Getting started on a research project
In the second of a two-part series on research in general practice, Dr Stefan Cembrowicz gives advice on how to get started on your qualitative project
Having identified your topic of interest, make sure you are aware of other work in your field. Avoid reinventing the wheel.
Start with a good literature search. Your local postgraduate centre librarian may help you turn up references that you wouldn't be able to find singlehanded. (Some qualitative purists suggest you shouldn't look at the literature until you have actually done a pilot, so that your ideas don't simply follow in others' footsteps.) Your own password will make searching scientific publications on the national electronic library for health network (NeLH) easy, using a selection of relevant keywords, though I am surprised how many journals are accessible simply via Google. Once you have put together your outline plan, stop and take advice from experienced researchers, such as those from your local research and development department. Will you need the expertise of an epidemiologist or other specialists? Be realistic about how much time is needed for obtaining and writing up your results. Will you be able to do this alone, or should you look for one or more co-workers – perhaps from non-medical fields – to be sure you get the project finished? Qualitative research is now published in mainstream journals but some editors will request independent double-coding of qualitative interviews to support objectivity, before considering publication. Where will the time come from for this project? Will you need grant funds for your time and other expenses?
Consider ethical approval
Ethics committee approval may seem a chore, but is required for any research work involving patients or even NHS staff. You will need to cross this hurdle to attract grants or to publish in peer-reviewed journals. Simple audits and performance reviews do not need ethical approval, but if in doubt and you feel you may have crossed the line between internal audit and research, look at your local ethics committee website or take advice from your local committee or the Central Office for Research Ethics Committees (COREC)1. Allow as long as three to six months for approval; the committees do have considerable research experience and may suggest helpful alterations to your project. Informed consent, anonymity and confidentiality are always key ethical issues to be considered.
Ethics committees may, however, be more familiar with quantitative studies and may expect to see your questionnaires – not realising that the content of these is less relevant than how you later process the text of the answers you receive.
The interview stage
At your first interview, draw up a list of open questions to start the discussion. Decide whether to record digitally or on tape. (As you will need to keep your recordings and mini-tapes cost £3 each, the former is a cheaper form of storage. Voice recognition programs are not infallible – you would still need to correct your transcriptions from the original recording.)
Will you transcribe the tapes yourself? This is time-consuming – consider using a professional typist, who could charge £10-12 per hour. Expect to need three to four hours' typing time for a one-hour interview. Make sure you and your subject have protected time and that informed consent has been obtained – you will need to ensure confidentiality and anonymity.A pilot is always sensible – look back and reflect on your data after several initial interviews. Start to code your texts – draw out themes, look for general similarities. Develop your questions to illuminate your evolving themes. As you begin the cycle of data collection and data analysis and interviews progress, reflexivity gives you new insights into your interviews and enables you to explore issues arising based on information from earlier interviews. Be sure to line-number your transcriptions before you start coding so that a third party can verify your sources of quotes. Make a grid to compare results. Groups of similar codes can then be grouped into broader abstract categories, and a core category – the heart of the study – can then be identified. At this stage your knowledge of the literature and your own experience-based insights are used for comparison of the data with existing and new theory, to explain your findings.Commercial computer programs such as NUD*ist2 and Atlas.ti3 can be used to help sort your text into codes, then process and group the codes into categories. However, Microsoft Word can simply be used to cut and paste text in the same way once you start to identify themes to code.The exception proves the rule. Exceptions and negatives (observations that contradict your current hypothesis) are important as they oblige you to form new theories. It is suggested that data gathering should continue until no further exceptions appear to contradict your theory – at which stage your final hypothesis should be able to explain all the variations found in your data.
When writing up a qualitative project, first choose your journal and make yourself familiar with its instructions to authors. Some journals will require double-coding of your data by another researcher before they will accept your qualitative paper. The context of your project can influence your data, so explaining your methodology is important.
Be sure to make explicit the context you worked in and the methods you used. Describe how you recruited your subjects, to show that your findings can be generalised. Describe how your research was explained to the subjects. Discuss how you sought out and dealt with contradictory or negative findings. Describe how you analysed your data and how you drew your themes and concepts from the data.Present enough of your original evidence (from numbered transcripts) to support and illuminate your deductions. Qualitative research has a different feel from the quantitative work that scientific medical publications have mostly published in the past. But it inherently echoes many features of the culture of general practice – listening to many voices, addressing care of individuals rather than cure of diseases, considering patients' narratives rather than the health of populations, and using intuition to develop theory.Stefan Cembrowicz is a GP trainer in Bristol