The negative health effects of hostile environment policies on migrants: A cross-sectional service evaluation of humanitarian healthcare provision in the UK

Background: Recent UK ‘hostile environment’ immigration policies, including obligatory charging and sharing of confidential data between NHS Digital and the Home Office, have created an atmosphere of fear and exposed already highly marginalised and vulnerable groups to significant health risks by increasing barriers to accessing NHS care. Methods: This is a cross-sectional observational study of patients accessing healthcare at Doctors of the World (DOTW) in the UK. DOTW is a humanitarian organisation, providing care to those excluded from NHS healthcare. We aimed to describe population characteristics of individuals using DOTW services and identify groups at greatest risk of facing ‘hostile environment’-related barriers to NHS care, specifically being denied healthcare or fear of arrest. Results: A total of 1474 adults were seen in 2016. Nearly all were non-EU/EEA nationals (97.8%; 1441/1474), living in poverty (68.6%; 1011/1474). DOTW saw a large number of undocumented migrants (57.1%; 841/1474) and asylum seekers (18.2%; 268/1474). 10.2% (151/1474) of adults seen had been denied NHS healthcare and 7.7% (114/1474) were afraid to access NHS services. Asylum seeker status was associated with the highest risk (adjusted odds ratio (OR): 2.48; 95% confidence interval (CI): 1.48-4.14) of being denied NHS healthcare and being undocumented was associated with the highest risk of fearing arrest (adjusted OR: 3.03; 95% CI: 1.70-5.40). Conclusions: Our findings make visible the multiple and intersecting vulnerabilities of individuals forced to seek care outside of the NHS, underlining the public health imperative for the government to urgently withdraw its ‘hostile environment’ policies and address their negative health impacts.


Introduction
There are an estimated 258 million international migrants worldwide, of which 25.9 million are refugees and asylum seekers 1 . The UN Secretary General described global migration as "one of the most urgent and profound tests of international cooperation in our time" 2 . Experts in the field have called for the active promotion and protection of the health of migrants 3 , while the Academy of Royal Colleges of Medicine has called publicly for the rescindment of policies that hinder migrants' access to healthcare 4 . The right to health for all is one of a set of internationally agreed human rights standards, detailed under the International Covenant on Economic, Social and Cultural Rights 5 .
In 2012, in stark contrast to this public health imperative and international human rights laws, the recent UK Prime Minister and then Home Secretary, Theresa May stated her aim "was to create here in Britain a really hostile environment for illegal migration" by denying individuals without legal documentation everything they needed to survive, including healthcare 6 . The UK's Immigration Acts of 2014 and 2016 extended 'hostile environment' immigration policies into far-reaching aspects of public life. This led to mandatory upfront charges to 'non-ordinary residents' for National Health System (NHS) secondary care, including undocumented migrants and those denied asylum. Some migrants are exempt from charges; however, a lack of understanding of the rules and difficulties in proving exemption means people are wrongly denied or charged for care 7 .
An extension of the hostile environment policies was highlighted by the discovery of a much-criticised Memorandum of Understanding between NHS Digital and the Home Office 8,9 which facilitated the sharing of confidential non-clinical information about patients for immigration enforcement purposes. Though this MOU was withdrawn in 2018, patient information collected by NHS trusts is still shared with the Home Office 10 and these extreme measures have extended an environment of mistrust, increasing the direct and indirect barriers that many migrants face in accessing healthcare: many fear arrest, are incorrectly denied access, or are unable to afford care 11 .
Migrants who are unable to access NHS services and in desperate need of healthcare seek assistance at the clinics of the international humanitarian organisation, Doctors of the World (DOTW). These clinics provide walk-in, free primary care and health advice, consultation for broader social problems, and advocacy to enable people to register with NHS GPs. Importantly, awareness of these clinics is often through word of mouth and via trusted contacts and as such these clinics provide services to a highly marginalised group of individuals who are unable or too fearful to access healthcare elsewhere. Attendance at these clinics is evidence of exclusion from mainstream healthcare provision.
Doctors of the World's recent Médecins du Monde Observatory Report 12 , based on data from 43,286 people attending clinics run by DOTW and partner NGOs across Europe in 2016, found 55.2% of attendees had no access to healthcare coverage at all, with 18.3% accessing emergency care only. Individuals sought medical attention for both acute and chronic conditions and over half of pregnant women were not accessing antenatal services. The report described a vulnerable and highly excluded population where social isolation was common. Most notably the highest number of responses reporting fear of arrest and denied access to healthcare was in the UK.
There is a need to identify the characteristics and healthcare needs of this highly marginalised group at the countryspecific level, in order to evaluate and improve the services provided by DOTW, hold national policy makers to account, inform future policy and advocate for meaningful public health interventions. This study contributes to this goal by looking specifically at data collected in the United Kingdom, during the preparation of the Médecins du Monde Observatory Report, and considers these data within the context of prevailing UK government policy. The objectives of the study are to evaluate the characteristics of the population using DOTW services, i.e. those excluded from NHS services, and identify those groups at greatest risk of facing 'hostile environment'-related barriers to NHS care.

Study design
We conducted a cross-sectional humanitarian health service evaluation using data collected during all consultations at DOTW clinics in Bethnal Green and outreach locations in London and Brighton, from 1st January to 31st December 2016. As this was a service evaluation all adults (>18 years old) attending the clinics were eligible for inclusion; those aged <18 years were excluded. Attendees at DOTW clinics underwent two consultations, one "social" and one "medical". The "social" consultation was conducted by casework volunteers and covered issues such as housing, immigration, income and health access. "Medical" consultations were conducted by doctors and focused on current health, medical history, pregnancy and vaccination. All information was entered into a database during the consultation. For this analysis we focus solely on data collected using the "Social" form (available as Extended data 13 ).
Age was grouped into age-band variables and EU national status was assigned based on self-reported nationality (where dual nationality was reported and one nationality was in the EU, EU nationality was recorded). Income was coded as 'above' or 'below' a poverty threshold of £1050/month. As income was missing for many (n=349) participants, these were assigned as 'income data missing' to allow them to be retained within regression models. Immigration status was coded as either 'asylum seeker' (defined as a person who seeks safety from persecution or serious harm in a country other than his or her own and awaits a decision on the application for refugee status under relevant international and national instruments), 'undocumented' (defined as a person who does not have the necessary documentation to enter or remain legally in a country) or 'permissions to reside (other than asylum)'. The 'permissions to reside (other than asylum)' immigration status group consisted of individuals whose status was categorised as one of the following: has valid residency permit and permit end date; tourist; student; short stay visa; work visa; humanitarian protection/ discretionary leave; permit to stay in another EU country and here for <3 months; specific situations giving a right to stay (specific per country): under process for sorting out papers; or unknown.
Where an individual attended a DOTW clinic multiple times, we selected data from the record with the most complete data for that individual, with a preference for the first attendance if data were equally complete.

Outcomes
Descriptive analysis was used to examine the characteristics of those using DOTW services. We interrogated data on age, gender, nationality, income, housing, parental status, immigration status and barriers in access to NHS healthcare. In further analysis we sought to investigate which population characteristics were associated with facing hostile-environment-related barriers to NHS care. The two hostile-environment-related outcomes that we investigated were whether people disclosed (a) being denied health coverage or (b) fear of being reported or arrested, as a reason for not accessing NHS care within the past 12 months. As each individual could give multiple reasons for not accessing NHS care, we created two separate variables, one for each of the responses noted above.

Statistical methods
Descriptive results are presented using counts and percentages. Data are suppressed where cell counts would be less than 10, or where column or row totals could be used to deduce suppressed cell counts. We conducted multivariable logistic regression analyses to investigate the effect of explanatory variables on the likelihood of experiencing each of the outcome measures, respectively, having adjusted for other explanatory variables. We conducted a series of sensitivity analyses to explore the effects of excluding missing data on the final logistic regression models presented in our analyses. Data were stored and analysed using Stata version 15 at University College London (UCL).

Ethics
Prior to undertaking this analysis, this project was reviewed by the joint chairs of UCL's Research Ethics Committee, who deemed it exempt from requiring ethics clearance on the basis that it constituted a service evaluation aiming to improve understanding of demand for and access to DOTW services and who deemed that participant consent was not required. DOTW provided anonymised data for this analysis.

Study population demographics
In total, 1543 individuals were seen at DOTW clinics in the UK in 2016. After removal of individuals aged under 18, we were left with 1474 individuals for analysis. Table 1   undocumented. Figure 1 shows the wide global distribution of individuals seeking care at DOTW clinics. The top three nationalities for individuals accessing the DOTW clinic were Filipinos (15.9%; 234/1474); Chinese (11.5%; 169/1474) and Indian (10.0%; 147/1474).

Barriers to accessing NHS healthcare
We examined barriers to accessing NHS healthcare in this population ( Figure 2), with individuals reporting one or more barriers where applicable. Administrative barriers were reported at the highest levels (15.9%; 235/1474), followed by a lack of knowledge on how to access NHS care (11.3%; 166/1474), previously having been denied access to care (10.2%; 151/1474), and language barriers (9.9%; 146/1474). We analysed the healthcare access barriers by immigration status. For undocumented migrants, fear of arrest (81.6%; 93/114) was the healthcare access barrier they were at greatest risk of experiencing compared to other immigration status groups ( Figure 2B). For asylum seekers, a previous bad experience was the healthcare access barrier they were at greatest risk of experiencing (however, as the total number was <10 this was not reported) ,  Table 2 presents the results of our multivariable logistic regression examining the risk factors for denial of access to healthcare and fear of arrest upon seeking healthcare. People who had made an asylum claim were more than two times likely to have been denied NHS care compared to those with permissions to reside (adjusted odds ratio (OR): 2.48; 95% confidence interval (CI): 1.48-4.14) after adjusting for age, gender and income level.
In total, 114 people (7.7%; 114/1474) hadn't tried to access the NHS healthcare due to fear of arrest. This barrier was more common among females (58.8%; 67/114) and those who were  undocumented (81.6%; 93/114). In a multivariable analysis, people whose immigration status was 'undocumented' were three times more likely not to seek care due to fear of arrest compared to those with permissions to reside (adjusted OR: 3.03; 95% C.I.:1.70-5.40) after adjusting for age, gender and income level. A sensitivity analysis was conducted to examine the effect of missing data and results were consistent with our primary analysis (Table S1, available as Extended data 13 ).

Discussion
Our study describes the characteristics of the population using DOTW services in London and Brighton, providing a unique and timely insight into a highly excluded population and the likely impacts of the increasingly 'hostile environment' imposed on some migrant groups in the UK. This UK-based humanitarian health service evaluation demonstrated that the majority of this population were young, non-EU/EEA nationals and living in poverty. Over half had an undocumented immigration status and close to a fifth had claimed asylum. Identifying groups at greatest risk of facing 'hostile environment'-related barriers to NHS care, we found a tenth of individuals have in fact sought and been denied NHS care and asylum seekers were at highest risk. Furthermore, seven percent of those presenting at DOTW clinics hadn't tried to access NHS care due to fear of arrest, and this was three times more likely for those with an undocumented immigration status.
The very fact that individuals are seeking healthcare outside of the NHS, at DOTW clinics, is itself evidence of marginalisation. Our findings are significant in highlighting the multiple and likely intersecting characteristics which may contribute to marginalisation and exclusion, such as poverty and precarious immigration status/undocumentation. The results are consistent with a number of recent studies in Europe concerning the characteristics of undocumented migrants and their healthcare needs 14,15 . The Médecins du Monde Observatory Report, which included 43,286 individuals, found that more than half of those reporting 'fear of arrest' as an obstacle to seeking healthcare, were in the UK 12 . More recently, health and legal professionals assisting at Grenfell Tower found victims reluctant to go to hospital because of concerns about their immigration status 16 . Individuals with undocumented immigration status are the express target of the UK government's 'hostile immigration policy'. This study found that undocumented individuals were three times more likely to be afraid of arrest upon accessing NHS care. This provides quantitative evidence supporting the assertion that UK hostile environment policy is deterring individuals from seeking healthcare, when access should not be dependent on immigration status or production of documentation 17 .
Women migrants are particularly vulnerable to these hostile policies since they are more likely to enter the country as dependents or to be employed in 'unskilled' professions 18 . This makes it more likely for them to be or become undocumented 18 , and subsequently afraid to access NHS care. Secondly, with the introduction of new charging regulations, the inability to pay a debt for receiving care can negatively affect immigration status and therefore act as a deterrent to seeking care. Our findings are consistent with recent findings that charging leads to avoidance of NHS healthcare because of a fear of charges and the humiliation of being refused care 19 .
Asylum seekers were found in this study to be twice as likely to have sought and been denied NHS care than those with permissions to reside, attending DOTW clinics. Asylum seekers are legally entitled to the same healthcare rights as ordinary UK residents, so our finding demonstrates that 'hostile environment' NHS policies are negatively affecting a wider group than their explicit intention. This is reiterated by a recently published paper that highlights the significant and multifactorial issues faced by asylum seekers when trying to access healthcare, including being wrongly denied access due to having no proof of address or because healthcare providers lack an understanding regarding their entitlement 20 . A recent systematic review of systematic reviews found migrants experience widespread racism, discrimination, stigma and stereotyping by healthcare professionals 21 . Poor understanding of migrants' entitlements and complex gatekeeping systems, little understood by clinicians and administrators, also have a part to play 22 .
Denying access to primary care has well-recognised public health and economic implications 11,23 . Migrants in the UK, for example, utilise screening and treatment adherence programmes less than locally born individuals 24 , whilst others describe barriers to booking and attending routine vaccinations 25 . Limiting access to primary care increases A&E attendance and hospital costs 26 . A recent study in Sicily found undocumented migrants had a higher risk of avoidable hospitalisations due to chronic health conditions, compared to those with documentation 27 . To further understand the economic impact of policies restricting access to healthcare, we refer to a study in Germany which found that per capita healthcare expenditure was 40% higher among those with restricted access to primary care compared to those with regular access 28 . The failure to provide good access to primary care and the likely costs associated with subsequent avoidable hospital admissions must be seen in the context of a government that has wholly failed to build a convincing economic rationale or to evaluate the public health cost of its hostile environment policies.
This health service evaluation aimed to improve the understanding of demand for and access to DOTW services. As such our analysis included a full year's-worth of attendees at DOTW's clinic, leading to a substantial sample size. DOTW clinics have provided a service, independent of the UK Government, for over a decade. It is likely their services are trusted by migrant communities, meaning this study represents those with the greatest need for healthcare who are excluded from NHS services. Furthermore. data was collected after implementation of 2015 NHS charging regulations 29 and provides important insights into the context and likely impact of this fundamental change in policy.
The total number of undocumented migrants in the UK is unknown, making it difficult to draw conclusions regarding the extent to which this group's health needs are being met, or to estimate the extent or direction of possible biases to those accessing DOTW services. For example, though DOTW offer travel expense reimbursement, services may be less attended by those further away. Newer migrants, more fearful or less knowledgeable, may also be less likely to attend. Lack of population denominator also prevents estimation of the differential 'risk' of population sub-groups (e.g. males vs females) of being excluded from NHS care. Our data only covers attendances during 2016. This means we are unable to investigate the effects of significant policy changes in 2017, when updated charging regulations were enforced, or when the information sharing agreement between NHS Digital and the Home Office was formally retracted in 2018. Furthermore, all information is self-reported, which may lead to inaccuracies; those who are more fearful, for example, may not truthfully report their true immigration status, and others might be embarrassed to admit they are too afraid to access NHS services.
The findings from our health service evaluation make visible the multiple, intersecting vulnerabilities of individuals forced to seek care outside of the NHS. Our results underline the public health imperative for the government to urgently withdraw all of its 'hostile environment' policies and address their negative health impacts on these vulnerable individuals, and the wider UK population. Further policy changes in 2017 (e.g. the formal data sharing between NHS Digital and the Home Office; and more charging regulations) have likely only increased the barriers to NHS care that these people face. These 'hostile environment' government policies represent an ideologically driven and reckless experiment with the health of these vulnerable people without any evidence base and are not subject to any proper ongoing impact assessment. There is a pressing need to better characterise, understand and meet the healthcare needs of this population and for ongoing research into the public health impact of these policies.

Data availability
Underlying data Underlying data cannot be shared owing to Ethical and Security Considerations. As we describe in our analysis, international migrants accessing DOTW humanitarian health services are often doing so because they are scared to access mainstream NHS services, due to data sharing between the NHS and UK Home Office. For this reason, we are unable to share the underlying data used in this analysis. Researchers wishing to use the data should contact Doctors of The World (email: info@doctorsoftheworld.org.uk) who will review applications on a case by case basis. Author contributions SJW, LJC, RWA, AM and LJ substantially contributed to the conception of the work, LJ, AM and RWA were responsible for the acquisition of data, while LJC and RWA conducted the analysis. All authors reviewed and interpreted the results, edited the manuscript and agreed to be accountable to all aspects of the completed work. The paper is well placed in the context of the UN and ICESR where Health is defined as a Human Right, and the issues of the relevance to 'Public Health' are quite well laid out -e.g. the impact on 'herd immunity' by effective exclusion from vaccination programmes. The cost issues for the national health system are also considered briefly -this is an issue which needs more hard data, but has been explored in a European context by PICUM (the programme for Undocumented Migrants). I would have liked a bit more reference to such relevant contextual research, but accept that the space is limited! Perhaps we can develop this via responses and relevant networks? There is also a useful resource in the Migrant Policy study of access to healthcare across national settings -(Declaration of http://www.mipex.eu/health interest -I helped compile this for UK).

Grant information
The paper shows very clearly that there are people who officially do have eligibility for health care and being (or perceiving that they are) refused service on the grounds of migrant status -the statistics might have been clearer displayed with proportions broken down by reasons for refusal against Migration status -rather than showing the proportions of each type of status against reasons for non uptake. It is significant that language is NOT the most important barrier to access, or indeed 'knowledge' -it might have been useful to have cited some exemplars of 'administrative barriers' (a benefit of qualitative research methods!). The link to the experiences of people affected by the Grenfell Tower disaster is significant and affecting. The ramifications of the 'Windrush Affair' might have come too late for inclusion, but that is another similar scandal.
It might (perhaps in a future similar analysis?) have been of interest to have recorded also 'Faith'/religion and some measure of 'race'/ethnicity -does 'perceived difference' add to the disadvantage or is personal awareness of being a migrant, in one of the at-risk groups, the dominant factor in self-selection for non-treatment and use of/reliance on this DoTW service?

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate? Yes 1 2 Yes migrants to denying access, but we know that in this under attendance, cultural factors as well as organisational problems not being denied access play an important role.
I have the following queries: Data were gathered in Brighton and London, but nowhere I can see the results per city; I wonder if there are no differences between those two cities.
I missed the information on the provision of interpreter services during the data collection and consultations.
I missed the information on the educational level of the participants; as we know, health disparities including barriers to access care, due to lower educational/limited health literacy are substantial, I think this infomation would be relevant.
Page 5, left column last sentence: "For asylum seekers, a previous bad experience...". For me it is not clear if having a bad experience was the most frequent barrier in accessing care for asylum seekers (and not e.g. fear, lack of money etc.) or that "having a bad experience "was mentioned most frequently by asylum seekers, compared to the other subgroups. All in all, an important, good paper, worthwhile for indexing.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? No

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
Reviewer Expertise: Migrant care, in particular health and access to healthcare for refugees and undocumented migrants, health disparities, general practice/primary care, health disparities.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.