It was Valentine’s day recently but instead of emails from secret admirers declaring their devotion I received some less than positive feedback on a paper that I had submitted; also known as a revise and resubmit. While a revise and resubmit decision always stings at first, there was a little more that bothered me about this one. The review that I received, misinterpreting the manuscript as a trial rather than a feasibility study, suggested that there is still some misunderstanding about the nature of feasibility studies that needs to be addressed.
What are feasibility studies?
Feasibility studies are defined in the NIHR guidance, as ‘pieces of research done before a main study in order to answer the question “Can this study be done?”’. At the Complex Interventions conference in Exeter last year I was lucky enough to hear Thabane talking about pre-trial studies and he said they reminded him of the quote: “Never test the river with both feet”. Feasibility studies are the toe in the water that allow us to test out an idea without committing to a two-footed RCT.
What do feasibility studies do?
The NIHR guidance suggests that feasibility studies can be used to estimate some key parameters for the design of the definitive study, including determining the number of eligible participants and establishing a willingness of participants to be randomised – however feasibility studies for randomised controlled trials do not necessarily have to include randomisation.
What do feasibility studies not do?
Feasibility studies for a randomised controlled trial do not evaluate efficacy or effectiveness; that is the aim of the subsequent definitive study. A feasibility study does therefore not have a primary outcome measure in relation to efficacy or effectiveness, but rather a range of outcome assessments which are explored to inform a future trial. Feasibility study therefore report descriptive results on feasibility issues. Since there is no formal hypothesis tested a power calculation is not carried out. However a justification of the sample size is required to ensure a methodologically sound study. This can be informed by papers such as that by Lancaster et al 2004 which suggested at least 30 participants to estimate a single parameter for a future trial (and that is 30 participants per arm at the end of the study) and the more recent work of Teare et al 2014 suggesting at least 35 per arm at the end of the study if estimating the standard deviation for a continuous outcome for a definitive study – although both of these refer specifically to pilot studies rather than feasibility studies per se. Further these papers refer only to the sample sizes required to estimate a single parameter and do not indicate the number of participants that would be required to assess other feasibility and acceptability objectives.
Applying for funding for a feasibility study
There are several NIHR streams which could fund feasibility studies and those considering applying should talk to their local Research Design Service. However, it might be worth considering the NIHR RfPB stream which has awards of up to £250,000 for feasibility studies. The RfPB scheme has recently undergone some changes and there are a number of RfPB events coming soon across England you can find out more from the local Research Design Service websites. The North East event for example is on 2 March 2016.
How should you report a feasibility study?
Extension to the CONSORT guidelines are being developed but a recent paper by Thabane and colleagues has some good suggestions for pilot studies which can be adapted for feasibility studies. Finally, I should say that you should make every attempt to publish your feasibility studies and hopefully you will not meet with any reviewers with misunderstandings about feasibility studies. The Pilot and feasibility studies journal was specifically set up to publish these studies and have reviewers familiar with the nature of the study design.