Background
International migration has increased significantly (by 41%) since 2000. In 2015, it was estimated that there were 244 million international migrants globally, the majority (151 million) with destination countries in Europe and Asia [
1]. In many recipient countries, international migration is becoming an increasingly important determinant of population change. For instance, in January 2016, it was estimated that 35 million residents (approximately 6.9% of the European Union (EU) population) in the EU were born outside of the EU, in addition to 19.3 million persons who were living in a different EU Member State from the one in which they were born [
2,
3]. Forcible displacement, as a result of conflict, persecution, violence or human rights violations, has also reached a record-high, with an estimated 21.3 million refugees globally in 2015; an increase of 55% since year-end 2001. This is largely attributable to the ongoing civil conflict in the Syrian Arab Republic [
4].
A number of countries have official resettlement programmes for refugees, including the USA, Canada, Australia, New Zealand, the UK and many others [
5‐
7]. The UK government accepts refugees under four different schemes, namely the Gateway Protection Programme, the Mandate Resettlement Scheme, the Syrian Vulnerable Persons Resettlement Scheme (VPRS) and the Vulnerable Children Resettlement Scheme (VCRS) (hereafter collectively the ‘UK programme’). The Gateway Protection Programme has committed to resettle approximately 750 refugees per year on the basis of their refugee status and need for resettlement [
8]. The Mandate Resettlement Scheme is much smaller, and applicable only to individuals who have been granted refugee status by the United Nations High Commissioner for Refugees (UNHCR) and who have close ties
1 to the UK. The VPRS and VCRS, on the other hand, represent specific resettlement schemes that the UK has devised to offer protection to people on a larger scale in times of crisis [
9]. For this reason, and due to its recent rapid expansion, the VPRS is probably the most high-profile of the UK schemes.
The VPRS was established by the UK Government in January 2014 in response to the Syrian crisis [
10]. It aims to enable vulnerable Syrians and other nationalities affected by the conflict to settle in the UK, prioritising those who meet the UNHCR vulnerability criteria, including women and children at risk, survivors of violence or torture, refugees with legal or physical protection needs, medical needs or disabilities, children and adolescents at risk, and refugees with family links in resettlement countries [
10]. Initially small, and with no fixed quota, it has increased in prominence following a pledge by the UK Government in September 2015 to resettle up to 20,000 people from the Syrian region by 2020 [
11]. This has attracted heightened media coverage and public interest (Additional file
1: Appendix III).
2 Subsequently, the VCRS was established in January 2016 to support and resettle up to 3000 vulnerable and refugee children and their families affected by the conflict [
12]. As of the last quarter of 2016, 20,878 refugees had been resettled through the entire UK programme [
13].
Under these schemes,
3 refugees are referred by the UNHCR and reviewed by UK authorities for resettlement in the UK. Prior to departure, a detailed health assessment (HA) is performed by the International Organization for Migration (IOM). The aim of the HA is to facilitate early integration of the refugee, promoting individual health, protecting public health where relevant and linking individual needs with appropriate health and social services in the UK. The UK HA protocol has recently been reviewed and updated with this in mind, to align it more closely with UK public health policy and best practice [
14]. The components of the HA are briefly outlined in Table
1.
Table 1
Components of the standardised pre-entry health assessment for refugee applicants
General assessment | Medical history |
Physical examination (vital signs, assessment of systems, oral and dental examination, skin examination, developmental milestones for children) |
Routine laboratory and radiological examinations, including urinalysis and chest x-ray |
Testing for specific conditions | Tuberculosis (according to the UK tuberculosis technical instructions [ 38]) HIV Syphilis Other sexually transmitted infections Hepatitis B and C Helminthic infection (as appropriate, according to protocol) Malaria (as appropriate, according to protocol) |
Immunisation | According to the UK immunisation schedule [ 39] |
Additional clinical assessments | Relating to other chronic, physical, psychosocial or mental health issues, as appropriate |
There is evidence that most migrants in Europe, at least initially, are relatively healthy compared to the host population, although migrants do face specific health challenges and may experience a deterioration in health over time in the host country [
15,
16]. It is possible that refugees, including those resettled through international resettlement schemes, may be at slightly higher risk of infectious diseases due to a higher prevalence of these diseases in their country of origin, specific circumstances of their residency and travel, and programme selection criteria which favour vulnerable migrants. However, there is limited information available on the exact epidemiology of infectious diseases in these groups. Therefore, there is a need to analyse these data and compare them with other sources of prevalence figures to ensure that appropriate public health measures, including HA, can be applied to these population groups most at risk and that individuals can be thus linked early to appropriate healthcare services in the UK.
This paper aims to analyse and describe, for the first time, data on the prevalence of all infectious diseases (tuberculosis (TB), HIV, syphilis, hepatitis B and hepatitis C) from a large cohort of refugees who underwent comprehensive pre-entry health assessments as part of the UK resettlement programme. It compares the recorded prevalence against published estimates in order to assess whether moving to risk-based testing would be feasible.
Discussion
This is the first study which reports on, and compares findings of, medical HAs for infectious diseases among a UK-bound refugee population. We found higher diagnostic yields than expected for a number of diseases, including hepatitis B.
For TB, testing yields broadly mirror WHO-estimated prevalence figures [
24]. The UK programme is particularly focussed on resettlement of vulnerable refugees and, whilst the possibility of testing bias cannot be ruled out (see below), it is likely this refugee population significantly differs from the general population of the respective country. In addition, the limitations of WHO prevalence estimates have been well recognised [
25] even in politically stable countries, and these limitations may be increased by political unrest present in many of the sender countries [
26].
TB testing results among refugees have been highly variable. Active TB yields for German-bound Syrian asylum seekers range between 93 and 153 per 100,000 [
27,
28], with some authors estimating significantly higher estimates [
29]. However, other countries found significantly lower yields, as illustrated by the Dutch (22 per 100,000) [
30]. Where reported in comparable pre-entry testing programmes, refugees tend to have higher TB testing yield than other migrants [
5].
In our analysis of confirmed cases of TB, we demonstrated an association between active TB disease and history of TB. This is not unexpected and could reflect recurrence or reinfection and the larger proportion of cases in this cohort that came from high TB burden countries, who may have been previously exposed to TB or not completed treatment. Based on the analysis of suspected cases (Additional file
1: Appendix VI), there is additional evidence that a number of other factors may be associated with TB, including increasing age and the presence of previous household contact with TB cases. Whilst these are largely expected findings [
31], they are important to help inform testing policy and guide clinical practice on the ground.
As with TB, HIV prevalence rates generally mirrored WHO figures, although we found that overall yields were higher than those that would be predicted by WHO figures. This may reflect a more vulnerable, high-risk population than expected based on the resettlement criteria of the UK programme. Among refugees from DRC, for example, HIV testing yield was five times higher than the WHO prevalence estimate, at 3.6% (2.3–5.6%) compared to the WHO estimate of 0.7% (0.6–0.9%). Overall, however, the prevalence of HIV was still relatively low (0.3%) but varied significantly between countries. Sub-Saharan African countries accounted for the majority of cases of HIV infection in this cohort, reflecting the higher prevalence rates of HIV in this region. On the other hand, the generally low prevalence seen among refugees from the Eastern Mediterranean region compared to those from sub-Saharan Africa suggests a potential practical advantage of applying a risk algorithm for determining which individuals should be tested.
Our analysis identified a number of factors, both demographic and behavioural, which increased the odds of HIV infection, including being female between the ages of 35–49, of sub-Saharan African nationality and with a history of STI. The increased vulnerability of women to HIV infection stemming from biological, social, behavioural and structural risk factors is well recognised [
32,
33]. The finding of syphilis being associated with a reported history of torture is interesting, however perhaps not surprising given that more than half (12/23; 5 female, 7 male) of syphilis cases are from Sudan, where experiences of torture were generally more prevalent (reported in 21.43% of Sudanese applicants). The higher odds of syphilis among individuals with HIV are not unexpected and again reflect the biological mechanisms and similar risk factors which facilitate transmission.
A large number of refugees were identified to have hepatitis B infection in this cohort. Overall prevalence was over 2% but, as with other infections, we found that prevalence varied substantially between countries and with respect to WHO estimates. Whilst sub-Saharan African countries had particularly high prevalence, Syrian refugees accounted for most cases. In particular, testing yields for refugees from Somalia (3.39%, 1.97–5.75%), Sudan (5.82%, 3.82–8.76%) and South Sudan (12.50%, 5.24–26.96%) were noticeably lower compared to prevalence estimates (Table
5), which could reflect the limitations of prevalence estimates from these countries, but also that the refugee population may be different to the general population. An additional consideration is that these countries may have WHO-recommended universal and selective hepatitis B vaccination programmes, although vaccine coverage is unlikely to be optimal in countries with fragile infrastructures and during conflicts. The disparity between observed testing yield and country prevalence estimates therefore needs to be explored further. The high yield from the other European region category (20.00%, 2.11–74.35) reflects higher rates associated with refugees from Turkey [
34,
35], an intermediate endemicity country, but the low numbers are noted [
36].
We demonstrated that a number of factors are associated with increased odds of hepatitis B infection, including being male, increasing age, sub-Saharan African nationality and a history of STI and blood transfusion. It is likely that male predominance may be due to adult exposures more associated with males, and should be explored further.
Hepatitis C testing yield was considerably lower (0.41%) than seen for hepatitis B, although again this varied by nationality and in comparison with prevalence estimates, likely reflecting different risk exposure. We demonstrated that the main factors associated with increased odds of hepatitis C were older age (> 50 years) and history of blood transfusion, yet unlike hepatitis B, there was no association with geographic region. The strong association with blood transfusion particularly among the older age groups likely reflects the lack of routine blood-borne virus testing in many low- and middle-income countries, possible iatrogenic transmission through reusing of needles and medical equipment, and potentially chronic infection in some cases. It is interesting, although not totally surprising, that illicit drug use was not reported, considering the high prevalence of this exposure in some countries; however, this is not necessarily a dominant risk factor in those countries from which the majority of screened refugees originate.
The current HA programme run by IOM undertakes a significant number of tests in this vulnerable refugee population. Our novel work highlights that prevalence of infectious diseases varies widely, raising the possibility of changing testing from a blanket modality to a more nuanced, risk-based model that targets those at highest risk. Our findings also demonstrate that refugees are not a homogeneous group and provides a baseline for further evaluation of the effectiveness of the HA in facilitating initial linkages with primary care and in the years following resettlement.
Given that the primary aim of the UK programme and HA is to facilitate early integration and linkage of the refugee to appropriate health and social services in the UK, it is important that the HA is tailored with this end goal in mind and conducted according to what is appropriate for the individual, performed with the voluntarily provided informed consent of the individual. Informed consent is a key element in the protocol [
14], yet it is important to be conscious that resettlement circumstances may potentially affect the freedom of consent.
Our study benefits from a large, well completed and comprehensive dataset of UK-bound refugee testing. Nevertheless, these observational data have a number of limitations, including data recording issues with potential for incomplete data or misclassification. Whilst the dataset was not subject to the routine IOM validation process, there has been rigorous data cleaning and validation before analysis to minimise such issues and it is therefore likely that the potential for misclassification is small and occurring at random. For active TB, IOM provided a rigorous case ascertainment exercise which verified status according to culture confirmation with the attending physicians in the field for all suspected cases.
Detecting diseases depends on the availability and quality of testing sites and it is possible that this may lead to testing bias. Most diseases require confirmatory tests and we employed very robust algorithms for case definition, so any testing bias would likely lead to case under-ascertainment. We have analysed the impact of assessment site on disease prevalence and think that these effects are likely minimal. For active TB, there is a possibility of confirmed case under-ascertainment owing to the methodology used.
Detected disease yields in our study are often similar, but sometimes different compared with WHO-estimated disease prevalence. It is expected that infectious disease risk and prevalence in this refugee population is different from the general population, not least because of differences in socioeconomic circumstances, access to care and/or accommodation, including overcrowding and camp conditions, which would minimise the generalisability of our yield as disease prevalence for the specific countries. Likewise, our study population is generalisable to UK-bound refugee populations and likely to refugee populations to other destination countries with similar programmes (e.g. USA, Canada, Australia, New Zealand). However, the generalisability of our results to other migrant or asylum seeker populations is limited due to differences in epidemiological profiles, socioeconomic status and possible selection bias (e.g. due to different selection criteria of resettlement programmes). Nevertheless, our findings provide an important snapshot into infectious disease risk of UK-bound refugees and yields important lessons to inform public health measures in this vulnerable population.
The limitations in self-reporting of risk factors, particularly if potentially considered criminal or stigmatising in the country of origin, should also be considered here. Whilst the null report of illicit drug use among hepatitis C cases may be real, it could also reflect lack of disclosure in response to fear of stigma or legal implications affecting rights to resettlement.
A further limitation is that the data recorded provided disease prevalence on a select group of refugees, predominantly from Africa and the Middle East, with fewer from Asia, who may have had a different infectious disease profile.