How to undertake research with refugees: lessons learned from a qualitative health research programme in Southern New Zealand

Introduction

Refugee health is an issue of global importance with significant implications for health systems and the health outcomes of refugees. Refugees often face significant and complex mental, physical and social health challenges (Hadgkiss and Renzaho 2014) and are often more likely to have increased morbidity, poor health outcomes and a reduced life expectancy compared to host country populations (Williams and Thompson 2011; Hadgkiss and Renzaho 2014; Chen et al. 2015; WHO 2022). Coming from countries in situations of long-term war or conflict, refugees have often experienced severe trauma which is commonly associated with cumulative vulnerability including deprivation, unhealthy environmental conditions and disrupted access to health care (Williams and Thompson 2011; Smith and Daynes 2016). They then face considerable barriers in accessing healthcare services and navigating through health systems in their host countries (Cheng et al. 2015; Mangrio and Sjögren Forss 2017), frequently experiencing gaps in access, poor coordination and service fragmentation (Joshi et al. 2013; Suphanchaimat et al. 2015; Phillips et al. 2017). The fit between health services and the needs of refugees is known to be challenging, and models of care for refugees remain poorly understood and researched (Joshi et al. 2013). Ultimately refugees can experience a decline in health status after arrival, even in resource-rich nations (Gabriel et al. 2017).

Clearly there is a pressing need to consider refugees’ experiences of health care to enhance the quality of care delivered, improve access and improve refugee health outcomes. However, engaging refugees in health research is known to be practically and methodologically challenging (Gabriel et al. 2017) and evidence on how to foster culturally appropriate research practice in this context remains scant. Much of the research about refugee health care focuses on health provider perspectives (Richard et al. 2019; Kennedy et al. 2021). A comparatively smaller body of research prioritises refugees’ perspectives (including other projects within Aotearoa New Zealand (Cassim et al. 2022; Coppens 2024) and others which have co-designed services and programmes (Power et al. 2022; McKeon et al. 2024)). This article adds to this smaller body of research prioritising former refugees’ perspectives and its novel contribution is the focus on reporting key lessons learned about the process of conducting such research.

This paper reports on our experience conducting several qualitative research projects over 4 years with former refugees in Southern New Zealand. Its aim is to reflect upon the methodological particularities of qualitative research with former refugees and to identify key processes of successful engagement – with the hope of guiding others who wish to undertake more research in this area of significant importance.

Methods

The primary study within which this methodological paper is nested, aimed to explore former refugees’ engagement with, access to and navigation through New Zealand’s healthcare system across two resettlement sites in Southern New Zealand. This qualitative exploratory research relied on a community-engaged approach. Service providers helped us develop research and interview questions and facilitated contacting former refugee communities with invitations to participate. The first author (MG) conducted 20 open-ended interviews with 31 former refugees in their homes. Participants came from four countries (Syria, Afghanistan, Colombia and Iran) and spoke three languages (Arabic, Farsi and Spanish). Interviews were always supported by an interpreter and lasted ~2 h. All interviews but one were audio-recorded and MG also recorded fieldnotes in spoken and written form after each interview. The interviews and fieldnotes were transcribed prior to thematic analysis (Miles et al. 2013). All research team members reviewed and discussed the transcripts; MG coded them with ATLAS.ti software. We brought early, anonymised findings back to our community partners (service providers working with refugees) and to two of the participating households for verification and feedback.

For the purposes of this methodological paper, we utilised our research team’s discussions of the process of community engagement that preceded fieldwork and the fieldwork experience itself. The fieldworker is a social anthropologist and practiced the continuous reflexivity that is good practice in her discipline by considering her own world view and assumptions and how aspects of her identity impacted upon the fieldwork and analysis. She took copious fieldnotes after each interview. The research team met regularly to review and discuss the process, the fieldnotes, the interview transcripts and the relevant literature. They reflected upon the specific particularities encountered in conducting this research. In this paper, we provide illustrative vignettes to elucidate some of these particularities and to offer guidance about conducting qualitative health research with former refugees. All names are pseudonyms, countries of origin are not listed to maintain anonymity and other identifying characteristics have been changed.

Results

We found that successfully engaging with former refugees in qualitative health research entails the following processes: establishing relationships, recognizing interpreters as cultural brokers, having available a responsive ‘suite’ of methods and finding meaningful ways to communicate. As shown through our vignettes, these categories are not discreet but inherently interwoven.

Find meaningful ways of communicating

We found it was important to be able and willing to communicate in a variety of ways – in person, by phone, via WhatsApp and by other social media – adjusting to whatever means of communication the participant used and effectively creating a ‘customised communications’ approach with each participant. For example, often living on tight budgets, some former refugees did not have credit on their mobile phones but used a WiFi connection and WhatsApp to communicate by text, voice messages or video.

All interviews included heavy reliance upon an interpreter. But outside of the formal interview, language needs varied. For those who spoke some limited English, but did not read, MG could voice-record or text simple messages. But for more complex messages and to allow participants to communicate in their own language, the process was this: MG would text the interpreter what she would like to say to the participant. The interpreter would translate this into either a text message or, for the illiterate participants, a voice message. The participant could then respond – in written or voice messages, in their language. The interpreter would translate to English for MG.

This may sound laborious at best, or even a barrier to communication and research participation at worst. But with a bit of trial and error and persistence, it worked. If the interpreter and participant were comfortable with their contact information being known to each other, a group chat would be established which made it more streamlined. Furthermore, as MG found out, warm, direct communication was possible with participants even without a common language or literacy – through emojis. Fig. 1 provides a screenshot of MG’s phone of exchanges with a participant who did not read or write in her own language or in English.

Fig. 1.

Screenshot of messages with a participant who does not read or write in her own language or in English. (When actual information needed to be conveyed an interpreter facilitated.).

The success of our careful approach became evident at the end of this research project when we approached several participants to be part of our next project. All but one agreed. As former refugees can be ‘hard-to-reach’ participants, this sort of successful engagement is a crucial step in mitigating the health inequities they face.

Discussion

The health inequities of and complexities for refugees and former refugees makes health research with this population imperative. Yet research with refugees can involve ‘some of the most difficult ethical and methodological challenges in the field of human research’ and the principles that should guide the researcher are less clear (Sieber 2009). There are ethical complexities associated with marginalised populations – such as past and present trauma and the inability to exercise a degree political/social/economic agency without being at the mercy of others (Mackenzie et al. 2007). Institutional ethics committees are often more specialized towards quantitative biomedical research and may have less understanding of the subtleties and complexities inherent in qualitative social science research (Mackenzie et al. 2007). However, the importance of ethical research – to address the needs and inequities of any marginalised population – is obvious, and so the discussion must continue. We offer here our experience and, upon deep analysis and reflection, our ‘lessons learned.’ The careful, unrushed steps we took and would recommend when conducting health research with refugees are: establishing relationships (this takes time), recognising interpreters as cultural brokers, having a flexible ‘suite’ of approaches for quick customisation and finding meaningful ways of communicating.

Our research team members established relationships with many in the greater refugee community over a period of months and years before beginning this research – service providers, volunteers, health professionals and more. Sometimes these community members are referred to as ‘gatekeepers’ of the refugee population who hold important background knowledge and information, helping the researcher grow necessary understanding and skills needed to work with marginalised populations (Mackenzie et al. 2007; Sieber 2009; Kabranian-Melkonian 2015). In our experience, a particularly important step was locating and forming connections with trusted interpreters. All of this is done before any fieldwork takes place. Building these relationships can be extremely time-consuming and tricky for some researchers to balance with funding and institutional requirements, however, we found this approach was imperative.

Interestingly, while these community members could help build bridges into the refugee community, when meeting with refugees it was also sometimes necessary to differentiate ourselves from these ‘gatekeepers’. In some cases, for example, a former refugee might have had mixed experiences and therefore mixed feelings about an organisation, so identifying ourselves as independent researchers (who would keep confidentiality) was then important. At other times, former refugees would think or hope that we researchers had power or connections that could help them, and explaining our role and limitations was also an important part of the consent process (see also: Mackenzie’s et al. (2007) discussion of honest declaration of researcher’s role and limited power).

Interpreters are active in the production of research – they may bring a complex mix of power based on ethnicity, class, race and gender (Temple and Edwards 2002; Jacobsen and Landau 2003; Mackenzie et al. 2007). When MG first contacted potential participants (typically by phone) the interpreter who helped with that phone call was typically the one who would be attending the interview as well. We were therefore able to tell the potential participant, in advance, who the interpreter would be. In small communities, like the ones where we were working, this was an important aspect of trust, confidentiality and emotional safety of the participants. Sometimes the interpreter was already known to the participant (who could then consent to participate, or not, knowing who the interpreter would be). Where the interpreter was not known, MG and the interpreter gave some identifying characteristics of the interpreter – for example, letting the Colombian participants know that the interpreter was NOT Colombian. Other researchers found, like we did, that using the ‘wrong’ interpreter could come in many forms – including in many cases using one from THE SAME community as the refugee (Jacobsen and Landau 2003; Kabranian-Melkonian 2015). Gabriel et al. (2017) wrote about how the right choice of interpreters is tricky because sometimes ‘outsider’ interpreters were met with some suspicion, but more often ‘insider’ interpreters – from the same community as the participant – caused the most hesitation due to confidentiality concerns or political backgrounds. Having the appropriate interpreter is paramount and is the kind of nuanced knowledge that can only come to relationship building before fieldwork begins.

Gabriel et al. (2017) found that having a time and place of familiarity, convenience plus childcare arrangements, made refugee participants much more likely to participate in research. MG always allowed the interviewee to choose a time and place of convenience for an interview and all participants chose their own homes. This had the added benefit, then, of the participant not having to find transportation or childcare. Allowing family life to continue is more than a matter of convenience and comfort for the participant, as important as these might be. MG has extensive experience conducting qualitative interviews in participants’ homes and has come to advocate for this as appropriate and respectful practice where appropriate. Embracing interruptions, family members or friends coming and going, sharing food or meeting the family pets, means it can be difficult to quantify numbers or demographic details of participants. We argue for embracing this ambiguity as lived reality; the significant strength of this approach lies in the detailed, personal accounts that emerge – and are witnessed – by ‘being there’ in someone’s home and ongoing life (George et al. 2021). This seemingly incidental ‘being there’ is an indispensable technique in achieving an ethical, flexible, humanistic approach as well as gaining nuanced data (Monaghan and Just 2000; Madden 2010).

However, it means that the fieldworker must be flexible, compassionate, mature and humble, ready to think on her feet rather than relying on a rigid protocol. The fieldworker must also continuously report back to the broader team about the unique set of circumstances of each interview and choices made. This respectful approach extends to following the lead of participants in matters such as being recorded or signing consent forms. Several researchers have found ‘signing a form’ to be particularly tricky for refugee populations – many of whom may be illiterate or concerned about signing forms, even feeling it is dangerous to do so (Kabranian-Melkonian 2015). This can be tricky for ethics approval processes which may prefer a more blanket, standardised approach (Mackenzie et al. 2007). Ensuring the safety of all participants, especially when dealing with vulnerable populations with complex lives, however, requires a methodological approach and a fieldworker capable of customised responsiveness to each participant.

The theme of flexibility and adaptability extends to communication outside the official interview – before and/or after it. People communicate in a variety of ways for a variety of reasons, and especially when the researcher and participant do not speak the same language, a variety of methods might be needed. We found no literature specifically discussing the use of emojis and stickers with former refugees, but in our case, we found this was not just a matter of enabling communication in a perfunctory sense, but also about making connections and finding shared meaning.

This study has limitations. This paper discusses a number of strategies that we found to be successful in engaging former refugees in qualitative health research, but we note that much more would be required to constitute a co-design approach (Moll et al. 2020; Power et al. 2022; McKeon et al. 2024). Furthermore, it is a small, qualitative study with 20 former refugee families in two of New Zealand’s smaller resettlement cities. Their experiences, and the experience of doing research with them, may differ for those in larger locales with more diversity and resources (though with potentially different challenges.) It is also a broad-level look at the experience of conducting research with former refugees who are not a homogenous group but our sample size and scope does not allow for differentiation by participant ethnicities, ages, genders, etc. Our discussion above also assumes the researcher does not share the same language and cultural background as the participants. It also was conducted with former refugees who have officially been resettled, rather than those still in transition.

Conclusion

More research that prioritises the voices of former refugees themselves is needed – doing so requires time and great care. Our hope is that our analysis of our own experience enables other qualitative health researchers to engage ethically and successfully with former refugees so that, together, we can contribute to the mitigation of inequities and improve health outcomes.