income, and lower education [35]. However, immigrants like those included in the present study may be more likely to be challenged by lower health literacy and language barriers in addition to discrimination. Several studies from the US have compared survival of hematological cancers for subgroups of patients based on race and ethnicity but studies are inconsistent. One study found no difference in survival for black patients with diffuse large B-cell lymphoma compared to non-Hispanic whites and another found worse 5-years OS for black patients with diffuse large B-cell lymphoma compared to non-Hispanic whites, however, both studies did not take socioeconomic factors into account [36, 37].

There have been contradicting results in the literature regarding immigrants’ health. In general, first-generation immigrants have shown better health compared to the host population in the receiving country. This phenomenon is known as the healthy immigrant effect or the healthy immigrant paradox [38]. The phenomenon is a paradox because immigrants often have lower socioeconomic position and fewer resources compared to the host population, which would otherwise be perceived as disadvantages in terms of health outcomes in general [39, 40], but also specifically in hematological cancers [41,42,43,44,45]. Some immigrants also originate from countries with lower life expectancy [46], but it has been shown that mortality for immigrants becomes closer to the mortality for native-born over time [47]. On the other hand, healthier individuals are more likely to emigrate than individuals with health issues, and host countries may also be more willing to accept healthy immigrants [48]. The magnitude of the healthy immigrant effect differs between European countries [49]. A systematic review from Denmark showed higher morbidity for non-Western immigrants compared to Danish-born citizens, but lower mortality [50]. The latter finding was supported by a Danish observational study of 27,134 immigrants followed from inclusion in 1993 to 1999 until 2008. The relative risks of all-cause mortality were 0.44 for women immigrants and 0.43 for male immigrants using Danish-born citizens as references and with adjustments for age and income. This supports the presence of a healthy immigrant effect in Denmark [34].

The main strength of this study is the population-based cohort from registers with almost complete follow-up. Emigration from Denmark is very limited, and all deaths are registered by the government, along with additional important information such as education level, cohabiting, occupational status, and income. In Denmark, visits to the general practitioner and the hospital are free, but co-payments are required for medications and dental care. Furthermore, treatment of hematological cancers is centralized to a few hospitals in each region of Denmark and hospital allocation is based on home region, with only very few exceptions of inter-hospital transfers. Choice or availability of hospital is therefore not dependent on living in a high- or low-income neighborhood, thus also contributing to the equalization of the impact of socioeconomic factors.

A major limitation of this study was the dichotomized stratification of immigrants into Western and non-Western immigrants as these are both very heterogeneous groups in terms of race, ethnicity, and host countries. No information on race and religion are allowed to be obtained in Danish registers and thus cannot be directly controlled for as confounders or risk factors in epidemiological studies relying solely on Danish register data. It is possible to identify information on the use of interpreters, used during hospital visits, in the Danish National Patient Register, as a pseudo marker for language barriers, but the information is not well-captured and would not contribute to the analyses performed here. Furthermore, patients with lymphoma were only stratified into aggressive, indolent, and Hodgkin lymphoma rather than specific lymphoma subtypes based on the International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3). This broad stratification was chosen to provide a comprehensive overview of the survival of immigrants with hematological cancers. Further studies are needed to provide insights into the specific lymphoma subtypes, as each subtype is associated with different characteristics such as age of diagnosis, treatment intensity, and survival probabilities [51, 52]. Consequently, the impact of immigration status and socioeconomic status may affect specific lymphoma subtypes differently through various interactions, which are not captured by the current analyses.

There may be a risk of selection bias, as the study cohort in the present study only consists of patients with a hospital-diagnosed hematological cancer. There may be differences in the incidence of undiagnosed hematological cancers. If immigrants were, hypothetically, less prone to being diagnosed due to difficulties navigating the healthcare system, the present study could potentially overestimate the relative survival of immigrant patients compared to Danish-born patients. The study showed some significant differences in subgroup analyses, but multiple tests were performed, and type 1 errors cannot be excluded. Since formal adjustments for multiple testing were not performed, the subgroup analyses should only be viewed as hypothesis generating. Lastly, there were generally small numbers of immigrant patients included in these analyses, especially compared to the large number of Danish-born patients included, and the results should be interpreted with this in mind.

In conclusion, the present study found no overall differences in survival when comparing immigrant patients, both Western and non-Western, to Danish-born patients in analyses controlled for imbalances in confounders. However, the results from subgroup analyses indicate that non-Western immigrant patients with lower income and reduced work capacity had better survival than Danish-born patients. Healthcare utilization was slightly higher among non-Western immigrant patients before and after diagnosis, but differences were small on an individual patient level. While results were generally encouraging and did not reveal profound impact on survival outcomes by immigrant status, the study provides no information on quality of life, patient experience, and shared decision making, which could still be compromised in patient groups with poorer language skills.