Ethical considerations in sensitive suicide research reliant on non-clinical researchers

Abstract

The impact on researchers of working with sensitive data is often not considered by ethics committees when approving research proposals. We conducted interviews with eight research assistants processing clinical notes on emergency department presentations for deliberate self-harm and suicide attempts during a suicide prevention trial. Common experiences of working with the data included feeling unprepared for the level of detail in the records, being drawn deeply into individual stories, emotional exhaustion from the cumulative exposure to the data over long periods of time while working alone, and experiencing a heightened awareness of the fragility of life and the need for safety. The research assistants also reported on some of the strategies they had developed to cope with the sensitive nature of the data and the demands of the work. The ethical implications for suicide research reliant on non-clinically trained researchers exploring sensitive data are considered. These include the need for research leaders and ethics committees to be aware of the potential adverse mental health impacts for these researchers examining sensitive data and to make appropriate arrangements to minimize the mental health impacts of such work.

Introduction

Researching suicide and self-harm is considered a ‘sensitive research topic’, in that conducting such research poses the potential for harm to participants of the research and/or the researchers involved (Dickson-Swift et al., 2006, 2007; Lee, 1993). In sensitive research, such harm may be due to the elicitation of ‘a range of intense emotional responses’ or other ‘emotional costs’ to those involved (Cowles, 1988; Dickson-Swift et al., 2008a, 2008b, 2009; Hubbard et al., 2001; Johnson and Clarke, 2003). To date, the key concern in sensitive research has been on the impact on researchers engaged in in-depth interviewing, with numerous accounts of field researchers’ reflective experiences (Coles and Mudaly, 2009; Hess, 2006; Lalor et al., 2006; Rager, 2005; Rowling, 1999; Warr, 2004). Specifically, this research has focused on the experiences of clinically trained health professionals/researchers who work with at-risk individuals. In the area of suicide, the international literature has consistently shown that a patient suicide can adversely affect the well-being of the professionals involved in their care, despite their formal clinical training (Chemtob et al., 1988; Hendin et al., 2004; Linke et al., 2002; Midence et al., 1996; Ruskin et al., 2004). Consequences may include burnout, guilt, shame, anger, helplessness or general stress, and these reactions have been shown to be more intense in less experienced clinicians (Chemtob et al., 1988). Research involving clinically trained health professionals working with people who self-harm has also shown that clinicians can experience significant distress, ranging from feelings of shock and anger to helplessness, inadequacy and guilt (Agulnik and Palmer, 2008; Fish, 2000; Thompson et al., 2008). These effects can be compounded by a lack of formal support and adequate training to deal with self-harming patients (Gibb et al., 2010).

Far less is understood about the impact of stories of suicide and deliberate self-harm on non-clinically trained researchers, and the risk of harm carrying out this type of work may not be considered by ethics committees when approving research proposals (Dickson-Swift et al., 2009). Although such researchers typically do not have any direct interaction with those at risk of suicide and self-harm they are often required to examine sensitive data documenting cases of suicide and self-harm often in the form of case files, clinical notes and coronial data (Fincham et al., 2008). In examining such data, researchers are exposed to narratives and intimate details of the trauma experienced by those who have harmed themselves (Dickson-Swift et al., 2009). Such exposure to stories of trauma is not limited to cases of suicide and self-harm, and includes case files documenting rape (Alexander et al., 1989), child maltreatment (Milling Kinard, 1996) and other traumatic deaths (Driscoll et al., 1997). Working with these different case files has been shown to affect researchers in a number of ways, including psychologically, emotionally and physically.

During the course of a community intervention trial for suicide prevention, we found that some unanticipated issues arose for non-clinically trained researchers working with emergency department clinical data on self-harm and attempted suicide, suggesting that such work resulted in the researchers feeling adversely affected to varying degrees. We report on some of the common adverse experiences of these staff based on interviews one year after the trial began and some of the strategies used by them in order to cope. To our knowledge, this is the first investigation of some of the impacts on non-clinical researchers from examining sensitive emergency department (ED) data on self-harm presentations.

Context

During a large community intervention trial for suicide prevention in New Zealand (the Multi-level Intervention for Suicide Prevention Study), eight (seven female and one male) non-clinically trained research assistants (RAs) collected and processed data from clinical notes of presentations to ED. All were university educated and their backgrounds and previous research experience varied (administration, policy analysts, marketing, freelance writer, physiotherapist). Only two had previous experience in an academic institution, and none had any formal clinical training.

The key task for the RAs was to identify all individual cases of Deliberate Self Harm (DSH) and suicidal ideation presenting to emergency departments in their region during the course of the trial. Cases of DSH included all presentations of attempted suicide and self-harm, including substance abuse, self-poisoning and self-injury (Horrocks et al., 2003). The data were obtained from a combination of handwritten and electronic records, including clinical notes, clerical notes and mental health records. The records often contained highly sensitive narratives/information on traumatic events and life histories of the patients (such as abuse and violence). Distinguishing cases of DSH from accidental harm was complicated by the fact that symptoms and diagnoses were often documented in ways that did not clearly reflect the suicidal/DSH intent, experience and/or behaviour of an individual. Thus RAs were frequently required to undertake meticulous examination of the clinical notes. This required examining in-depth information in the electronic databases, and reading through all related clinical notes held on the patient by the hospital for generalized diagnoses and descriptions (e.g. abdominal pain, wrist trauma and burns), surgical notes, and assessments by Mental Health Service staff.

Method

At the end of the first year of the trial and using a semi-structured interview guide the eight RAs were interviewed by an independent researcher about their experiences of collecting and working with the sensitive data. The guide was designed after reviewing existing literature on researching sensitive topics and consisted of open-ended questions and prompts which asked about their experiences of working with the data and the perceived impact of the work on them. The interviews lasted between 50 and 100 minutes and were recorded. The interviews were deidentified, transcribed verbatim and coded using QSR NVivo 8 software. Each transcript was annotated line by line, using constant comparison between transcripts with open coding (Gibbs, 2002). This allowed for identification of initial codes, which were revised and added to after careful reading and re-reading of the transcripts to ensure that all ideas from the data were encapsulated. When all data had been initially coded, the different codes were sorted into potential themes and all the relevant coded data extracts collated. These themes were reviewed against each of the original transcripts to determine fit with the original data and to ensure they best described the experiences of the RAs (Attride-Stirling, 2001; Braun and Clarke, 2006).

Findings

The RAs talked freely about their experiences. Overall the RAs found the data collection process stressful. Experiences common to all the RAs included: feeling unprepared for the level of detail and the sensitivity of the information in the data; being drawn deeply into individual stories; and the stress associated with the cumulative exposure to sensitive data over a long time period. Some also reported frustration with the health system and particular cases of DSH. Others noted having a heightened sense of the fragility of life resulting in an increased awareness of the need for safety in their everyday life. In response to the difficulty of the data collection process, the RAs also talked about how they had to develop their own strategies for coping with the work.

Detail and sensitive nature of the data

None of the RAs felt emotionally or psychologically prepared for the sensitive content and detail in the case files, describing them as ‘shocking’, ‘distressing’, ‘grotesque’ and ‘disturbing’. Most talked about the difficulties they had reading graphic descriptions of physical bodily harm, whether intentional or accidental. Some RAs described having to ‘make themselves’ read physical descriptions and ‘pretending’ they were not ‘hearing’ it. A number of RAs read clinical notes with direct quotations from patients about wanting to die or hurt themselves. Others read suicide notes and emotive details of sexual and physical abuse which were ‘horrific’. Some of the RAs found they were visualizing images of such events long after reading a particular file, one RA described having vivid dreams about particular cases ‘like it was something from a movie’. This occurred especially with cases that involved serious assault or significant pain such as poisoning or burning. All the RAs reported that the reading of the intimate details of clinical notes was extremely challenging, and felt they were unprepared for how much personal information about people they had access to. For example, one RA stated ‘you see everything about people … nothing[s] left to the imagination … you know if you had a conversation with this person, you wouldn’t find out as much as you do about the whole situation that you do in the clinical notes … everything’s down there on the page’.

Being drawn deeply into individual stories

All the RAs felt it was impossible not to be drawn into individual stories. In order to distinguish the cases of DSH from the non-deliberate, RAs often had to read extensively about an individual patient’s history, especially those with repeat presentations. The amount of information in the patient histories often allowed the RAs to piece together the events leading up to the self-harm drawing them into individuals ‘stories and hardships’. A number of RAs experienced a strong desire to know what happened to people after their discharge from ED. For example, one RA described reading about an overdose case but not knowing whether the individual had survived. It became really important to the researcher to find out. It appears that not knowing the outcome of specific cases was a struggle for some RAs, although at the same time they acknowledged that they could not invest time in finding out because they had thousands of cases to document. Occasionally RAs would inadvertently find answers to their questions whilst going through records. The social context documented for many of the self-harm cases, including the distressing circumstances in people’s lives leading up to a self-harm event, was as emotionally challenging for the RAs as reading the descriptive information about how an individual had self-harmed. For example, one RA reported ‘you’re not looking at happy family holiday snaps … you’re looking at dysfunctional families, its women getting abused and beat up by their partners … the boyfriends finished with them, so they’ve taken an OD, they’ve cut themselves’. Many of the RAs could recall details of specific cases. In most instances this was because they felt some personal connection to the case or the case had some personal relevance for them. For example, an event might have taken place on a date significant such as a birthday or anniversary, or the individual may have been the same age as themselves or a family member, or it may have involved children of a similar age to their own, or occurred in a place familiar to the participant.

Stress associated with the cumulative exposure to the sensitive data over a long time period

All the RAs felt that the impact of the data collection was made worse by the amount of time they had to spend looking at ‘every single possible thing that could go wrong to make you sick or hurt’. There was clearly a cumulative effect of looking at the data daily for months, and many of the RAs appeared to be experiencing some degree of emotional exhaustion, talking about ‘reaching saturation point’ or feeling ‘emotionally drained’. One RA reported that some days ‘you just can’t do anymore … you can’t look … you literally just cannot think about it anymore’. Some of the RAs appeared to have become desensitized to the content of the files and reported they ‘no longer feel shocked’ and there was a ‘sense of normality’. They had read so many difficult cases and circumstances that they no longer felt affected by them.

Frustration with the health system and particular cases

A number of the RAs reported being frustrated with the health system itself, whilst others reported being frustrated by particular cases. Some RAs noted, in retrospect looking at the data on a case as a whole, that there was apparent lack of coordination or communication within services. For example, one RA reported ‘I’ll read these things and say how can they be saying that the patient is not suicidal and there’s no risk or low risk, when two hours just previous … there’s a nursing note that says the patient stated that they still want to die?’ Some RAs recalled specific cases where they felt that gaps in communication between health professions or lack of documentation had affected the quality of care received by an individual and people were ‘just patched up and sent home again’. Many of the RAs became frustrated by repeat DSH presenters because of the frequency with which they represented to emergency departments, the volume of notes and the complex histories in each of the cases, which made their data collection process even harder. These ‘frequent flyers’ became known to RAs as they had ‘overwhelming’ amounts of medical files. A number of the RAs expressed anger towards these cases because of the distress they were causing to friends and family and the on-going cost to the health system. One RA asked, ‘why are [we] wasting money on people who are injuring themselves when there are people down the road who are really ill?’

Heightened sense of the fragility of life and increased awareness of the need for safety

Some of the RAs reported becoming more vigilant and aware of physical safety and the safety of others, including strangers. They had become sensitized to the possibility of harm, both accidental and intentional, occurring around them. As one RA explained, ‘you just see accidents after accidents you know people falling off their roofs people getting hit by cars … you start to cycle home going “I could get hit by a car”’. Some seemed to be more aware of dangerous activities than they were before doing the work. For example, one RA reported, ‘I looked at some kids the other day when I was walking along and they were pretending to hang over a bridge and I thought ‘if only you knew’.

Need to develop coping strategies to manage the work

All the RAs had developed their own coping strategies both during the work day and in their personal and social lives. This was seen as essential owing to the feeling of physical isolation they felt because of the geography of the hospital locations. The RAs talked about the need to ‘know their own limits’ and ‘when to stop and walk away’ from the work. Many felt that the flexibility in their work patterns was a really important aspect of the job as they were able to decide when to leave the data collection process and go home. This was especially important if they were working with particularly difficult cases. Many had devised ways to manage their time, creating reward systems such as coffee breaks, walks, catching up with friends or checking emails, depending on how many cases they had completed. All the RAs emphasized how important it was to ‘cut off their job from the rest of their lives’, although some found this easier than others. As one RA explained, they would have to be ‘inhuman … and [have] no compassion to be able to totally cut off’. For many of the researchers the coping strategies they employed outside of work involved changing habits at home such as limiting or avoiding exposure to violence in the media or the reporting of traumatic events, not reading ‘serious books’ or watching ‘depressing movies’. Some RAs reported that they had stopped reading newsprint and only read the news online so that they could screen and limit their exposure to sensitive content and images. One RA reported, ‘my husband [would] say … did you see this headline today? If it had something to do with domestic violence or murder … I don’t want to listen to it. So he knows not to tell me about it anymore.’ Another strategy was hiding the nature of their work from others in social situations. Most of the RAs did not disclose details of their jobs to people outside of work as they did not want to talk about it. Although general supervision was provided for in the design of the wider study (the RAs had regular contact with each other via email, and face-to-face meetings were held at the supervising institution every 3 months) and this was utilized by the RAs, a number of them reported relying on talking to a partner/spouse if they had one. However, some felt they had had to build up their own tolerance because there were not enough people to talk to and they were not allowed to discuss the data with anyone outside of the trial because of patient confidentiality.

Lessons learned and implications for practice

We described some unanticipated issues arising for non-clinically trained researchers during a community-based suicide intervention trial. Although the RAs did not report any significant emotional issues at the time of interviewing, we learnt that the emotional impact of the research on the RAs and the potential for harm had not been fully considered when planning the study. We believe that the emotional impact was exacerbated by the lack of clinical training and the associated unpreparedness for the level of detail within the sensitive data.

The emotional impact experienced by the RAs was consistent with that previously documented by researchers working with sensitive data such as rape or child abuse (Alexander et al., 1989; Milling Kinard, 1996). Through working with the clinical notes and databases in such a meticulous fashion, the RAs were exposed to the many ways in which people hurt themselves and others and the traumatic personal histories that lead to these events. This appears to have resulted in over-identifying with individual cases, increased sensitivity to violence portrayed in the media, difficulties separating work from personal life, lowered tolerance of those who frequently self-harm, frustration and anger towards health services, and a heightened awareness of the fragility of life and the need for safety. A number of the RAs spontaneously described symptoms which could be consistent with burnout, including emotional exhaustion, experiencing work as a heavy burden, developing a cynical attitude towards cases, indifference towards work, and feelings of low professional efficacy (Pross and Pross, 2006). Although the RAs appeared to have experienced emotional effects that could be suggestive of secondary traumatization or burnout, it remains unknown whether this reached the level of clinical significance as no formal assessment of this was made. What our findings do suggest is that the RAs were unable to employ the types of emotional detachment strategies described in the literature (Sorensen and Iedema, 2009; Tutton et al., 2008) and used by trained clinicians. The lack of formal clinical training or experience working with this type of data meant that the RAs were unprepared for the work and as a result had to establish their own personal coping strategies during the process. These included creating reward systems for themselves, limiting their exposure to other forms of sensitive media, not disclosing their jobs to others, and confiding in a partner (which was not recommended). In comparison it is customary practice for clinicians: to manage caseloads and limit the number of ‘intensely traumatic’ cases they see each week; to have regular access to individual supervision; to access one-to-one or group peer supervision; to receive ongoing professional development, education and training; to ensure they maintain a balance of work, play and rest; and to take regular leave.

Some previous authors have questioned whether non-clinically trained researchers should undertake sensitive research, such as this, given that the emotional costs conferred on them by such research might be difficult to establish (Smith, 1992). Our findings suggest that it is possible to establish the impact on non-clinically trained researchers involved in sensitive research and that research leaders and ethics committees need to ensure that strategies are put in place to minimize the risk of harm. Based on our learning from this trial we can make a number of recommendations.

A number of the researchers reported using informal support of partners or other RAs. Although access to informal networks is valuable, researchers should not rely solely on friends and family to provide support. We recommend that non-clinical researchers are provided regular formal professional supervision arrangements, in the same way that clinicians are, to help deal with the emotional impacts of undertaking sensitive research and to avoid burnout. Participation in clinical supervision is now a requirement in the code of ethics of many professional associations in New Zealand. In addition we recommend that researchers required to work alone are given opportunities for managed reflection on their work experiences (as a group) to support them. If the number of researchers does not permit this, then working in pairs is an arrangement which has been suggested by other researchers (Alexander et al., 1989; Driscoll et al., 1997).

Given the detailed and sensitive nature of the data collected in such studies, we also recommend training for non-clinical researchers so that they understand what is involved in undertaking this kind of work, the type of information they are likely to encounter, and how this might impact them. Further, to support researchers to maintain a work–life balance we also recommend that study leaders ensure flexible working hours and patterns. Finally, we recommend that ethics committees are made aware of the potential harm to non-clinically trained researchers in undertaking similar research to minimize the risk of harm to the researchers. This is especially pertinent given that non-clinically trained staff are often hired because they are cheaper to employ than clinical staff. Despite being able to do the task, they may be more vulnerable.

Conclusion

Non-clinically trained research staff engaged in research using sensitive and detailed data are at risk of secondary trauma, and need to be provided with support, training and supervision similar to those in clinical roles. The challenges faced by non-clinical researchers working with sensitive data warrant further research to inform ethics committees and research leaders to understand the implications of this type of work and develop strategies for harm minimization.