Our survey shows that the practice of earmarking resources to develop a central system for appropriate methodological support is not widespread in the NHS Trusts in England Although this problem in health care organisations was highlighted more than a decade ago, little seems to have changed since then in England.
If resources are limited, health researchers may have to perform their own analysis, relying only on short courses or a limited number of advisory sessions. Although these arrangements are better than having no methodological support at all, they may not always produce satisfactory results.
Ideally all health care organisations hosting research should develop a central resource for methodological support, for example a Medical Statistics/Biostatistics Unit. This should be formed in collaboration with their academic partners in order to attract high quality methodologists with their own research programmes. This should enable support from experienced methodologists or supervised support from the junior ones. It will also facilitate specialist support if needed. The number of methodologists should be appropriate for the volume of quantitative research carried in the organisation. Unlimited support free of charge should be provided at the initial stages of a study: study design and preparation of grant applications/protocols. A methodologist's involvement in data analysis should be funded by research grants where available or possibly subsidised by the R&D or a similar department, depending on the quality of the research and the research priorities of the organisation. For the data analysis to be done effectively, the methodologist should be an integral part of the research team. Limited free support could be provided for supervising data analysis if the researchers sought advice at the study design stage and the analytical requirement is simple. Ideally the waiting time to see a methodologist should not be too long, possibly not exceeding a couple of weeks. Additionally, the methodologists should contribute to the organisation's research governance activities through training programmes, methodological audits and the development of methodological guidelines. Audits of research protocols, statistical needs etc. could be used as means to target support to researchers. Our own Trust is working towards developing a central resource of methodological support similar to the one described above.
However, having a structured system for methodological support in organisations which host health research may not always be sufficient to ensure high methodological quality of studies. Organisations should also adopt strategies to ensure that the service is being used effectively by their researchers. This could be done by raising awareness among their researchers to recognise the need for support and to monitor whether research projects have appropriately involved methodologists where necessary. For example, the RG framework in England requires that all research in NHS Trusts should have an independent expert review either through the peer review process of grant bodies or an internal review process within the Trusts . These review processes should ensure that a research team has the necessary expertise to conduct their proposed research to a high standard. The Royal Statistical Society recommends that a statistical review should be provided for research projects during the approval stage by the R&D department of each Trust . At the same time the researchers should assess when and to what extent their projects need methodological support and to budget for it appropriately when applying for research grants.
Our survey is focussed on support available from NHS Trust's R&D Departments or via their links with other supporting bodies. This maybe seen as a limitation as some health researchers may work in departments which employ their own methodologists. However, this does not address the problem for those researchers who do not have this facility. Furthermore, if a junior methodologist is working in isolation within these departments the level and quality of support may not be always be adequate .
Reviews of published literature in health journals show that the use of poor methodology in health research is not confined to England only (1–5). In the USA, following calls from the Institute of Medicine of the National Academies of Science for the overhaul of research approval processes, many institutions are implementing systems of research oversight that separate out procedures for scientific review from the ethics review process .