One fifth of HSR studies funded by the Department of Health in England between 1994 and 2004 were mixed methods studies, with some evidence that this had increased over time. The frequency of use may have been different if fellowships and programmes of research had been included in the study. Evidence elsewhere around the frequency of use of mixed methods research is based on publications rather than funded studies and thus there is no evidence to support or refute this finding. Other research indicates both lower and higher use of a mixed methods approach in the health field. For example, only 1% (37/3830) of papers in clinical journals were found to use both qualitative and quantitative methods, although the focus of the study had been to identify qualitative research [35
]; and 10 of 26 (38%) evaluations located in health journals in 1995 used a mixed methods approach [32
]. In other research fields, it has been cited that 13% (145/1156) of articles in education journals used a mixed methods approach [5
], and 8% (14/170) of primary research studies on long term conditions were mixed methods research [36
]. Despite the lack of supporting evidence, it appears that mixed methods studies are common enough in HSR to be identified as an important methodological approach warranting further consideration. Thus established and emerging literature on mixed methods research is highly relevant to the HSR community.
Researchers justified the use of a mixed methods approach on pragmatic rather than ideological grounds – they worked in an applied field studying complex issues in complex environments. Qualitative research helped them to address and understand these complexities, and bring in the voices of users and providers of services to help them to do so. A pragmatic justification for using mixed methods research has been found recently among social researchers in the UK [8
]. Thus this is not unique to the HSR community. Indeed although much of the established and emerging literature on combining qualitative and quantitative methods discusses the philosophical challenges of taking this approach, this literature also addresses a pragmatic stance [10
], making it highly relevant to HSR. However, a potential downside of the driver of mixed methods research being the practical need to use a range of methods is that researchers may not view their research in the context of the body of learning about mixed methods research, thus limiting the knowledge they yield from this approach [27
The main justification for using a mixed methods approach in HSR was comprehensiveness rather than the range of justifications discussed in the literature. In particular researchers did not use a justification of increased validity when different methods with different strengths offer convergence of findings. This lack of use of classic triangulation, or confirmation, in HSR is welcome because concerns have been expressed about the difficulties of using methods with a purpose of confirmation [13
], particularly as an indicator of validity [37
]. Having said this, confirmation is a key focus within much of the mixed methods literature and the HSR community will need to be aware of this when reading the literature about mixed methods research.
Researchers in HSR discussed the intrinsic value of mixed methods research but also discussed its use for strategic purpose. The main strategic use was to gain funding and researchers expressed concern about this because it might result in poor quality components of studies. Again, this was not confined to the HSR community because social researchers have also stressed the centrality of the research question in determining methods and expressed concerns that mixed methods research has become a fad because of funding bodies' desire for this approach [8
Researchers in HSR drew on a range of both quantitative and qualitative methods but tended to make a lot of use of surveys and interviews. This frequency of use of interview studies in mixed methods research is very similar to social research where 71% of mixed methods articles were based on interview studies compared with 67% here [5
]. Of course the questions posed in HSR may be best addressed by individual interviews, but this may also indicate a neglect of potentially useful qualitative methods such as observation and documentary analysis. Researchers appeared to be drawing on a wide range of roles for the different methods within studies in the contexts of evaluations, survey and fieldwork, and instrument development. A lack of use of some roles might be due to the fact that they are best suited for use within programmes of research rather than single studies but nonetheless researchers may wish to consider the range of roles of methods detailed here and the relevance of these roles within their future research. A lot of use was made of predominantly quantitative designs which again may be most suitable for HSR or may indicate a lack of variety of designs.
Perhaps the most interesting finding here is that a combination of qualitative research with a randomised controlled trial is a minority of the types of mixed methods studies undertaken in HSR yet this is where health services researchers are making a strong contribution to the literature on mixed methods research. This may not seem surprising given the significance of the randomised controlled trial within HSR. However, it is important that researchers in HSR note the frequency with which they combine methods in the context of non-randomised evaluations, survey and fieldwork studies, and instrument development, and recognise that they can make a contribution to the development of mixed methods research in these contexts too.
This paper is based on an empirical study of mixed methods research in a specific research field. It focuses on why and how this approach is undertaken within HSR. The empirical study covered a number of issues of importance to undertaking mixed methods research in HSR which are not reported here, including paradigm differences between researchers, the effect of team working on research outputs, assessment of the quality of mixed methods studies, and the facilitators and barriers to integration within studies. These will be reported in further papers.
The responses to requests for documentation were representative of the specified population of mixed methods studies. However, there was a lower response to requests for proposals than reports. Assessment of the study documentation did not include double coding to check for inter-rater or intra-rater reliability. The studies here did not include all mixed methods studies in HSR funded in England between 1994 and 2004. HSR was funded by Regional Health Authorities, charities, and research councils over that time period. In addition, programmes, initiatives and fellowships were not included. There is no reason to believe that HSR funded through other sources is different from the HSR funded by the Department of Health, except in terms of the extent to which it was policy related. However, the research cultures of different countries in North America and Europe may be different from England and the findings are likely not to be generalisable outside the UK. Transferability is relevant to the findings from the qualitative component [37
]. The context of the qualitative research has been described to allow readers to consider the transferability of the findings to other settings. The context was researchers in HSR, mainly in England, where quantitative methodology has dominated historically.