Making Gendered Healthcare Work Visible: Over-Looked Labour in Four Diverse European Settings

Healthcare has long been a gendered enterprise, with women taking responsibility for maintaining health and engaging with service providers. Universal healthcare provision notwithstanding, women nonetheless undertake a range of healthcare work, on their own account and on behalf of others, which remains largely invisible. As part of a multi-method comparative European study that looked at access to healthcare in diverse neighbourhoods from the point of view of people’s own health priorities, the concept of ‘healthcare bricolage’ describes the process of mobilizing resources and overcoming constraints to meet particular health needs. Bricolage mediates between different kinds of resources to meet particular challenges and describing these processes makes visible that work which has been unseen, over-looked and naturalised, as part of a gendered caring role. Drawing on 160 semi-structured interviews and a survey with 1,755 residents of highly diverse neighbourhoods in Germany, UK, Sweden and Portugal, this article illustrates the gendered nature of healthcare bricolage. The complex variations of women’s bricolage within and beyond the public healthcare system show how gendered caring roles intersect with migration status and social class in the context of particular healthcare systems.


Introduction
The gendered nature of healthcare work is inherent to the way that health and welfare systems have developed over the years: supporting the health of others, both in private settings and mediating with professional service providers, has been taken for granted as a natural part of women's roles as mothers, daughters, sisters, aunts, wives, partners and neighbours. While health and welfare systems provide services and sup-port to avoid dependence on family members, women's labour has never been entirely replaced by paid services (Annandale, 1998). The rise of globalized migration (Castles & Miller, 2009) has brought new gendered family regimes into contact with European health and welfare services, as well as disrupting expectations of cultural and linguistic familiarity and recognition (Brochmann & Hagelund, 2012;Crepaz & Lijphart, 2008). Migration is a gendered process, favouring variously aged and gendered groups, depending on the migration route and regime of the sending and receiving countries (Donato & Gabaccia, 2015;Yuval-Davis, Anthias, & Campling, 1989), such that gender, migration and health interact in complex ways. This article considers how gender plays out in populations characterized by migration-driven diversity, in seeking healthcare appropriate for their self-defined health needs, in four different European welfare settings.

Background
Theoretical approaches to gender have developed apace since binary conceptions of men and women's health gave way to relational (Connell, 1987) and intersectional (Crenshaw, 1991) approaches. Theoretical approaches that have sought to understand the complex contingencies of how gender plays out alongside other aspects of identity and structure have yet to be regularly operationalized in health research (Annandale, 2013), despite efforts to establish working definitions (Hammarström et al., 2014).
Recognition that gender is fundamental to health outcomes through the lifecourse is represented in the routine presentation of mortality and morbidity figures disaggregated by gender and age. While health outcomes are patterned by gender, our interest here is in making visible the gendered processes involved in getting access to healthcare when living in diverse neighbourhoods, including their intersection with other dimensions of identity and structure, particularly migrant status and racism.
Ideally, healthcare would be both gender equal, in avoiding sexist discrimination, and gender equitable by "meeting the needs of women and men, whether similar or different" (Hammarström et al., 2014, p. 188). Access to healthcare is gendered both structurally, in terms of the political economy, and interactionally, in terms of how identity and discrimination play out interpersonally (Ettorre, Annandale, Hildebrand, Porroche-Escudero, & Rothman, 2017). Gender is both structural and individual, but is neither organizationally nor interpersonally determinant, since it intersects with other aspects of identity and of distribution (Fraser, 2000). This level of complexity is acknowledged by intersectional approaches which "go beyond additive analyses to study complex intersections as well as ensure that gendered power relations and social context are included" (Hammarström et al., 2014, p. 189). Intersectional approaches hold the promise of allowing the complexities of gender as they play out in different cultures, to extend beyond binary gendered categories (Richards et al., 2016).
This study was designed to map, both qualitatively and quantitatively, informal and formal work that is undertaken in diverse neighbourhoods, in order to meet self-identified healthcare needs. Our analyses have explored the rules of access for healthcare resources  identifying bricolage as a process undertaken by both healthcare providers (Phillimore, Bradby, Doos, Padilla, & Samerski, 2018) and users and defined as: A creative mobilisation, use and re-use, of wideranging resources, including multiple knowledges, ideas, materials and networks in order to address particular health concerns. (Phillimore, Bradby, Knecht, Padilla, & Pemberton, 2018, p. 6) Our quantitative analysis shows patterns of healthcare bricolage whereby young people, women and those with more education are more likely to engage in bricolage (Phillimore, Bradby, Brand, & Padilla, in press). The current article examines the gendered dimension of bricolaging healthcare, to explore how this work consists in persistent advocacy work, often in the face of dismissal or discrimination, that interacts with migration and family status to look at the effect it has on women's lives.

Methods
The study utilized a parallel sequential mixed method in which each phase of the research informed the next (see Phillimore et al., 2015). One city in each of four different countries was selected to represent four different welfare regime 'ideal types': Germany, Portugal, Sweden, and the UK, with each country sufficiently different to allow comparison. Ethical approval was obtained from the relevant ethics committee in each setting.

Qualitative Research
The study began with an ethnographic phase wherein researchers walked two selected diverse neighbourhoods in each city, observing how different healthcare resources were used while engaging in conversations about healthcare with residents and service providers. Insights from observations were used to develop a semistructured interview schedule (see supplementary file for schedule in English). Trained community researchers interviewed some 160 residents across the four countries, using their multi-lingual abilities and local networks. The community researchers were paired with academic researchers and together identified interviewees via networks, local organisations, and snowball sampling through street mapping and interaction with locals. Maximum variation sampling was used to ensure heterogeneity in terms of country of origin, age, gender, education level, income, ethnic and linguistic background. This comparison-focused sampling approach selects cases to identify factors explaining similarities and differences. Commonalities that emerge, despite many intersecting axes of difference, have increased validity because they do not result from sampling by predetermined characteristics.
Residents were interviewed in their preferred language, having signed consent forms stressing confidentiality and the option to withdraw from interviews. Names used in this article are pseudonyms. All interviews were digitally recorded, transcribed and where necessary translated. Data were coded collectively using a systematic thematic analysis approach to identify key issues. This involved interpretive code-and-retrieve methods wherein the data were read by the research team who collectively identified codes and engaged in interpretative thematic analysis. A shared codebook was devised between teams using MAXQDA software with the project lead checking inter-coder reliability across sites.
Analysis of these semi-structured interviews identified five types of bricolage that were used as models in the survey.
These types are summarised as follows: first, no bricolage, where people used only the public healthcare system, such that all services and treatments were either covered by health insurance or were publicly financed; second, within-system bricolage, whereby respondents used the public healthcare system plus informal support from family and friends or information sources such as the internet, magazines, the radio, to address the health concern; third, added-to-system bricolage, where respondents added advice, services or treatments that were not covered by the public healthcare system, included out-of-pocket services, alternative or complementary medicine or services from another country; fourth, alternative bricolage where respondents did not use the public healthcare system but used privately contracted services or informal and informational support; and fifth no resources used, where respondents did not use any resources to address the health concern.

Survey
Potential respondents were randomly selected from the eight neighbourhoods. All persons that were over 18 years old and residents of the respective neighbourhood were eligible for participation. The aim was to sample at least 300 individuals per neighbourhood. The fieldwork was undertaken between January and October 2017. Respondents were approached via invitation letters, phone calls and door-to-door visits. Response proportions ranged between 53% in Birmingham to 14% in Uppsala. The interviews were conducted by multilingual staff either face-to-face or over the telephone. All participants provided written, or for telephone interviews, verbal informed consent.
As we were particularly interested in the healthcare work of residents living in diverse neighbourhoods, respondents were asked how long they had been resident in the neighbourhood and whether they experienced any health concerns while living there. Their health concern was recorded and they were asked which resources they had used to address their concern. The full sample comprised 2,692 respondents. Of these, 937 were excluded from the current analysis because they had not experienced any health concern while living in the neighbourhood or because of missing values in the variables of interest, leaving 1,755 individuals for the analysis.

Statistical Analysis
Frequencies were used to describe the sample characteristics, in total and by gender. Gender differences in bricolage behaviour were assessed using multinomial logistic regression models with type of bricolage being the outcome variable. Four different models were run. The first model analysed the unadjusted relationship between gender and type of bricolage. In the second model we adjusted for survey country, age, education, income, migration background and self-rated health. In the third model we added a gender-by-country interaction term and in the fourth model a gender-by-migration background interaction was included. Since coefficients from multinomial regressions are difficult to interpret, we calculated marginal effects using the postestimation margins command in Stata. This command estimates the effect of changing the values of the factors on the probability of observing an outcome. Further, for the illustration of the interaction effects we estimated predicted probabilities with 95% confidence intervals. All analyses were carried out using Stata 15 (StataCorp, College Station, Texas).

Survey Findings
As Table 1 shows, the men and women who were included were similar regarding the covariates. Both in the crude and in the adjusted model, men were less likely to bricolage than women and, if they do report bricolage, then it tends to be contained within the healthcare system, with less adding to the system (see Table 2).
Considering the differences between bricolage patterns across countries, there is notable variation (Table 2, Figure 1). In the UK, there is comparatively little bricolage; women tend to do more within-system bricolage than men and, to a small extent, more adding to the system. Germany reflects the pattern that we found in the whole sample. The pattern in Portugal differs in that women tend to do less bricolage and less adding to the system bricolage than men. In Sweden, men do less bricolage, specifically less within-system bricolage (see Figure 1).
When analysing the tendency to bricolage by migration status we see this gendered pattern among those classified as non-migrants and also among the secondgeneration migrants (Table 2, Figure 2). However, among first generation migrants there is no difference between women and men's reporting of bricolage. This suggests that there is something about the migration process that, at least for a period of time, disrupts the standard patterning whereby women tend to bricolage more than men. Newly arrived migrants who find themselves in a new and unfamiliar healthcare system, may have a restricted ability to negotiate with service providers both because of limitations to linguistic and cultural knowledge. By the second generation both socio-cultural    ( Notes: a Adjusted for survey country, age, education, income, migration background, and self-rated health; b Adjusted for age, education, income, migration background, and self-rated health; c Adjusted for survey country, age, education, income, and self-rated health; * p < .05. and linguistic knowledge has been gained, permitting women, once again to take on the bricolage role.

Semi-Structured Interview Findings
Turning to the semi-structured interview material, below we present cases of women reflecting on the process of healthcare bricolage to address their own and family members' health conditions. We identify intersections of gender with migration status and other aspects of identity, which play out within the healthcare system in the process of seeking access. The gendered tendency for women to negotiate healthcare for themselves and their family members was disrupted by migration to another country where women had to acquire a new language and develop new networks of support. In some cases this made women highly dependent on their children or partners for translation. In caring for their children, partners and parents and themselves, women drew on whatever resources were to hand whether private, public or overseas, driven by the need to find a solution. For women who were also migrants, taking on a caring role in a new context implied learning new ways of navigating the system and negotiating care including new conceptual languages for illness, new vocabularies and practices of care and therapy and, sometimes, a new language. We spoke to some women, both migrants and non-migrants, who negotiated the care that they needed successfully and expressed their satisfaction. However, there were many women who de-scribed having to make significant efforts in order to get access to care and we heard about some situations that could not be overcome, despite these efforts.

Care as Persistence and Mediation
Women take on healthcare communication, both as mediators and as linguistic translators. Mok, a man of Chinese origin in his thirties (UK35) who had been in the UK for more than twenty years working in catering, relied on his wife for translation, saying that her English was better than his. But it became apparent that Mok relied on his wife not only for translation and mediation, but also for her persistence in seeking a solution to a debilitating pain in his shoulder. Mok said that his wife accompanied him to every General Practitioner (GP) appointment, resulting in various referrals and treatments, which did not cure his pain. Ultimately, Mok's wife consulted the internet and located exercises for a frozen shoulder on YouTube which effectively addressed the pain.
Women were dependent on their extended networks of other women for support to address their health concerns. Tinka (38 years, originally from Bulgaria) had multiple health problems including eczema, anaemia and abdominal pain. Unable to speak German, she had several unsuccessful GP consultations and refused to attend appointments without a supportive German-speaking friend. Tinka became dependent on her extended female network for support at medical appointments, treatment advice and sourcing medication. Tinka's mother in Bulgaria was particularly important in providing transnational care including medication and advice about traditional treatments.
Persistently making demands of whatever the resources were available, until a solution is eventually identified, was apparent in Ema's account of her son Tiago's diagnosis with schizophrenia. Ema (Lum44; 63 years, Portuguese) had multiple morbidities (including cancer, diabetes, hypertension, gout, kidney stones) that she coped with through bricolaging within the public healthcare system. She described illness episodes when Tiago, originally been diagnosed when employed in Germany, had been violent towards her, but despite these problems, Ema said she never gave up trying to get care for him. Eventually Ema brought Tiago to a doctor who persuaded him to accept treatment in preference to being hospitalised. Ema hoped that Tiago would "go before" (predecease) her, otherwise her daughter would have to take on caring for him, which would be difficult, not least because the daughter's children (Ema's grand-children) had serious health problems too. When asked what she did when she had a health problem, her response was: "I resolve it". Ema and her daughter were both divorced from the fathers of their children who seemed absent from the family support system.
Bricolage as persistence in searching for treatment for a dependent's serious symptoms was described by Annie (UK30), also a single mother. Having arrived as a refugee in 2000, she had British citizenship and spoke English well, but it nonetheless took Annie more than a year to get her son's symptoms diagnosed as tuberculosis (TB). She described her intense concern as her child was unable to keep food down, becoming weak and showing dramatic weight-loss. Her son's low mood and vomiting were interpreted as an eating disorder, and his stomach pain was medicated. The TB was not diagnosed by a doctor in Kenya, while visiting relatives, nor by the GP on her return to Birmingham, when Annie's son could hardly walk. The diagnosis finally came when he was admitted to hospital as an emergency case, with intense pain that led him to say "I am dying, Mum". Annie said that appropriate tests were finally run in the hospital when she refused to let her son be discharged. Once a blood test revealed TB, treatment started and within a month Annie's son was eating, gaining weight and, according to Annie, "he became normal".
Like Annie, Olga (32 years, married and originally from Cameroon), struggled to get her child appropriately treated although she had an excellent grasp of the German medical system. Disappointed by the treatment that her daughter received, Olga knew she had the right to change specialist and, after some research (withinsystem bricolage), she selected a new paediatrician and a new Ear Nose and Throat specialist from whom her daughter received satisfactory treatment. Olga was married and her husband was not mentioned as being involved in the process of finding suitable care for their daughter.

Barriers to Accessing Support
As a single mother, Annie bore the burden of care for her three children, and she was unable to name another source of support, apart from the GP whom she held responsible for the year's delay in diagnosing her son's TB. Olga was a University graduate, employed as a geriatric nurse and confident about navigating the German public health system to get care she deemed satisfactory. As a Black single mother, and care worker, with little formal education Annie's concerns about her son's wellbeing were apparently dismissed by health service providers in a way that she found difficult to challenge.
The challenges of negotiation and re-negotiation were described by Yuming (UK10), caring for her father, who had been diagnosed with dementia 10 years earlier. When first diagnosed, the father could be left with relatives, especially an aunt, but as the condition progressed, this became more difficult. Yuming recalled this same aunt pushing for formal care via social services, to replace the informal arrangement: My aunty, she threatened us with social services, because she looked at our situation and she said, we have some kind of disagreement or argument actually, and she said: "I am gonna report social services about what you are doing with your Dad", and then I thought, okay, I'm going to ring them myself. So I rang…and I explained the situation to them. They came and they assessed the situation.
Even with support from social services, the difficulties of meeting her father's needs had necessitated Yuming giving up her employment. Her father had a Chinese and Vietnamese background, but Yuming always hired people who were "100% Chinese, either Hong Kong or China", because "Chinese people work better for [them]". Yuming said that Cantonese speaking people always responded to her advertisements, and, although her father did also speak Vietnamese, she would not want to hire a Vietnamese carer because: It's too personal because it's too close to our background. Too much part of the community, possibly. You know, communities, they talk, they talk to one person and then…I don't like that, I don't like that. But with Chinese people, they don't know us so well.
Despite difficulties in recruiting suitable carers, Yuming did not want social services to appoint carers on her behalf because "they would be English people, English speakers" and so "they won't be able to do the job very well. They are not ideal for the job". If carers could not speak with her father (who "doesn't speak Englisg"), she would get no respite since she would have to be on hand to translate. The difficulties of recruiting and retaining carers who were the 'sort of people I want in my house' left Yuming with her own health problems (as described in the next section).
According to Aliyah (in her early 60s, originally from Sudan), getting good healthcare was significantly harder for migrants compared to 'native Swedes' for two reasons: first, migrants did not have access to much family support and; second, migrants were discriminated against in the public healthcare system. Married with two children, Aliyah said that before she had arrived in Europe, she had been told that Europeans had weak family relationships: I used to hear before that when you come to Europe, people here don't have strong social relationships with their families-the extended families-they are weak. This is a big lie! I found out that they have strong ties with their families and they get a lot of support. Like the women who work with me, they have their mothers and their grandmothers who help them with their kids: when someone brings a cake to work, their father makes it because he visited her at home and she [the colleague] has small children. So they have lots of social support from their families from their aunts and so on.
While migrants would help each other out, Aliyah reported that the support available to migrants was much more limited compared to "the native people here, they get the support they need". Aliyah felt that she was "definitely affected", that her "choices are limited" compared with "native Swede" colleagues: she said this "affects my life because we don't have that type of support". Although Aliyah had been in Sweden for twenty years, had citizenship, a good job and fluent Swedish, she nonetheless felt discriminated against in terms of primary healthcare, where nurses act as gate-keepers for access to the GP: Since I'm not Swedish, they would think that I don't speak the language well. I know the issue with doctors in this country: the nurses will try, as much as they can, they will try to make you not see a doctor. I have heard about cases-Swedes and non-Swedeswhere nurses will say: "You don't have a problem"…so you don't meet a doctor. So it's up to you and your skills to convince the nurse to let you meet the doctor, to insist.
Having recently been admitted to hospital for an operation, Aliyah had observed that as an African woman, she received a poorer service than other Swedes: "I felt that the Swedes get the established doctors while the foreigners get interns. It's a feeling I got, but like I said I don't have proof. I could be wrong". She said that "because [she's] a foreigner" the healthcare professionals "don't treat [her] like a patient who needs to be treated". The extra barriers for those who look or sound foreign when seeking access to a doctor, to an experienced doctor and to appropriate treatment, all added up to extra work, at a time when help was most needed. As Aliyah explained: I just want to say that we as foreigners when we try to meet a doctor, we need to make an extra effort to get help. I don't think Swedes need to make the same effort because they get the help they need and they ask the doctor and the doctor asks them about many things. I always feel that we need to make more effort, so even though you are ill, you need to be alert and focus so you don't miss anything. When you are ill, you may not be able to talk or focus but you have to…you have to be sick and focused at the same time.
Aliyah describes a double-bind for those who are new migrants and who have a 'foreign' appearance in that their healthcare needs are deprioritised and they also lack the supportive networks and cultural knowledge to negotiate better care via an alternative route.

Lack of Appropriate Support for Carers
It was not only migrant women who found accessing suitable healthcare problematic. Åsa (Got10), a 60 year old Swedish woman, divorced from a man from Eritrea, with whom she had had three children, felt let-down by the healthcare system. Åsa was coping with thyroid and joint symptoms that she attributed to severe mental distress. She explained that two of her adult children had died within 3 months of each other and so she was dealing with great "sorrow" which gave rise to "somatic ailments" due to having "lost [her] children, which [she] took hard". Both deaths had been unexpected and probably related to existing chronic health problems. Swedish born and bred and employed in healthcare, Åsa was confident about accessing services, but found that none of them responded appropriately to her loss. A mindfulness group, identified and paid for by her employer, "had its own agenda" which did not take her needs into account, while a counsellor recommended a self-help book that she felt belittled and even trivialised the scale of her loss, although he sometimes said "sensible things".
Åsa felt that she bore her sorrow alone, reliant on the support of her remaining daughter. Perhaps the death of two offspring could never be lessened by professional services, however responsive and sensitive that provision. Åsa's account reminds us of the limits of professional services to alleviate suffering and also the limits of women's own efforts to overcome restricted resources and services.
Despite her enormous determination Ema, who had had to retire from her cleaning job due to illness, had few alternatives to the public healthcare system, on which she was dependent to meet her many needs. Ema wanted her cataracts treated, but could not afford private treatment, so had to endure the long waiting time in the public system. Maria Alice, a migrant from Santo Tome and Principe, in her thirties, had spent most of her life in Portugal, was employed as a cleaner, and, having been diagnosed HIV positive, had significant healthcare needs. Like Ema, she relied exclusively on the National Health Service (NHS), despite her initial HIV diagnosis being delivered in a humiliating way by a doctor in a public hospital. While the Portuguese NHS is acknowledged as offering the best treatment for serious conditions such as HIV and cancer, Maria Alice nonetheless named specific treatment that she could not access. Specifically, she wished she could get interventions for her failing sight and treatment for depression that she could not afford herself and that were covered neither by her work-based health insurance nor the public healthcare system. These women's accounts describe the efforts necessary to secure suitable healthcare for themselves as well as parents, partners and offspring, despite language and cultural barriers and in the face of having concerns and symptoms dismissed and deprioritised by professionals, which sometimes amounted to racism. The intimate nature of women's care for family members means that an absence of suitable services can have serious consequences for her own health. Maria Alice was a single mother to two young children and wanted to stay healthy for them. Yuming felt she was barely coping with her father's care and described intense anxiety associated with her father's variable ability to cooperate. The anxiety: Coincides with my Dad's dementia, the deeper his dementia, the deeper my anxiety, because then, when his dementia was only mild, then I can cope with it, because he can still communicate and cooperate with me. But it's when his-it's the time when he doesn't cooperate, you know, he refuses to do things; that's when I feel anxious.
The anxiety disrupted her daily life, but she did not take these problems to the GP any longer because he had not offered much in the way of support in the past.
Interviewer: Did you ever bring it [anxiety] up with him?
Yuming: Well I saw him a few times and I said, "Look, doctor, I got this chest pains and I think it is brought on by anxiety and just pressure and stress", and he gave me tablets for it. That's what he does. They don't advise any kind of counselling or whatever….Yeah, it could be stress but then you can't really get tablets for stress, can you? You can't kill stress with tablets.
Cristina, 60 years-old, was born in Angola, had lived in Portugal for decades and suffered with many health issues. She attributed chronic leg wounds to an accidental fall at work. Despite private insurance through her employment as a cleaner, this option was later withdrawn leaving her dependent on the NHS. She said: I don't really know what happened. I was good with the insurance and they took it away to put me back to the public system, something like that….They are treating me as if the problem I have is something I was born with. They say it is a prolonged illness but I was not born like this.
Cristina was distressed that her leg wounds did not heal and that their cause was misunderstood by her healthcare providers. She reported dismissal from her occupation-related private provision when her injuries became a chronic problem and, as a migrant, with little education and having been dismissed from her low-paid employment, Cristina had very few alternatives.

Discussion
Drawing on material from semi-structured interviews and a survey, this account renders visible gendered work undertaken to meet healthcare needs in diverse neighbourhoods. Using a typology of healthcare bricolage, the survey showed women more likely to report bricolage than men and, where men reported bricolage, it was more likely to be within the healthcare system. This pattern, consistent with gendered expectations of women as carers, can be seen among non-migrants and secondgeneration migrants, but was much less marked among migrants across the whole sample, suggesting that the difficulties of navigating an unfamiliar healthcare system disrupt women's bricolage. Unfamiliarity with the health-care system could explain why gender differences seen in non-migrants and the descendants of migrants are reduced in the migrant generation.
The gendered bricolage pattern in our survey data is complex-with the gender differences varying by not only migration status, but also by country, suggesting the influence of the national healthcare systems on the opportunity for bricolage. In countries where the tendency to bricolage was more pronounced (Germany, Sweden), gender difference comes into play with women doing more bricolage than men and perhaps with women bricolaging on behalf of men, although our survey questions did not ask about this. By contrast, in Portugal and the UK where there is much less bricolage overall, gender differences play only a marginal role in explaining the survey data, but are clearly important in making sense of the qualitative interview material.
In the UK and Portugal there may be less opportunity for bricolage because of the configuration of the NHS with limited points of entry for patients to be referred on for care. While those with professional social networks can negotiate better care, for women in unskilled labour or who are unemployed, their networks do not offer the resources with which to bricolage. The Portuguese system offers very few alternatives to the NHS, so while it is possible to contract private services for tests and basic treatments due to public subcontracting, the infrastructure, treatment and physicians available through the NHS for serious diseases such as AIDS/HIV, diabetes and cancer are better than private provision. Women on low incomes have little incentive to bricolage so as to add to the healthcare system, especially when they do not have a social network that could support such an endeavour.
In the UK, as in Portugal, the costs of getting treatment outside the NHS are prohibitive and so women are less likely to add to the system through out-of-pocket treatments or private health insurance, although many state their desire to do so in order to access care more quickly in the face of lengthy waiting times. The universal but rigid healthcare systems of Portugal and the UK do not afford much scope for within-system bricolage, whereas in Germany the insurance-based system requires people to make a choice from a range of potential providers, the cost of which is subsequently reimbursed. Healthcare access is not controlled by providers in Germany and women bricolage more than men, who, if they do bricolage tend not to add to the system. Sweden has a universal tax-based healthcare system, but the range of different entry points and referral routes affords opportunity for within-system bricolage, which is undertaken by women to a greater extent than men, although Aliyah felt it was more difficult for foreigners. Sweden has very little private healthcare provision, limiting opportunities to add to the system.
The accounts of women's attempts to access suitable healthcare for themselves and their family, illustrate how gender, social class and racialisation intersect with migration in the context of national healthcare sys-tems. Women provided translation and mediation services for partners, parents, offspring, persisting in the search for solutions in the face of unsuitable or ineffective treatment. Women's networks, largely consisting of other women, supported them in seeking suitable healthcare through advice and the provision of materials, including medication, where men feature as recipients, but not providers of support. The ability to insist on alternatives to inadequate or inappropriate care cannot be taken for granted, but when in a caring role, even women who were not confident in negotiating with providers, persisted on behalf of others. This persistence could have costs in terms of their own health, as illustrated by Yuming's anxiety. Women's resilience in the face of enormous challenges was illustrated by Ema's assertion that a health problem simply had to be resolved.
Women's persistence in seeking access to care was important for getting suitable treatment, despite concerns being dismissed by professionals. Professionals' reluctance to engage with women's health concerns must be seen in the light of embodied identities as Black, migrant, old, disabled and/or working class women intersecting with the structures of class and race. Specific needs around minority language and ethnic identity were not always respected by service providers and foreigners' healthcare needs were routinely deprioritised, according to some women. The intersections show why no single aspect of identity can be solely determinant. Thus, Black, migrant, single mothers in low-paid work described different outcomes: Annie repeatedly had concerns about her son's health dismissed; Maria Alice accepted treatment through the public health system, despite having been humiliated by a doctor; while Olga was able to negotiate more acceptable treatment for her daughter. Olga's success in getting appropriate healthcare was perhaps due to the German social insurance system that requires bricolage from healthcare users, alongside Olga's university-level education and confidence in German.
The ability to bricolage was enhanced by good social networks and the experience of depression and anxiety hindered it. The limited networks of migrants, as compared with 'natives', were said by Aliyah to constrain women's life choices. Some women who could not access suitable healthcare withdrew from further contact with healthcare provision without lodging any protest or official complaint; a tendency to express gratitude for inadequate services suggestive of a sense of expectation to behave like a "good migrant" .

Conclusion
Healthcare and welfare systems have been important in shifting caring work that had largely fallen on women in domestic settings, away from the private to the public sphere. However, caring remains gendered in ways that are revealed by looking at bricolage patterns across four countries. In order to bricolage healthcare access, women need supportive networks and cultural knowledge and even then, they may not find their way to suitable treatment. The complexities of gender's intersection with racialised aspects of identity, with ethnicity and social class and with the specificities of healthcare systems, make intervening in this area complex, although the unmet need for care and support is apparent. Women have told us stories of the exhaustion and frustration involved in getting healthcare, but the details and contours of that exhaustion are not easy to generalise and so are likely to continue to be under-acknowledged as a hidden feature of accessing healthcare.