Working With Community Research Partners to Undertake Sensitive Public Health Interviews: A Qualitative Protocol to Enhance Rigour and Safety

Abstract

Qualitative research interviews are invaluable in understanding participants’ experiences within public health systems and driving systemic change. These interviews can be sensitive in nature, particularly when exploring how certain groups are failed by systems. Considering this, it is crucial to conduct interviews in a way that prioritises the psychological safety of both participants and interviewers. While there is extensive literature theoretical approaches and methods to minimise risk, practical guidance to enhance safety for both research interviewers and participants throughout the data collection process remains limited. This article outlines a protocol for undertaking research interviews within the specific context of children and young people who have experienced housing insecurity and homelessness in Aotearoa New Zealand. The protocol draws on both the wider literature and the experiences of our multidisciplinary research team, with a focus on supporting less experienced qualitative researchers to conduct psychologically safe interviews. Key features of the protocol include: ensuring appropriate cultural responsiveness, providing researchers with opportunities to practice sensitive and/or challenging interviews, carefully considering interview location, forecasting content to participants, and sharing appropriate support resources once the interview has finished. In particular, we describe closure practices that support researchers to ‘leave’ an interview. We reflect on the processes we put in place with the aim of minimising risk to both interviewers and participants when discussing sensitive topics, while maintaining research rigour. By making this protocol available to fellow researchers and community partners we hope to encourage further discussion of best practice for engaging in sensitive interviews in community-based research.

Introduction

Public health research frequently requires the collection of qualitative data that is sensitive in nature—especially where participants may be in a position of vulnerability—with the potential to cause distress to both participants and researchers involved. There is a wide body of literature discussing the theory and methods of different qualitative research approaches to data collection and analysis (see for example (Denzin et al., 2023; Patton, 2002). Participant safety and wellbeing is a key ethics focus and has been extensively researched (Dempsey et al., 2016; Hall et al., 2021; Perri et al., 2018). However, the discussion of practical steps to ensure researchers remain physically and psychologically safe is more limited (Bashir, 2017; Garrels et al., 2022; Kinitz, 2022; Sikic Micanovic et al., 2020; Silverio et al., 2022; van der Merwe & Hunt, 2019). Using multiple interviewers can spread the burden, but also requires careful design (Sattin-Bajaj, 2018). This paper presents a protocol designed to increase interviewer and interviewee safety, as well as building interviewer skills and managing multiple interviewers’ contributions to projects.

Our current programme of research seeks to understand the housing support service needs of children and young people who experienced homelessness and housing insecurity in Aotearoa New Zealand (hereafter, Aotearoa). In Aotearoa, homelessness is defined broadly as severe housing deprivation, where: “people with no other options to acquire safe and secure housing are: without shelter, in temporary accommodation, sharing accommodation with a household, or living in uninhabitable housing” (Statistics New Zealand, 2014, p. 1).

A vital aspect of our current research is undertaking qualitative interviews with young people with lived experience of homelessness, necessitating a considered approach to managing psychological and physical safety and minimising harm for both participants and researchers involved. This paper sets out a brief background to our study, then focusses on our protocol. At each stage of the protocol, we have included strategies and safeguards to ensure interviewer and interviewee safety. We explain the sampling strategy and recruitment into the study, working with our community-based practitioner partners. We describe our approach to training relatively inexperienced interviewers to undertake data collection using a semi-structured interview guide, and steps taken to maximise rigour when multiple interviewers are involved. We outline the ethical considerations, particularly focussing on the safety plan developed for the research, including a resource we developed on closure practices for qualitative interviewers (Appendix 1), and steps taken to enhance cultural and psychological safety for participants. We also detail data handling principles and proposed methods for analysis. In providing this detailed account of our methods protocol, we aim to contribute to the development of a literature base around enhancing researcher and participant safety when undertaking sensitive qualitative public health research.

Context of Our Research

Half of those experiencing homelessness in Aotearoa are under 25, with a higher prevalence rate for Māori (Indigenous population of Aotearoa) young people compared to Pākehā (NZ European) young people (Amore et al., 2021; Ministry of Social Development, 2024). However, children and young people are largely invisible in research on homelessness and housing precarity in Aotearoa (Pehi et al., 2025). International quantitative evidence suggests that being exposed to homelessness as a child is associated with impacts on physical health, mental health, and education (Shinn et al., 2008). Local quantitative evidence shows that Housing First programmes can be a beneficial form of support for young people, showing a reduction in interactions with the healthcare system and an increase in incomes (Fraser et al., 2023). Government has extensive obligations to children (United Nations, 1989), but these mostly assume that parents or caregivers will meet the housing needs of children. Housing support services are not designed, and are often not allowed, to provide support directly to young people (Pehi et al., 2025). It is clear that children and young people experiencing homelessness have unique experiences and needs compared to those of adults, but these are not well understood.

Māori experience disparities in health and mental health, higher poverty rates, lower rates of homeownership, and a higher prevalence of poor-quality housing with dampness and mould, compared to Pākehā (New Zealand European) people (Beukes et al., 2023; Fasoro et al., 2024; Johnson et al., 2018; Oliver et al., 2017). These unequal outcomes for Māori reflect Aotearoa’s colonial history, marked by economic and cultural oppression, which has caused intergenerational material, cultural, and psychological impacts (Cram et al., 2022; Rout et al., 2019). Housing inequity is acknowledged as the result of colonisation and raupatu (confiscation) of land and resources (Lawson-Te Aho et al., 2019; Miller, 2016; Olin et al., 2022). Despite comprising just 17% of the general population (Stats NZ Tatauranga Aotearoa, 2019), Māori make up 33% of the total number of people experiencing severe housing deprivation/homelessness in Aotearoa (Amore et al., 2021).

Te Tiriti o Waitangi (The Treaty of Waitangi) (Manatū Toanga - Ministry for Culture and Heritage, 2023; WAI1040 Waitangi Tribunal Report, 2014), the founding document of Aotearoa guarantees equal care of Māori, however the current system is failing to provide this (Waitangi Tribunal, 2024). A disproportionately high number of Māori whānau (families), tamariki (young children), and rangatahi (young people) are represented in the housing support system (Kukutai et al., 2018), yet their housing support needs are not being met. This is evidenced by half (50%) of new applicants on the housing register identifying as Māori (Ministry of Housing and Urban Development, 2020). Further, of those households that subsequently exit the register without successful placement, 47% identify as Māori (Ministry of Housing and Urban Development, 2020). Our review of existing housing sector data and official information requests from government departments with responsibility for housing support found approximately 323,257 children recorded within the housing support system in Aotearoa, with around 40% (131,776) identifying as tamariki or rangatahi Māori (Pehi et al., 2025).

This study forms part of a larger research programme that aims to explore how to provide the best possible housing support system for children and young people. In this qualitative interview study, we specifically focused on young people who had experienced homelessness, and aimed to explore not only their experiences of the housing support system and changes needed to make the housing support system work for children and youth, but also their aspirations and dreams for their ideal housing futures. We also determined a need for our project to ensure consistency across multiple interviewers. Given the population of study, a secondary aim was to develop methodological protocols for conducting sensitive qualitative interviews appropriate for public health research. We had a particular goal to outline effective interview closure practices and detail how to provide adequate supervision of less experienced researchers to allow them to develop qualitative research skills, while providing the safest work environment possible.

Explanation and Justification of Method

Seeking to understand young people’s experiences of homelessness in Aotearoa before the age of 25, we undertook qualitative semi-structured interviews (Ethics approval reference: University of Otago Human Ethics Committee 24/0149). The interview schedule (see Appendix 2 for interview guide) was designed collaboratively across the team in an iterative process; we began with a brainstorming session noting down all possible questions we wished to ask, and discussing the reasons why we felt they were important. Two members of the team refined these based on study aims, before meeting back with the larger group to discuss. This iterative process occurred several times until we had agreed on an interview schedule that met our goals, was an appropriate length, and, importantly, ended on an aspirational, positive note so that participants were able to end the interview reflecting forward, rather than reflecting back on potentially painful experiences. The schedule began with a broad open-ended question (see Appendix 2 for the interview schedule), became more specific to cover key points such as sources of support, and ended with aspirational questions.

Our interviews were also being conducted by multiple interviewers, and we wanted to ensure rigour in this process. One way of approaching multi-interviewer data collection is to use a framework to ensure that interviewers follow a standardised protocol (to ensure consistency in the data collected) while at the same time leveraging each interviewer’s unique perspective, positionality, and skill set. Rigorous training, procedural consistency, ongoing feedback and self-reflection are considered best practice for multi-interviewer projects (Rosenblatt, 2012; Sattin-Bajaj, 2018). Building on these tenets, each of the three interviewers underwent training (further described below), allowing those who had not been in the field to experience interviewing, and critically, allowing the interviewers and wider team to compare each interview and provide support when needed. This reflexive practice highlighted those areas where interviewers diverged in their questioning style, or their explanations of the topic differed when follow up questions were prompted by the interviewee during role play. A similar process occurred after each of the interviewers had conducted their first study participant interview, with the whole team coming together to discuss and reflect on both the audio recording and transcript. This collaborative accountability will continue into the coding process, with regular check ins and a collaborative codebook used to establish consistency. We note that all interviewers were closely involved in the larger research programme meaning they were well-versed in the subject area. In situations where this is not the case, further background work introducing the area of enquiry and rationale for the interview schedule may be needed (Sattin-Bajaj, 2018).

Ethical Considerations

We have considered both participant safety and researcher safety carefully in the design of this study.

Trauma-Informed Lens

We know from previous research that children and youth experiencing homelessness and housing instability are at high risk of having experienced potentially traumatic events including sexual, physical, emotional abuse and/or neglect (Wiewel & Hernandez, 2022). Complex trauma is characterised by disruptions to psychological, social and biological systems as a result of unmet psychological or physical needs, and/or prolonged and repetitive exposures to traumatic events (Williams, 2006). Additionally, intergenerational trauma (Chou & Buchanan, 2021) and historical trauma (Paradies, 2016) both describe the enduring impact that events such as colonisation have on the health and wellbeing of indigenous and minoritised populations. We recognise that our participants may have trauma histories and talk about their experiences during interviews. We understand that talking about these experiences may cause harm for some people. We have taken a trauma-informed approach to interviews to allow participants to share their stories, while avoiding re-traumatisation and promoting choice, collaboration, and empowerment.

Alessi and Kahn (2023) provide trauma-informed research guidelines, to which our protocol adheres (Alessi & Kahn, 2023). These guidelines include: learning about the impacts of traumatic events and historical trauma on individuals and communities; establishing safety and trust in the research environment; extending safety and trust into the qualitative interview or focus group; knowing when to change course to avoid re-traumatisation in the interview or focus group and; committing to regular and radical self-reflection and self-care in the research process (Alessi & Kahn, 2023). We have taken care to put in place clear boundaries to ensure that we only talk to people for whom it would be safe to revisit their experiences of homelessness. We have also established procedures and resources to make the interviewers and interviewees as safe as possible. Conolly and colleagues (2023) describe how interview processes can help participants with trauma histories to feel heard and to achieve catharsis (Conolly et al., 2023). We attempted to make interviews a positive experience for participants who had trauma histories. We also sought to mitigate the risk of secondary traumatisation of our interviewers (Williamson et al., 2020), by preserving emotional safety and providing adequate tools for self-reflection, support from senior researchers, and ensuring the availability of professional counselling services should they be required (Silverio et al., 2022).

Participant Safety

We have taken care in the development of our interview questions, information sheet, and consent form (see appendices) to ensure that the participants can be as empowered as possible when sharing their stories. We emphasised that participants could choose to skip any questions, take a break, stop the interview, and withdraw from the study before results are prepared for publication. Our interview schedule was semi-structured to allow interviewers to respond to the participant and follow the conversation if needed, however, it was well-developed so that the majority of questions and prompts that we anticipated (based on our previous studies) were available for ethical review. Our schedule included closing/grounding questions to help participants re-enter their day at the close of the interview (Sowicz et al., 2019), and we include offers of opening and closing karakia (ritual incantation or prayer used in this setting for opening and closing the interview in a culturally safe way), to promote cultural safety (Mooney, 2012). We offered the option of a Māori interviewer should Māori participants feel more comfortable with that option. We also followed tikanga-a-hui (protocols and customs for meetings), which included other practices such as sharing kai (food) and upholding participant’s mana (prestige/spiritual power/ authority/ charisma) to encourage cultural safety throughout the interview process (Mooney, 2012).

Interviews took place in the premises of our community partners. This helped to ensure both participant and researcher safety, so that if either party needed support during or after the interview they were in a safe environment. The community setting of the interview was also aimed at mitigating any power dynamics. We provided a list of resources to all participants explaining where they can access a range of support—we have used these resource lists previously and our previous participants have indicated that they found these useful and gained new information from them.

Interviewer Safety

Our protocol is designed to maximise both participant and interviewer safety. We engaged in processes of interviewer training for this study. The training workshop allowed reflection on the kinds of responses we might expect to arise from the interview schedule, and understanding how our interviewers might respond to these, so that we could ensure they are cared for with ongoing supervision. We also took particular care to work through how to best ensure the cultural safety of our Māori interviewer, such as grounding and supervision with an adjacent Māori research team in our department. It also allowed us to discuss options for responding if someone looks uncomfortable or becomes upset during an interview. We discussed closure practices for interviewers (Ihidero, 2020), and they had access to a resource sheet to step through these in the field. All interviewers have been trained in Mental Health First Aid.

An interviewer’s own awareness, conduct and self-care affects interactions with study participants. Personal wellbeing is a precondition for trauma-informed practice and care. Researcher wellbeing, which includes individual and organisational (i.e., the research team), reduces the risk of vicarious or secondary traumatisation (Sheen et al., 2014; Stoler, 2002) fosters empathy and limits the likelihood of destabilising interactions in interviews. As part of our methodology, we drew on theatre methods to create a set of closure practices to guide our interviewers to ‘leave’ the interview (Appendix 1), which we believe will have broader application for qualitative researchers conducting sensitive interviews. Closure is a “closing practice” at the conclusion of an event such as a performance, or in this case interview, to transition between the event to de-role, create distance, and return to self (Ihidero, 2020). We recognise that using such closure practices following an interview can help researchers preserve their wellbeing, separate research from their personal life, and facilitate more effective and sustainable research practices. We use closure practices to provide a ‘bookend’ structure to an interview process allowing the researcher to uphold boundaries that might otherwise cause residual emotional and psychological stress and fatigue. This allows researchers to contain, process, and categorise feelings and emotions that might emerge during an interview. In this way, closure is an active and intentional process or set of processes that offer researchers tools to complete the stress cycle.

All interviews were scheduled so that there were a maximum of two interviews per day per interviewer, a few hours apart to avoid fatigue. After each interview, the interviewer had at least text contact, and a phone call option always available, with senior members of our team with experience in qualitative and homelessness research. Senior team members always gave interviewers a supervisory phone call at the end of each day in the field. They then stepped through closure practices as outlined in the resource sheet. Interviewers participated in regular group discussions with the team where they had the opportunity to discuss any feelings or issues arising from particular interviews. They were regularly reminded that our university’s Employee Assistance Programme (i.e., free counselling for a set number of sessions per year) is available if they feel they need extra support. Interviewers knew that if they felt they needed to withdraw from undertaking further interviews or involvement in data analysis, that was acceptable and something we could manage.

Interviewer Training

Our team for this project consisted of three younger team members, only one of whom had previously undertaken qualitative interviews; three experienced qualitative researchers, two of whom have undertaken several interviews for studies with people lived experience of homelessness and/or mental illness; two senior research project leaders; two who have worked in housing support service provision; and one who has extensive experience in theatre performance and pedagogy. Recognising that the interviews for this project were sensitive in nature, with relatively vulnerable participants and an increased potential for secondary traumatisation to the interviewers, we first provided the opportunity for the interviewing team to take part in ‘lower stakes’ interviews on a less sensitive project together. This also helped to develop the multiple-interviewer team and encourage reflection about how to manage rigour with multiple interviewers using the same interview schedule for our projects.

Before beginning interviews on this project, interviewers undertook a mock interview with a team member who has a background as a professional theatre practitioner. They acted as a mock participant, akin to standardised patient training used in health care for training and assessment as a bridging tool (Hillier et al., 2023). They improvised encounters with trainee interviewers based on background details of the mock participant they had created from high-level details we gave about the purpose and content of the interview schedule. Each interviewer proceeded through the full interview process, including going through the information sheet and consent form, draft interview schedule, and closure practices. The interviews were audio recorded, and reviewed by the qualitative project team (i.e., the authors).

We then held a wānanga (workshop) to reflect on the data and stories that arose during the interview, the use of prompts, how the interviewers felt about using the draft interview guide, and any differences between the interviews or reflections on interviewer style. We discussed room setup and interviewer positioning, body language, and the use of karakia and how to help participants with grounding themselves to re-enter their day after the interview. We also discussed the closure practices for interviewers (see Appendix 2), and workshopped how to use these, prompting the interviewers to consider their plans to intentionally ‘leave’ each interview. We created a process for booking interviews that included senior project leaders in the dairy booking so that the interviewer’s location was known and phone calls to debrief were scheduled for after interviews. Ongoing internal mentoring support of interviewers and the safety plan, including self-care, was formalised and put in place.

Sampling/Recruitment

While our intention was to speak with young adults who have experienced homelessness or housing instability, we only interviewed those who were currently in a stable housing situation, to enable them to talk reflectively about their experiences from a position of emotional distance. Participants were eligible if they were between 16 and 30 years of age, with experiences of homelessness before age 25. We intended to achieve a minimum of 40% of our participants identifying as Māori to ensure that our results and strengths-based recommendations reflect the experiences, needs, and aspirations of rangatahi who have experienced homelessness (as per the official NZ definition of severe housing deprivation/homelessness) during their childhood and youth. Our community partners were supportive of this recruitment goal. One of our interviewers is Māori, and on recruitment our participants were asked if they would prefer to speak with a Māori interviewer. Given the discrimination experienced by Māori in the housing system and our awareness of this, we intended to provide a culturally safe option for interviewees to engage if they felt more comfortable with a Māori interviewer (Samples et al., 2012).

Our participants were recruited through community and Māori organisations providing housing support services for those experiencing homelessness and housing instability. They were referred for invitation into the study based on the knowledge of our community partners of their current situation and their childhood and young adult experiences of homelessness. Because they were being supported by our community partners, we were confident that they were receiving active support. As a duty of care, we provided a list of resources and contacts for participants to refer back to after the interview if needed. The continued support being provided by our partners also allowed our researchers to leave the interview without feeling responsible for continued care of these potentially vulnerable participants.

Potential participants were briefed by our community partners, who talked to the participant about the process of the interview, what would be expected of them and any questions or concerns they had. In some cases, the community partner contacted the study team to relay the young person’s questions in advance, so that we were able to address any concerns before the interview. Our community partners invited participants who were eligible and interested and arranged a time for the interview (in consultation with the interviewers), which were held in our partners’ offices or community rooms. On arrival, participants were greeted by the community partners who already had an established supportive relationship with them, and these trusted individuals introduced the participants personally to the interviewer, helping to establish rapport. The participants were provided with a contact number for the interviewer’s study phone (institution rather than private phone, with a separate number) in case they needed to get in touch.

Interview Process

Once the participant was introduced to the interviewer by the community partner, the interviewer explained their role, and the intentions of our research. At this point they explained the process of research, why we audio record interviews, and how we would handle and protect their data after the interview. The aim of this conversation before starting the interview was to answer any questions or address any concerns the participant may have had and to build rapport with the participant. This conversation served to partially offset power imbalances between the participant and the interviewer, by sharing personal information and motivations for the interviewer to be involved in the study, and emphasising the participant’s choice and control.

At the start of the interview, the interviewer asked the participant if they would like to open and close with a karakia. During the interview, the interviewer offered food, water and breaks, and reminded the participant that they did not have to share anything they were not comfortable with sharing. At the end of the interview, the participants were offered to stay in the space for as long as they would like to, and with the permission of the interviewee, made the support workers present in the office aware that the interview was finished so that support workers could check in and farewell the participants before they left.

Informed Consent

The information sheet and informed consent discussion covered the purpose of data collection and the research more broadly. Participants were given the option of being kept up to date with the project, including being sent a plain-language summary of the results. Interviewers explained how their details and names would be kept anonymous, but that we will continue to meet with our community partners to discuss the results with them and publish a research summary on our website afterwards. Participants were told that there was a window of time in which they were able to withdraw from the study or remove any information that they had shared that they were not happy to be included in reports. They were also made aware that we will publish and present this work in academic journals, at conferences, and engage with the media and social media to ensure the broadest possible dissemination of the research.

Shared Transcription

Interviews were transcribed verbatim and deidentified by members of the named team to be used for analysis (e.g., using ID codes, and names of people or places removed/given ID codes or descriptors e.g., family member, friend, teacher). Transcribing was shared by the interviewers so that they did not transcribe their own interviews. This allowed for reflection on interview style and helped the interviewers to continue to learn from each other and grow their qualitative research practice. It also helped to relieve the burden of single custodianship of a participant’s story where more sensitive material had been discussed. Interviewers warned each other about the nature of sensitive stories so that they could ask not to transcribe these if they preferred. We also requested that interviews were transcribed in the office, rather than when working from home, to enable separation of the interview content from our researchers’ home environments. Interviewers knew that, if necessary, senior members of the team could perform transcription and analysis.

Confidentiality

With regards to confidentiality, we have and continue to make every effort to ensure that data that we share will not identify individuals. The wider project team who may be involved in publications will only see selected deidentified data. As this includes our community partners, there is some risk that if they know a particular participant well and knew that they had taken part in the interview study, they may be able to identify them from their quotes. However, the community partners have worked closely with the participants to support them into more stable housing situations and continue to support them, and we are confident that there will be no detrimental changes to the relationship between the housing support providers and the participants as a result of this research. Participants were made aware in the consent process that these providers would be receiving anonymised results of the research. The shared goal of the study is to highlight the challenges that children and young people experiencing homelessness face and to suggest recommendations for designing a more supportive housing system that will help to meet their aspirations. Our community partners have been involved from the development stage of the research. For these reasons there is minimal risk from the community partners identifying the participants from interview quotes.

Analysis

For this project we are using reflexive thematic analysis methods for data analysis (Braun & Clarke, 2006, 2019, 2022). Consistent with our consideration of researcher safety, coding is being undertaken by two of the interviewers, with input from the wider team. An initial codebook was developed after the first five interviews were completed and transcribed, with a core group of the team, to create a common language between the coders for the codes we expected would arise. Our team are now undertaking data analysis using an inductive approach to generate themes.

Summary

We have described in detail the methods protocol being used to undertake qualitative research with young people who have experienced homelessness before the age of 25 in Aotearoa. Our protocol contributes a discussion of steps taken to equip multiple and relatively inexperienced researchers to undertake qualitative interviews with attention to ethical principles that promote rigour and maximise safety for both interviewees and interviewers. Our project benefited from careful attention to the development of an interview guide, the use of role play in training, using detailed closure practices to allow interviewers to ‘leave’ the interviews, and a strong focus on mitigating harm to both participants and researchers. This paper adds a practical example of how to undertake public health interviews in populations that have experienced trauma and provides a guide for closure practices for interviews. We hope that describing our protocol encourages further discussion of best practice for engaging in sensitive interviews in community-based research.

Supplemental Material

Supplemental Material - Working With Community Research Partners to Undertake Sensitive Public Health Interviews: A Qualitative Protocol to Enhance Rigour and Safety

Supplemental Material for Working With Community Research Partners to Undertake Sensitive Public Health Interviews: A Qualitative Protocol to Enhance Rigour and Safety by Kimberley Clare O’Sullivan, Mary Buchanan, Lori Leigh, Rachel Kowalchuk Dohig, Tīria Pehi, Matthew Jenkins, Nevil Pierse, Lynda Ryan, and Brodie Fraser in International Journal of Qualitative Methods