‘It’s All Public Anyway’: A Collaborative Navigation of Anonymity and Informed Consent in a Study with Identifiable Parent Carers

ABSTRACT For qualitative researchers seeking the perspectives of people with unusual characteristics or circumstances, compliance with expectations about participant anonymity can be difficult, if not impossible. In the age of internet communications and emerging research methodologies, traditional strategies require ongoing re-examination to ensure cohesion between a project’s ethical framework and its research practice. This paper reflects on the approach to informed consent used in a study with parent carers whose children had high-level support needs. A two-step process of written consent was developed in response to concerns about the possible re-identification of these parents as a result of their highly individual circumstances. This approach acknowledged the potential for identification, and maximised participants’ agency in choosing the level of risk that they were comfortable to accommodate. The paper discusses the researcher’s and participants’ responses to the adapted consent process and recommends that researchers and ethics review committees remain open to the development of collaborative and innovative approaches that are also culturally and contextually relevant, to enable people to contribute perspectives that might otherwise be silenced by the very ethical frameworks that purport to protect their interests.


Introduction
A routine component of human research ethics approval is an assurance that participant information will be de-identified (Ogden 2008;Walford 2005). Anonymisation of data is a fundamental mechanism by which to ensure that participants' right to privacy and confidentiality is protected (Vainio 2012), and is a cornerstone of the assessment of research quality, but in some situations such assurances are difficult, if not impossible, to give. Geographically small communities, communities connected by interest or characteristics, and participants with a public profile, for example, might all be identifiable in certain contexts, even if names are removed from the information they provide. Some participants express a desire to be identified (Wiles et al. 2008), and in some methodologies, such as autoethnography, anonymity is impossible by virtue of the methodology itself.
Although the potential identification of participants is a concern to every researcher, qualitative researchers in particular can be caught between competing pressures to create new knowledge and to protect participants' identities while doing so (Damianakis and Woodford 2012). If it is likely that participants will be identifiable, the researcher is confronted by a 'conflict between conveying detailed, accurate accounts of the social world while simultaneously protecting the identities of the individuals who live in that particular social world' (Kaiser 2009(Kaiser , 1639. Complicating this situation is the issue of anonymity as a 'methodological given' in human research ethics guidelines (Tilley and Woodthorpe 2011, 199). Ethics guidelines commonly have their roots in positivist biomedical research approaches (Doyle and Buckley 2017); in that context, randomised controlled trials, aggregated data, or studies of large, homogenised samples allow researchers to confidently promise participant anonymity, although it should be noted that developments in genomic research have led to emerging concerns about potential re-identification of individual data (Savage 2016). When using qualitative methodologies in work with groups of people who may recognise each other in research reports, or undertaking studies in which the heterogeneity of participants is a feature of recruitment, the capacity of the researcher to make such a commitment may be particularly limited.
This article considers the potential identification of participants in a research project that explored the perspectives of parent carers of children with ongoing high-level support needs. Anonymisation was complicated by the unique characteristics of families, particularly in rural and regional areas. Informed by Kaiser's (2009) suggestion of a twostep consent process, this study demonstrated that concerns about identification should not automatically exclude potential participants or aspects of their views. By engaging parents in conversations about the use of their information at multiple points in the production of data and dissemination of findings, with the level of involvement determined by them, I sought to achieve the 'dual mission' (Damianakis and Woodford 2012, 708) of creating knowledge whilst maintaining ethical standards.
This paper offers my reflection on the experience of operationalising the ethical concept of informed consent, in the context of a study exploring parent carers' perspectives on their relationships with complex service systems. While the consent process was not itself a formal study within the overall research, its value and potential relevance to other research became apparent as the study progressed. Data such as parents' views on anonymity, consent, and the process described here were not produced systematically, but were offered spontaneously by some parents; thus, the learnings presented here warrant further, focused, exploration.
The paper commences with a review of literature addressing concerns about anonymity and confidentiality of participant information. It then describes the consent process used to facilitate the informed consent and research participation of parents of children with high-level care needs (often due to rare conditions) who might be identified through their unusual circumstances or characteristics, and draws on field notes and study information to reflect on this experience. The article concludes with a recommendation that researchers and ethics review boards develop innovative and power-sharing approaches to issues of anonymity and consent, in order to keep pace with rapid developments in information dissemination, and to give voice to people whose perspectives might otherwise be excluded.

Literature review
Current debates about anonymity in research occur in an increasingly risk-averse context (Busher and Fox 2021), highlighting concerns both for participants as a result of identification, and for researchers and their institutions should they be held to account for a breach of confidentiality. Although research institutions create policies and codes of conduct and offer training to early career researchers to assist them to avoid pitfalls as they conduct their studies, the content of such training often develops in an ad hoc way in response to high-profile cases of impropriety (Comstock 2013). Comstock argued for a more coherent, systematic, and philosophically based approach, enabling a considered response to possible future risk rather than compliance with a rigid and possibly burdensome set of rules.
In a challenge to traditional principles of universalism and impartiality, ethics of care suggest a more contextual and collaborative approach (Botes 2000). In the research context, this allows a focus on co-production of knowledge (Kara, Edwards, and Brannelly 2017). In contrast to positivist emphases on researcher distance and objectivity, a collaborative approach acknowledges vulnerability, relationship, and power (Turner and Webb 2014) and prioritises principles of context, dialogue, and reflexivity (Linabary and Corple 2019). Building these concepts into research design can be transformative, allowing ethical research practice with members of marginalised populations that is inclusive, culturally responsive, and presents a direct challenge to existing power structures (Mertens 2021). Ethics of transformation open up opportunities for research methods to respond to a rapidly developing research landscape that features unprecedented ethical challenges through, for example, information sharing via social media. Such challenges require qualitative researchers to be innovative and agile, as they navigate coexistent concerns about justice and care (Banks 2012, 78).
Practical guidance on effective and ethical research methods that respond to the complexity of rapidly expanding and increasingly diverse environments remains underdeveloped, however. Online environments, for example, pose complex new ethical challenges around understandings of privacy (Linabary and Corple 2019). When faced with concerns about participant identification, the literature offers researchers a number of possible strategies. The use of pseudonyms or numerical identifiers is perhaps the most common and familiar approach (Allen and Wiles 2016;Saunders, Kitzinger, and Kitzinger 2015a). Other strategies include creating composite accounts under a single pseudonym (Saunders, Kitzinger, and Kitzinger 2015a); disaggregating data (allocating multiple pseudonyms to data from a single participant) (Saunders, Kitzinger, and Kitzinger 2015a); de-contextualising data by separating discrete pieces of data from their context, even if this reduces capacity to find meaning in the data (Bickford and Nisker 2015); changing non-essential details (Saunders, Kitzinger, and Kitzinger 2015a); emphasising participants' rights to decline questions or withdraw consent, and clearly informing them of potential risks of identification; engaging in reflective research practices to foster awareness of emerging ethical concerns (Damianakis and Woodford 2012); engaging in collaborative research (with other researchers and/or with participants themselves) to increase mutual accountability (Damianakis and Woodford 2012) and offering participants the choice to be identified (Damianakis and Woodford 2012;Mukungu 2017).
None of these strategies is without its ethical or methodological challenges. Some strategies run the risk of leading readers to form unfounded assumptions or make unwarranted judgements. For example, the use of a particular pseudonym might imply gender, age, culture or religion, or could have associations for a reader who has known another person by that name (Lange, Zaretsky, and Euler 2016). Similarly, removing a quote from its context could impair readers' ability to make sense of the information, or invite them to form unfounded generalisations (Bickford and Nisker 2015). Other strategies, such as determining non-essential details for removal (Saunders, Kitzinger, and Kitzinger 2015a), or changing details such as gender, risk affecting the integrity of the data.
A number of factors increase the urgency of addressing ethical concerns about anonymity and consent. The Internet is becoming increasingly influential in practical terms, both as a source of data and as a means of dissemination of findings. The rise of e-citizenship as a concept accompanies challenges to traditional governance practices in society generally, and in research specifically. Electronic 'communities' simultaneously globalise and localise the sharing of information (Gruzd et al. 2016;Polat and Pratchett 2009). Working across geographic boundaries to reach a global audience, these networks may also function as if they were local communities (Polat and Pratchett 2009). This globallocal interface has particular relevance for research ethics, as the potential exists for participant information to be collated from a range of publicly available electronic sources, leading ultimately to identification. As Saunders, Kitzinger, and Kitzinger (2015b) point out, the ethical responsibilities of researchers are further complicated by the possibility that data, de-identified at the time of the research, may be re-identified through this process at a later date, with the possibility of harm that could not have been anticipated at the time of the original consent to participate.
At a deeper level, the role of the Internet in sourcing and disseminating data requires researchers to consider diverse understandings of private and public. This is not a new area; rather, it is built on the foundations of philosophical debates spanning centuries. Neither is it a simple question of a mutually exclusive public/private binary, but a complex mix of considerations (Weintraub 1997). However, the debate has specific implications in an age in which information from a range of sources is widely available, often in an unfiltered format. It also requires an understanding of changing community perceptions about the boundaries between private, public, and the areas between (Gerrard 2021;Linabary and Corple 2019). An example of this shift in community attitudes to privacy is seen in the phenomenon of the 'selfie' digital image, which 'challenges the traditional demarcation between "public" and "private" social life' (Walsh and Baker 2016, 14).
The concept of risk pervades this discussion, at ethical, legal and practical levels. Although the consequences of unintended identification potentially affect researchers and organisations (Dickson-Swift et al. 2008), the risk of harm to participants is especially multi-faceted. Illustrating this point, Kyriakakis et al. (2015) described a study of women experiencing family violence who were also residing in the country illegally. For these women, risks included both the subjective fear and the objective act of deportation; the consequences of identification for these women could be physical, psychological, economic and/or legal (Kyriakakis et al. 2015, 262).
The potential for, and consequences of, participant identification can also affect family members, co-workers, service providers or communities (geographical or of common interest), who might become known once a participant's identity is recognised. The impacts of this affect individuals personally, and depending on the views attributed to the participant, could also affect their relationships with the participant. Furthermore, negative consequences can destroy trust in the researcher, potentially alienating people from engaging in other research projects in the future, even if unrelated (Kaiser 2009).
In considering the possibility of identification of participants, Saunders, Kitzinger, and Kitzinger (2015b) advise qualitative researchers to consider the potential identification of participants when information available through dissemination of research findings is added to information held elsewhere, such as in media stories or on social media sites. So much personal information is now held online that it may be possible to triangulate information held in diverse locations to create an identity even if the individual's name is unknown, a process of 'jigsaw identification' (Saunders, Kitzinger, and Kitzinger 2015b, 125). For example, the friends of a family affected by a rare genetic disorder might launch a fundraising webpage; the same family might later be featured in a television program about the child's condition. Publication of research data, even with the use of a pseudonym, could contain enough contextual description to identify separate pieces of information that, alone, would be considered anonymous. This complexity highlights the importance of research approaches that, while grounded in long-held ethical principles of benevolence, non-maleficence and autonomy, are also able to meet new challenges with innovative strategies.

Methodology
The study's methodology was founded in a commitment to take the concept of informed consent beyond familiar, routine processes; it invited parents to engage actively in confirming consent not only before interviews, but to reconfirm following the interviews and potentially at later dates if ensuring a person's anonymity in the presentation of the study's findings might be more difficult in certain contexts. Having said this, the intention was also to avoid the process in itself becoming onerous or contributing to undue anxiety.
Prior to undertaking this research, I had worked with families affected by rare conditions and had therefore anticipated that the small number and unusual circumstances of the group of parents who were offering their information through this study would require careful attention to potential identification not only of participants but, by extension, others in their networks. As the research question explored parents' understandings of their relationships with service systems, participants were recruited through support groups and word of mouth to minimise the possibility of gate-keeping by service agencies. However, this also increased the possibility that participants would be recognisable to other community members when findings were disseminated. The study's approach to discussing the potential risks and consequences of identification, and the process for gaining informed consent, were informed by these concerns.
I was also conscious that research participants are routinely asked to give their consent to the use of their information at a time when they do not know what will emerge during the interview. Not only are the questions unknown to them, but their responses occur 'in the moment'. If trust and rapport are well-established, it is possible that a participant will disclose something they had not intended to share. Similarly, interview questions might lead them to articulate a view that they had not been aware of themselves prior to the discussion. Routine ethics approval can be expected to cover these possibilities through stated capacity for participants to withdraw their consent at any time, but this places the onus firmly on the participant, who may not be clear about future dissemination of findings. Unanticipated disclosures have the potential to pose ethical dilemmas for researchers, necessitating decisions guided by, but not necessarily articulated within, prior ethics approval. Ensuring that participant consent is current in this fluid environment requires a certain amount of skill to detect emerging ethical concerns and the possible need for review of the existing consent either at the time or later (Saunders, Kitzinger, and Kitzinger 2015a). In response to these concerns, I sought an approach that fostered collaborative research relationships with parents and that maximised their participation in decision-making about the use of their information at multiple points throughout the research process. This stance reflected a step towards transformative ethics (Mertens 2021), to the extent that it was underpinned by a commitment to operationalising familiar ethical principles, such as non-maleficence and autonomy, through building collaborative relationships with parents to increase their opportunities for active decisionmaking.

Method
Twenty-one mothers and six fathers were recruited to a qualitative research project exploring their views and experiences of interacting with complex service systems as a result of their children's health conditions. Recruitment occurred initially through parent support groups, and then via word of mouth. Parents participated in a single face-to-face interview, with an option to create additional visual data, using a portable whiteboard, markers and magnets to map their service systems. Recognising the heightened risk of inadvertent identification due to parents' unusual circumstances, the consent process adapted a two-phase model (Kaiser 2009), which incorporated active engagement with parents in the handling of information at each step, from production to dissemination. Parents were encouraged to choose their level of participation in this process; while the traditional pre-interview consent was required, all further involvement was determined by parents themselves, including the use of the second consent form in full, in part, or not at all.

Consent documentation
The process of informed consent comprised three documents, approved by the University of Sydney Human Research Ethics Committee (approval: 2014/508). The Participant Information Statement and Consent Form (Pre-interview) used a traditional format and wording. A second, optional, consent form was offered to parents to confirm their consent or to document any constraints on how their information might be used. They were also invited to articulate any further involvement as the study proceeded and findings were shared more broadly. The innovation of this approach was not only in the design of the second form, but of the actively power-sharing relationship and discussions held between researcher and parents.
At the first contact, usually by phone or email, I explained the process and rationale to the participant, including some examples of possible uses of the information and hypothetical instances of identification. Potential participants were encouraged to take their time to consider this, and to choose the degree of control they wished to exercise. They were also encouraged to seek additional information as required before making any decisions about further participation in the study. While this is not a novel approach in itself, the potential for identification in this study warranted a more explicit discussion of risks and benefits, and a more active involvement of parents in decision-making not only prior to interviews but as equal partners (to the extent that they wished further involvement) in later steps through to dissemination of findings.
Each initial contact was followed up in writing. This correspondence included a copy of both the pre-and post-interview consent forms for parents' information. One to two weeks later I called each person to answer any further questions they might have, and to arrange an interview should they wish to proceed. The timelines for this phase were flexible, and parents were encouraged to have further conversations with me by phone or email as needed prior to deciding whether to proceed. However, it was important to ensure that this did not in itself become an onerous process; the ability to listen closely and gauge parents' responses played a significant part in tailoring individual consent processes and any follow-up.
A hard copy of the pre-interview consent form was given at the time of the interview, with a further opportunity for explanation and questions prior to signing. The interview was audio-recorded, and an optional visual 'map' of their service systems, completed by parents as they shared their stories, was photographed at the conclusion of the interview. Both of these data forms contained identifying information, however parents were advised that the names of people and places would be removed before being analysed. The map itself was wiped completely clean in the parents' presence at the conclusion of the interviews, with the exception of any non-identifying information written on magnets that might be relevant for future interviews (e.g. 'palliative care').
A hard copy of the post-interview consent form was also given at the time of the interview. I reiterated the earlier explanations, and encouraged parents to take a few days to reflect on the interview if they chose to, prior to returning the form. All parents were offered the opportunity to receive a de-identified interview transcript and image of the map, prior to deciding on any constraints in how their information could be used. In addition, parents were assured that they could complete and revise the form at any time up until the completion of the study and were encouraged to contact me if they had questions at any point. They were also assured that the usual options for withdrawal of consent still applied. Importantly, parents were asked whether they wished to be contacted for consent in specific contexts, such as presentation of research findings to a local support group where risks and consequences of identification might be greater.
The post-interview consent form was optional, and invited parents to specify their wishes about how the information could be used, and what further involvement (if any) they preferred: It is the goal of the researchers at all times to be responsible and trustworthy in our handling of the information you have provided. Now that you have completed the interview, we would like to give you the opportunity to provide us with additional feedback on how you prefer to have your data handled. I, … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … .[PRINT NAME], give consent to the use of my information as follows: (Please tick one of the following statements) ⃞ You may share the information just as I provided it; however please do not use my real name. I understand that others might identify me based on the data, even though my name will not be used. or ⃞ You may share the information I provided; however please do not use my real name and please change details that might make me identifiable to others. In particular, it is my wish that the following details not be shared without first altering the information so as to make me unidentifiable (describe this information in the space below): Throughout each step of the consent process, parents were encouraged to stipulate their preferred level of participation. While every effort was made to deidentify their information, the process stopped short of the assurances of anonymity that are commonly given and, rather, invited parents to collaborate in a dynamic relationship with the researcher in which they determined the scope of their further participation, based on their own informed assessment of risk and consequences. It was equally acceptable for a parent to sign an initial consent and then withdraw from further involvement as it was to request contact and collaboration in decision-making about dissemination.

Findings
All 27 parents lived with unusual or unique circumstances, and accessed very individual systems of support across service sectors, such that their information might be identified in some contexts. For example, one parent had two children who were affected by a condition that had been diagnosed in less than ten Australian children. Other parents lived in small rural towns, and features of their stories would be easily recognised if the research was presented to local carer groups or service providers.
Parents' choices about viewing interview transcripts and completion of the post-interview consent form varied more than I had expected. Conversations with peers at the time had indicated that few, if any, participants asked to see their transcripts. In contrast, 25 of the 27 participants in this study requested a copy, even if they chose to complete the post-interview consent form before the copy was received. Their requests did not appear to be closely related to their confirmation of consent, however; parents chose to complete the second form at various points in the process, even prior to the interview itself. One parent who signed the post-interview consent form immediately after the interview still requested a copy of the transcript and explained that this was not for purposes of modifying their consent but out of curiosity about the process. This conversation illustrated a curiosity about the research process itself, beyond the topic under exploration, supporting the notion of co-production of knowledge. It also suggests an interesting line of further enquiry, to explore more deeply the meaning that participants give to transcripts and other documentary evidence of their stories.
It is a routine expectation of qualitative research that adequate information will be provided to enable potential participants to make an informed choice about their involvement. However, the added step of post-interview documentation foregrounded the process of gaining informed consent, and implementation of the focused explanatory phase resulted in new insights into the way parents in the study viewed their own privacy. It should be noted that the parents who participated did so because they had identified themselves as interacting with 'complex' service systems. For the majority, this meant a prolonged involvement with service providers across a range of sectors including, for example, acute and community health facilities, paediatric and/or adult disability programs, education and social security. As such, it could be expected that they had been asked to sign consent forms in many contexts and for many reasons.
A number of parents had already printed and signed the pre-interview form prior to the interview meeting. In these instances, I offered the usual opportunity for further questions, before verbally confirming their consent to proceed. Although there was significant diversity in parents' choices about whether or not to complete the post-interview form, there was less variation in the actual contents of the completed forms. Two parents noted in the course of their interviews that certain information should be considered off the record, and this information was removed prior to sending them the transcript. Although all but two parents asked to see the transcripts of their interviews, none stipulated further constraints. All parents confirmed their initial consent at some point following their interview.
The opportunity to test the practicality of this collaborative research approach emerged before the study had been completed. Hoping to include a specific image from a very recently completed service 'map' in an upcoming conference, I sought the consent of the parent, Lucy, as her transcript and post-interview consent were still in preparation. Lucy asked to see the image that I was proposing to use, asked about the event, and then provided her consent based on details about the information I would use.

Discussion
The focus of this discussion is a reflection on the researcher experience of foregrounding questions of anonymity and potential re-identification with parent participants, as the implementation of the two-step written consent process was not a study in itself and there was no specific research question or data production. Having said this, parents' responses to the process were noted as they emerged, either as recorded comments within interviews, or as field notes, and these contribute to the discussion. In addition, the consent forms themselves constituted data for this purpose.

Sharing power
Questions of anonymity and informed consent highlight the power-infused relationship between researchers and participants. At a fundamental level, the concerns of participants, researchers and institutional ethics boards are assumed to align: the pursuit of benefit, the avoidance of harm, and the autonomy of participants. As noted previously, I was conscious that I could not confidently assure anonymity for this group of participants in all contexts, despite my best efforts to do so. Kaiser's (2009) proposed process of confirmatory consent was attractive in offering a means of increasing collaboration through explicit discussion of risk as the research continued, and through this increased collaboration, a transformative shift, however minor, in the power held by parents to remain actively involved in decision-making to the extent that they chose.
Having two separate written consent forms may be a novel approach but is not in itself transformative. If unsupported by changes in the way that power is experienced by both researchers and participants, particularly through decisions to include or exclude participant perspectives, a two-step consent process becomes simply another piece of paperwork, likely onerous to both participant and researcher. The importance of the procedural innovation rests in the relationship that underpins it; the commitment of the researcher to actively and intentionally engage participants who might experience consequences from reidentification in choices about whether to contribute their stories and under what circumstances. Implicit in this choice is the option to choose a low level of engagement, signing only the standard consent form and maintaining little or no further activity. Also implicit is a willingness by the researcher to exclude information despite its usefulness, if so requested.
This requires researchers to enact traditional ethical commitments in specific ways, building on familiar conversations about informed consent to open additional dialogue about what information might be shared, when, and with what possible consequence. As I undertook these conversations with parents in this study, I noted that I might have occasion to share their stories with other parent support groups (for example). In doing so, I drew on both my own practice knowledge and my research and practice skills in communication to be open, but to avoid causing unnecessary anxiety. I was conscious that an interviewer's capacity for reflexivity 'in the moment', and metacognition, an awareness of one's own thought processes, is particularly important in this context. Scholars such as Kyriakakis et al. (2015) have described the potential for people identified as 'vulnerable' to be precluded from research out of concern for potential harm arising from their participation. The parents I interviewed could be considered vulnerable, in terms of their life circumstances and their heightened risk of re-identification. It would have been possible (and easier) to turn my attention to parents of children with more common conditions, whose stories are also valuable. Having come to this study from a practice background, it was also important to me to provide a research space in which the voices of this group of parents could be heard. In seeking knowledge that benefits the majority, or 'the greatest good for the greatest number', the perspectives of minority groups may be over-looked (Christians 2011;Mertens 2018). As a result of their long-term engagement with complex service systems, the parents in this study had experiences that were not necessarily representative of 'typical' service users. Their narratives indicate unique relationships with service providers and long-term impacts from policy and service provision that are unlikely to be experienced by families whose interactions with services are episodic and generally time-limited. Ultimately, however, this study demonstrated that parents whose experiences sit outside the prevailing narrative added richness, depth and integrity to the emerging knowledge, despite the uniqueness of these parents' 'outlier' experiences also rendering them the most likely to be identifiable.

Being trustworthy
Virtue ethics assumes the capacity of researchers to act in a 'trustworthy and ethical manner' (Mertens 2018, 2), regardless of culture, relational and contextual influences. Furthermore, Mertens (2018) points out the historical evidence that ethical conduct cannot be ensured by relying on researchers' character alone. From the outset, my philosophical position in this research was as a participating co-producer of knowledge (hence my preference for the language of 'parent' rather than 'participant' in references to this study). Parents' willingness, even passion, to contribute their experiences to knowledge highlighted their trust in me, and the power therefore I held use this information with integrity. I questioned how I, as a researcher, could be confident that my level of comfort with the risk of identification in a given context aligned with that of individual parents, or that our understandings of concepts such as trust and privacy aligned culturally. A single point of consent offered a limited mechanism for checks and balances, in contrast to an actively collaborative relationship. In foregrounding conversations about anonymity, I noticed a converse increase in trust. The openness to acknowledging risk seemed to help forge a collaborative relationship in which we could explore risks and strategies together.

The importance of flexibility
The diversity in parent responses to the optional second consent form also highlighted the value of a collaborative approach. Had parents been uniform in choosing how their information was used, this would support the development of a standardised (albeit revised) consent process. This study demonstrated that when given options to consider an approach that reflected their own circumstances, interests, and choices, parents exercised their power in different ways, from use of the initial written consent only, to confirmation or revision at different points. In this, attention to both process and content strengthened the integrity of the research design. Explicit discussion about the future use of parents' information, engaging them actively in choosing a level of decisionmaking that was comfortable for them, intentionally transferred power to those who would be most affected by the study's outcomes.
Most importantly, attention to these questions not only illustrates the power-infused relationships between researchers and participants but creates a space for reciprocity in the exploration and dissemination of hitherto undocumented knowledge. The findings in this study suggest that questioning traditional practices around anonymity, within the wider activity of seeking informed consent, involves extending a qualitative researcher's repertoire, rather than a replacement of established principles. Foregrounding ethical issues through offering an opportunity to provide written confirmation of consent increased parents' choice and control and deepened their engagement as decisionmakers in the unfolding process of data production and anticipated future dissemination.
Nuance and diversity is not only seen between participants, but can occur within an individual over time. An illustration of this was a parent who had specified more than once during the early part of the interview that he wished to focus on factual information about his experiences and was reluctant to discuss emotionally charged aspects. In the latter part of the interview, however, he shared very sensitive personal information. It is possible that the early assertions represented a mechanism for asserting a boundary between publicly accessible information and that part of his experiences over which he was still able to exert control, but it also suggests an internal re-negotiation of boundaries as the interview progressed, again emphasising the limitations of a single point of consent.
If this interpretation has value, it underscores researchers' ethical responsibility not only to consider the implications of dissemination of their findings, but also to treat participants' boundaries with care and respect when building rapport. This challenge was emphasised by Damianakis and Woodford (2012): Researchers purposefully foster trusting relationships with participants to learn from them and achieve study aims. As this trust develops, especially in studies that involve prolonged engagement, participants might entrust the researcher with deeply personal information. Within a context of ethical research processes, the researcher must decide whether to draw on participants' full or partial narratives to generate knowledge. (709) Reflexivity is crucial to this endeavour, as we must interrogate our own motivations and biases as well as our understandings of the context and culture of two groups of people: those whose stories we hear and from whom we learn, and the audiences who receive our insights and add their own subjective interpretations.
'It's all public (knowledge) anyway' An additional layer of complexity emerges when undertaking research with people whose health or social circumstances require them to provide frequent access to highly personal information. It appeared to me that parents in this study had become so used to sharing information that many seemed almost unaware that this would be more than a tick-thebox exercise. I noted in my field notes after an interview with a father whose child had a rare genetic disorder: (Father) had commented, when I talked to him about consent and confidentiality, that he is not worried about this, because 'it's all public anyway'. It occurred to me (on reflection) that this had been a common response. With a couple of exceptions … I've been far more concerned about confidentiality than the parents. It seems that the concept of privacy itself is socially constructed and that this is particularly the case when there has been contact with multiple services over a long period of time.
It was not uncommon for parents to express surprise when I explained the possible uses of their information. It appeared that most, if not all, had been willing to give an 'informed consent' without a full understanding of the possible implications of this. Kaiser (2009) questioned the possible negative impacts of identifying risks of identification so overtly, however without exception the parents who participated in this study demonstrated the opposite. Rather than refusing further participation, my experience was that they responded positively to the discussion and, instead of increasing anxiety, parents seemed appreciative of the efforts to ensure that the research design was respectful and trustworthy. This suggests that if undertaken sensitively, researchers can use the consent process as a powerful mechanism for building rapport and demonstrating trustworthiness. In this way, we are able to enact ethics of care and reciprocity, approaching the research as a collaborative venture in which interviewer and interviewee work with important information and establish (and sometimes re-negotiate) boundaries around what is to be produced and, subsequently, shared. My own discussions with some parents highlighted that although all were aware that the information would form the basis for my doctoral thesis, they had not considered the possibility that I could be asked to present my findings to a local carers' group, for example. A framework that includes follow-up consent is a step towards ensuring that consent remains informed, current, and relevant to the context within which it might be shared.

Conclusions and recommendations
This study demonstrated the value of looking beyond traditional single-step, pre-interview consent processes, which may not be adequate in studies involving people from small or otherwise identifiable groups. In contrast to research designs that use quantification and participant homogeneity to enable a sound commitment to anonymity, qualitative research can confront researchers with additional methodological challenges. Researchers may be faced with an apparent contest between, on one hand, the responsibility to protect participants identified as vulnerable and, on the other, the generation of new knowledge. Couching this as a binary dilemma, however, echoes positivism; instead, this study demonstrates that a more nuanced, creative and empowering response is possible.
Importantly, the rise of electronic media has highlighted the importance of acknowledging and addressing the limitations researchers face in committing to protecting anonymity and confidentiality of participant information. Researcher control over use of the data, once it is released into the public realm, is diminishing; there are increasing opportunities for third parties to access and share information, and/or combine research data with information from other sources, so that future identification may be possible even if risk-reduction mechanisms are in place at the time of the original dissemination of findings (Saunders, Kitzinger, and Kitzinger 2015b). The two-step written consent process explicates this and shares power with participants in determining the future use of potentially identifiable information. By engaging participants as collaborators in decision-making, not only about production of their information but about its dissemination, researchers are able to work from an ethic of care while remaining faithful to traditional ethical principles.
The study is somewhat limited by the number of parents who participated, and the potential bias associated with their self-selection. Having said this, the inclusion of voices not widely represented in cohorts of more 'typical' parents is central to the study's aims. It should also be noted that the study did not set out to explore the question of anonymity. Rather, this aspect of the research emerged in the course of designing its methodology. As a result, a more focused exploration may provide a more comprehensive understanding of the meanings that parents of children with complex care needs give to privacy and anonymity, together with the meanings they give to documentary evidence (such as transcripts) of their stories.
As a wide range of methodologies become more established in the qualitative researcher's suite, further studies are required to develop a deeper understanding of the emerging ethical implications. Ethical approaches themselves are becoming more diverse, challenging established unilateral power relationships as, for example, the development of ethics of care and reciprocity have demonstrated over recent decades. It is therefore timely to ensure that ethics review boards and researchers work together to respond to an ever-growing range of new challenges. If this does not occur, there is a risk that the voices of people whose information is potentially identifiable, or subjects that are particularly sensitive, may be excluded from research. For social work researchers and their colleagues in other related disciplines, this is not only a methodological and ethical concern, but also a political one, as the exclusion from research participation of people in small communities or whose experiences are highly sensitive, even for their own protection, opens the way for uncontested dominance of other perspectives. An openness by both researchers and ethics review committees to the development of methods that are innovative, collaborative, and respectful of individual culture and context, opens new possibilities for socially just knowledge production.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Pam Joseph. With qualifications in nursing and social work, Pam Joseph's research interests span intersecting aspects of care, community and complexity. Prior to her academic career, Pam worked for a small peer-led genetic support group, an experience that led to her doctoral studies into the relationships between individuals, families, and complex formal support systems.