The use of qualitative evidence in clinical care
- Department of Caring and Public Health Sciences, Mälardalen University, Eskilstuna, Sweden
Evidence-based nursing provides nurses with the means to use results from research in a systematic way, informing care that is balanced with and tailored to patients’ specific needs. Nursing research comprises far more qualitative than experimental studies. Findings from qualitative research can provide nurses with information to develop and improve professional care. However, many qualitative studies provide few clues as to the ways in which their findings might be applied to clinical practice. Some authors argue that the responsibility for providing such clues should not rest with researchers: “It is the reader who has to ask, what is there in this study that I can apply to my own situation and what clearly does not apply” (p34, emphasis added).1
Nevertheless, the question of “application” remains. How can the findings of studies based on narratives from a few informants be applied to caring situations with other patients? In other words, can the findings of qualitative studies be regarded as generalisable?
ASSESSING GENERALISABILITY OF QUALITATIVE RESEARCH
There are 2 main approaches to assessing generalisability of qualitative research. The first is to acknowledge that many quantitative techniques fail to adhere to strictly probabilistic sampling (eg, stratified random sampling). Qualitative studies are simply alternative forms of research, albeit forms that deviate from probabilistic assumptions. Nevertheless, researchers must still demonstrate that the findings are in some way representative of the population to which the findings are to be generalised.2 Hammersly3 suggests that researchers can draw on statistics and deliberate diversity in study settings to establish similarities between the study and wider populations.
The second means of making sense of generalisability in qualitative research is to argue that “phenomena are [so] intimately tied to the times and contexts in which they are found” that generalisability, as we normally make sense of it in a clinical trial, is impossible. …