General practitioners can play a vital role in the detection and resolution of domestic violence ("intimate partner abuse") because their patients are likely to include twice as many partner abused women than in the general community. They are also in a key position to detect and screen for partner abuse in their patients who may present to them with related symptoms such as depression and anxiety. General practitioners have access to the whole family, including any children, who will also be affected by partner abuse in the home. Clearly, intimate partner abuse is a serious family health issue, but its management may be complicated. Any guidelines aimed at helping general practitioners to deal with the problem should take into account the issues and dilemmas that doctors are likely to encounter with advice on how best to deal with them. However, studies of general practitioners' experiences of attempts to intervene in situations of intimate partner abuse are rare.

The aim of the study was to make an in-depth exploration of how general practitioners respond when a woman presents whom they suspect is being abused by her partner. For example, what interventions do they attempt; how do they react to the patient, her partner, and their children; and how does the problem affect the doctors themselves? From this evidence the researchers hoped to be able to make useful recommendations about how to improve the education and training of general practitioners and to develop strategies for dealing with the problem in the best way, from both the family and the doctor's perspective.

The researchers used a triangulated qualitative case study of two general practitioner training projects about the management of domestic violence plus nested case studies of doctors and their patients. A case study is an in-depth intensive study of a small number of cases. These do not necessarily need to be people; they can be anything the researcher is interested in.

In this study, the cases were the two training projects plus the doctors undertaking the training and their patients who were the nested cases. Triangulation is often used in case studies and normally involves using several different research methods to study the same problem, usually to give many perspectives or to cross validate the findings from the different research methods. In this study, the researchers used triangulation to compare the doctors' strategies for dealing with intimate partner abuse with current recommendations in the medical literature.

The data collected in the study were qualitative which means they are non-numerical. Qualitative research can take many forms, but most commonly, as in this study, researchers collect verbal data in interviews. The researchers used a semi-structured interview which means that they had a set of research questions which they hoped to answer through a roughly fixed set of interview topics, but would have allowed the interviewer a fair degree of flexibility in the way the interview was done. This kind of flexibility is important because qualitative interview data are often used, and are ideally suited to, explorative research where little is currently known about the research problem.

When there is a degree of flexibility about what the person being interviewed can talk about, the interview data can be generative-used to generate new ideas and theories about the research problem. On the other hand, if the interview schedule was completely fixed, or structured, then the research might fail to reveal important ideas and information not covered by the interview questions. In this study the issues that the researchers specifically wanted to explore in their interviews included how general practitioners identify domestic violence and what strategies they use to manage it, as well as how useful they felt their training had been in helping them to deal with the problem.

Of course, the downside of case study and qualitative research is that it usually involves small samples of respondents which have probably not have been selected on the strict statistical grounds usually used in quantitative studies that allow the researchers to generalise their findings to a certain population, or group of people. In this study a purposive sample of doctors was used which basically means that the doctors were selected because they were undertaking training in the area that the researchers were interested in studying.

Consequently, the researchers cannot justifiably claim that the responses of the doctors in their study are representative of all doctors in Australia. But what qualitative research gains is a richness and breadth of information which hopefully leads to theoretical generalisation, permitting the researchers to draw up preliminary theories or practical solutions to the problem which can then later be tested using quantitative research methods.

Although qualitative data may lack some of the scientific virtues of numerical data, they still requires some careful analysis to be placed firmly within the realm of scientific inquiry. There are a variety of ways of analysing verbal data, each associated with a different school of scientific philosophy and also depending on what kind of information you want to take from the data.

In this study the researchers used a grounded theory approach which involves an iterative or repeated process of intensively studying the interview data to find themes, categories, patterns, and ideas about theories, while constantly comparing these back to the original interview data and using illustrative and supporting verbatim accounts.

As with statistical analyses, computer software can be used to help with some elements of this kind of analysis. But ultimately all decisions about what themes and patterns are present within the data must be made by the researcher, based on an intimate knowledge and feel for the data.

The kind of approach that the researchers used is particularly useful when you want to study interactions between people by looking at the meanings that those individuals attach to their interaction and the situations in which it occurs. This idea is important because it serves as the basis of one important school of thought within social science (phenomenology), which states that the way individuals make sense of a situation determines the way that they behave and so the outcome of that situation. In other words, if you want to start to make sense of what people do in a certain situation, you need to first understand how they think about that situation, and one of the best ways to do that is to ask them about it in an in-depth interview.

This way of thinking is particularly relevant for this study because you can see that the way that a general practitioner might respond to a patient subjected to intimate partner abuse could have an impact on the outcome of the situation for both parties. In fact, if you read the results section of the paper you will find a great deal of examples of how the doctors' thoughts and judgements influenced the actions they took and how these actions ultimately affected both the doctor and the patient, in some cases with unfortunate consequences.

So verbal data are useful for understanding how your respondent thinks and feels about a situation, but the key word really is understand. Researchers in any field often make the mistake of equating what interview respondents say with what actually happens. For example, one doctor who posted a rapid response to this paper claimed that because the emergency physicians in her own study report that they see a lot of patients affected by domestic violence that their awareness of the problem must therefore be high.¹ But until we know exactly how many partner abused patients the doctors are detecting out of all those who come to see them, we cannot say for sure that their judgment is necessarily accurate, only that the doctors themselves believe it to be so.

As with any good qualitative research, this study has produced a large volume of detailed information, some of it surprising, perhaps even shocking, but always interesting. The interview data have generated several themes indicated under the subheadings in the results section of the paper, some of which throw light on previously unexplored issues such as how general practitioners manage children of partner abused mothers.

Glossary

Case-study: An in-depth study of a small number of ‘cases’, which can be anything from individuals and groups of people to events and behaviours. Usually involve small sample sizes and qualitative research methods.

Qualitative research: Research which involves the collection and analysis of non-numerical data. Usually this means verbal data gathered in interviews but also includes many other types of research such as analysis of text, diagrams, and pictures or behavioural observation and mapping. Particularly useful for initial or explorative research on problems where little is currently known.

Triangulation: Comparing two or more different types of data about the same research problem to give a range of perspectives, or to cross validate findings from different research methods.

Semi-structured interview: An interview with a roughly fixed set of interview topics but which allows flexibility in the way the interview is conducted and the exact phrasing of questions. Useful for when the researcher has a specific set of research questions but also hopes to generate new ideas from the interview data.

Structured interview: An interview with fixed questions in a set format. May also involve fixed response categories from which respondents must choose their answers.

Purposive sample: When researchers choose participants for their research based on their suitability for the study, rather than random selection.

Statistical generalisation: When researchers make statements about a ‘population’ of people in the real world based on findings from research involving a much smaller ‘sample’ of people taken from the population. Requires adherence to strict statistical criteria covering sampling and data collection and analysis.

Theoretical generalisation: When the data collected about a subject are rich enough in depth and breadth for the researcher to be able to propose a preliminary theory about the subject.

Phenomenology: A scientific philosophy within the social sciences which states that reality is only knowable and constructed through subjective experience and the meanings that individuals attach to it. This is opposed to ‘positivism’ which states that events can be objectively observed and are governed by laws which can be discovered through scientific inquiry.

Grounded Theory: A means of analysing verbal data from a phenomenological perspective which results in a set of themes, categories, patterns and preliminary theories drawn by the researcher from the interview data. Involves iterative, or repeated, comparison of verbatim interview data with the researcher’s findings.

From the discussion section of the paper we can see that the research has shown some problems with the way that general practitioners manage and are trained to deal with domestic violence which sometimes has important consequences for all parties. From these findings the authors have been able to put together a series of recommendations about policy and training of general practitioners that will hopefully lead to better management and a better resolution of the problem for patients subjected to partner abuse and the doctors who must negotiate this difficult problem.

1. http://bmj.bmjjournals.com/cgi/eletters/328/7440/618 (accessed 16 Apr 2004).

1. Tipper EK, Boyle AA. Domestic violence: the difference between emergency physicians and general practitioners [rapid response to Taft A et al. General practitioner management of intimate partner abuse and the whole family: qualitative study]. BMJ 2004.
   http://bmj.bmjjournals.com/cgi/eletters/328/7440/618 (accessed 16 Apr 2004).