Reviewing Qualitative and Quantitative Studies: Mixed-Method Reviews

Some review questions do not fall into the neat categories of being qualitative or quantitative, instead being a mixture of both. Sometimes problem can be overcome by breaking a question down into qualitative and quantitative components which can then be analysed separately. Alternatively, one can undertake a mixed-methods review. There are three broad approaches to this. The first is to convert the qualitative data into quantitative data, or the quantitative data into qualitative data and then combine them directly; the second is to undertake separate analyses on each and then combine the results; and the third is to analyse the two data sets sequentially one after the other. These are known as integrated/data-based convergent designs; segregated/results-based and parallel-results convergent designs; and contingent/sequential designs, respectively. The main guidance for these types of reviews is provided by JBI. Research questions are often seen dichotomously as being either quantitative or qualitative. However, it is sometimes more complex than this, and many questions involve consideration of both types of data. For example, many modern review and guideline methodologies mandate that patient values and preferences are considered alongside the strength of the evidence when making recommendations. Some research questions naturally call for a mixture of qualitative and quantitative data, for example, a proper evaluation of a drug will look not only at its quantitative effect but also at how people feel when they are taking it? In such cases, it would be perfectly proper to undertake two separate reviews, but another option for the reviewer who wanted to undertake an all round review would be to conduct a mixed-studies review encompassing both types of data in one review.