Resolution of deliberate self harm: qualitative interview study

Leanne Tite takes you through a paper about the reasons why people stop self harming and explains the use of a qualitative research method

This month's paper is Sinclair J, Green J. Understanding resolution of deliberate self harm: qualitative interview study of patients' experiences. BMJ 2005;330:1112-5.

Why do the study?

Deliberate self harm is a serious health problem that can present itself as a response to personal difficulty, as well as across more pervasive mental health problems including personality disorders, depression, and psychoses. In some cases, deliberate self harm is a "cry for help," or a means of outwardly expressing internal anguish, but in other cases it can lead to considerable disability, even suicide.

Despite the extent of the problem, services for treating people who self harm vary widely across the United Kingdom. In addition, there is considerable uncertainty surrounding the best way to treat patients who self harm because of evidence of the limited effectiveness of current treatments and limited research on the problem itself. What little research that has been done so far has tended to take place in the time immediately after an episode of self harm, allowing no insight into long term outcomes for people who have in the past and continue to self harm. Evidently, there is a need for greater research on the problem to help inform new ways of managing self harm and to evaluate existing treatment options. The researchers of the paper sought to obtain this insight by asking a series of patients who no longer self harm to give their account of how they had overcome their own self harming behaviour.

What it means to them: when to use qualitative research

The researchers used a qualitative research method to gain insight into the problem of self harm from the perspective of those who have experienced the problem first hand. Qualitative research involves collecting non-numerical data, which can take many forms but most commonly means verbal data collected in interviews. Interview data can provide a large amount of rich and personal information, but the information is always subjective. Firstly, because unlike numerical data it cannot be readily analysed or broken down into objective straightforward results, such as a predefined age band or a test score. But secondly, and perhaps most importantly, qualitative data is subjective by its very nature. Verbal data obtained in an in-depth interview does not try to give facts or an objective account of an event but, instead provides an account of an event from a personal perspective. The aim of this kind of research is to obtain an insight into an individuals' experience, how they interpret that experience, what they have drawn from it, and the meanings that they attach to it. As such, qualitative verbal data is essential when the researcher's aim is to understand a behaviour or situation from the perspective of the person involved.

You can see then that in the current study, a qualitative interview method was essential because here the researchers wished to find out how patients who previously self harmed understood the events and changes in their life that led them to stop self harming, hopefully gaining some useful insight into treatment strategies that might help others overcome their own self harming behaviour.

Who did the researchers interview?

The kind of interview used in this study was a semi-structured interview, which means that a brief and very general interview schedule (set of questions) was used to prompt participants to talk about certain times and experiences in their life, while allowing almost complete flexibility in the kind of information that patients choose to give. This ensures that the interviewer obtains the information they need without overly influencing the participants' responses. This freedom in the course of the interview is important in qualitative research because it allows information to emerge that the researcher might not anticipate, potentially shedding a whole new insight onto a problem.

Notice that when selecting their interview sample, to look at long term outcomes, the researchers deliberately sought patients who had a history of self harm but who had not self harmed for at least two years before the interview. In numerical research, the aim is usually to generalise findings from your sample to a larger population, which means that participants must be randomly selected from that population. In contrast, when participants are purposefully selected from a population because of some characteristic, as in the current study, the generalisability of any findings is limited, because of the bias in the sample. The goal of qualitative research is rarely to generalise findings to a large population, however, not least because sample sizes are usually small, but instead to give breadth and depth of information on a relatively unknown topic. In this case, it is acceptable or even desirable to select participants based on a characteristic of interest.

All participants were interviewed by a psychiatrist, but the researchers deliberately set the interviews in participants' homes. This was to try to reduce the "interviewer effect" - that is, when participants respond to the interviewer in certain ways according to their role as respondent. In this case, the roles of doctor and patient could have easily emerged in the interview situation, perhaps considerably influencing the kind of information that participants choose to give as well as the way they portray it. By using an informal setting, the researchers hoped to reduce the interviewer effect and obtain more valid data.

How was the data analysed?

To make sense of qualitative data, it must be analysed rigorously and methodically just as numerical data is. There are many different ways of analysing interview data depending on the researchers' disciplinary perspectives as well as what they hope to take from the data. In this study, the researchers conceptualised the interview data as narratives or life stories.

Broadly, the narrative approach to research says that people make sense of their lives by attaching meaning to their experiences by constructing stories from them. Researchers can, in turn, then understand the meaning of experience for the individual by listening to the stories or narratives they tell of their lives. To help this approach to analysing their interview data, the researchers drew upon a previous theory about the narratives people use to make sense of illness, devised by sociologist Arthur Frank. Frank argues that people's narratives of illness usually take one of three types - of recovery, chaos, or quest (see below). These theoretical frameworks gave the researchers a "listening device" to draw upon when trying to make sense of the stories provided by the participants in their own interviews.

The researchers used techniques from grounded theory (a method of analysing verbal data) to organise and analyse the vast amount of data generated in each interview. This involves an iterative or repeated process of studying the interview data intensively in search of themes or ideas about recurrent categories and patterns of responses that the data fit into, while constantly referring back to the original interview data to check the accuracy of the emerging themes.

What did the study find?

Using Arthur Frank's theory, the researchers found that 18 of the 20 participants' interview stories fitted into one of two illness narratives across three different themes. In the first group, self harming was seen as a response to a loss of control within the home during a chaotic adolescence. For these individuals, admission to hospitalwas a traumatic experience and only worsened their sense of lost control. Here, the key to resolving self harm was the gaining of autonomy later in adulthood. The researchers interpret this as Frank's quest narrative, which emphasises a journey into becoming someone new after recovery from illness.

JOE PARTRIDGE/REX

In the second group, self harming was seen as triggered by participants' use of alcohol as a means of escaping their problems. For this group, the restitution narrative describes a sense of recovery, in which respondents saw giving up drinking as a way of regaining their lives and their pride. In the final group, the restitution narrative again describes how participants saw their self harming as a cry for help that led to treatment of a more pervasive problem (depression), again emphasising a sense of recovery in participants' stories.

You can see that by conceptualising the different accounts of self harm using different narratives, the kind of treatment that might be helpful changes, depending on the context of the self harming behaviour. The researchers were able to conclude that for the first group, hospital admission might actually worsen feelings of lack of control, but for people in which alcohol and depression are a triggering factor, health services are essential in dealing with the problem by treating first the underlying cause.

Was it a good study?

By analysing narratives, the researchers were able to draw conclusions about how patients who previously self harmed accounted for the resolution of the problem, allowing tentative suggestions about treatment options for people who self harm. However, this research gives just one side of the story and begs the question of what happens to other patients who overcome the same issues but continue to self harm? An absence of comparison does not, in itself, invalidate this research by any means but emphasises that, like all research, it is limited in the conclusions it can draw and to whom those conclusions apply. On the other hand, what this study gives is a perspective from the patient. Analysing figures about self harm rates after certain types of treatment is one way of evaluating treatment options, but really under-standing why patients self harm can give a much needed humanistic slant on a sensitive and highly personal problem.