The historic lack of diversity and representativeness among public contributors in research is well recognised. When study design is informed predominantly by people from populations already well represented in research, barriers experienced by other groups may remain unidentified and unaddressed.
An intersectional perspective can be valuable at this stage. Barriers to participation in research may arise through the interactions between ethnicity, age, socioeconomic circumstances, disability, geography and other factors. These issues are better considered when research is being designed than trying to do so through additional subgroup analyses after data has been collected.
Report and share but don’t necessarily analyse
There is a strong case for study populations to be reported in greater detail, regardless of whether the characteristics described are subsequently analysed, to facilitate the wider and better reuse of research data.
Consistent reporting of certain characteristics allows better metadata to be created. When it’s clear who has been included in a dataset, its relevance to other research questions can be assessed more readily and appropriate datasets can be identified for reuse. With suitable consent and governance in place, data can then be responsibly shared and, where appropriate, combined.
Populations that were too small to be reliably analysed within one study need not remain scientifically invisible. The retaining and sharing of sufficiently detailed data can facilitate individual participant level meta-analysis and other pooled approaches to allow for important research questions to be examined with greater statistical power, and at a scale at which robust answers can be obtained.
Inclusive research has rightly been prioritised, but scientific rigour should not be compromised. Sustained and informed discussions among all members of the research community about diversity, representativeness and intersectionality, the distinction between reporting and analysis, and the importance of the role of data reuse, are essential. Its only by doing this that we’ll ensure increasingly diverse datasets produce reliable and meaningful evidence, and that public investments in research are maximised.
Source link
