Hi Timothy, thanks for your thoughts on this, it certainly raises an interesting point and one which was carefully discussed in the preparation of our manuscript. The point was also raised and carefully considered during the peer review process which included review from a well respected and prominent Professor of Behavioural Science. I think on this occasion our definition has perhaps been misunderstood out of the context for which it was intended. I agree that classification of a behavioural intervention as cognitive based or otherwise is far from straightforward and also lacks clear definitions. In this study we simply aimed to differentiate between interventions which included a component to alter a patient's thought processes, motivations and feelings (which we defined as cognitive-based) and those which followed a more traditional (for the field) approach based on provision of education or behaviour modification.

I hope this helps to clarfiy, I wonder if you perhaps have any suggestions for a different terminology that could have been used? Your feedback is greatly welcomed.

Claire

Thanks for taking the time to respond. In some ways it does clarify things. It does however highlight the issue of being clear about what how our interventions are designed to work. But collapsing interventions I don't think helps. There are clear distinctions between trying to influence someones thought content, thought process, their emotions, etc.... I not sure how it helps move our science forward to collapse them into one rubric. It would possibly have been better described as psychological interventions. Rather this paper seems to point everyone to a CBT model of dealing with things, I am not sure you are intending to put that message out.

Hi Timothy,

Thanks again for your thoughts on this, I really value your feedback and the opportunity the have my work critically evaluated, it's certainly prompted some reflections and I believe this is an important part of research. I both recognise and agree with your comments on the distinctions between the differing psychological processes that interventions target, ultimately, we do need to break it down and figure out which approaches work when and for which patients. However, I still feel a meta-analysis was an intuitive starting point, for the specific application (medication adherence) in question here. Prior to this work there was no consensus as to whether a 'psychological' approach was effective in improving medication adherence, just a handful of studies that indicated it could be. My hope is that in synthesising this research and looking at the collective overview, medication adherence researchers can be more confident in pursing 'psychological' techniques and getting down to the specifics of which techniques work best.

Claire

Did you attempt any meta-regression to tease out any variability within studies by technique, or was there insufficient variability in the study interventions to make that viable. I get your reasoning, and I think the study is important, it is the use of CBT as the overarching term I struggle with. We are constantly bombarded with the message that CBT is the only evidenced based intervention. So superficial readers will see support for this again in this BMJ article. So I would be interested in why you used the terminology, rather than just psychological interventions.

Hi,

I also read this meta-analysis and I had the same questions. However, rather than cognitive-behavioural therapy, I think the authors were more likely to treat about behaviour-change techniques such it is described by Mitchie et al. (2011 ; 2012) in their taxonomy. For example, in page 4 (and also in the table 1), the authors mentioned several elements from this taxonomy. This interpretation helps me in the understanding of the article.

Thanks again for your responses, it's great to know my work is being read, even if it is evoking a few debates, actually saying that, I welcome the debates too.

I think the fundamental point here is that at no point in the manuscript have we actually used the term cognitive behavioural therapy or the abbreviation CBT, we were very careful not to do so as I'm fully aware that this is a very specific intervention. We chose the terminology of cognitive based behaviour change techniques to reflect that we were interested in those pertaining to altering patients cognitions; their thought process, attitudes and beliefs about their medication. Ahmed is quite right, it is closely linked with Abraham and Michie's BCT taxonomy, we're very familiar with this work. However, the taxonomy rightly includes behaviour change techniques of a purely behavioural (e.g. demonstrating the behaviour, providing aids to make it easier) and educational (explaining link between behaviour and health) nature. These were the very techniques we wanted as a comparator as these techniques are the mainstay of current medication adherence interventions and are particularity ineffective for intentionally non-adherent patients, those who willing chose not to take their medication as prescribed due to their thoughts, values, motivation and so forth. Based on a dictionary definition, the word cognitive appeared an intuitive summary to encompass these aspects but on reflection, I can see that our term cognitive-based behaviour change techniques could be confused with cognitive behavioural therapy, even though they are very different. We actually considered using the term psychological based techniques and in hindsight (which truly is a wonderful thing!) perhaps this would have been less of a challenge for our readers. Our reluctance in doing so was the concern that by using the term psychological we would a) infer that the patients had some form of psychological problem and b) these were techniques to be utilised by psychologist and this couldn't be further from the message we wanted to portray.

In response to Timothy's question about teasing out the variability in techniques used, yes and no is the honest answer. We were able to compare interventions using motivational interviewing (MI) with those that did not which showed no real difference but the only other defined behaviour change technique was Implementation Intention Interventions (III) and there were just three studies using this, so not enough data to do much with. All the other interventions were non-specific in so much as they described an intervention to alter patients thoughts processes, perceptions and motivation to adhere but did not use a specific BCT like MI or III. In all honesty Timothy, we were plagued by a paucity of information in so many of the studies as reporting of intervention details was often poor. Our take home message therefore has to be that it looks like these psychological or cognitive type interventions have promise and a re likely worthy of further investigation to establish which techniques are effective and when they are effective