Prof Edwards: an unblinded trial such as PACE simply does not provide useful information

@BMJ: Re: Tackling fears about exercise is important for ME treatment:

I am a retired academic physician and biomedical scientist who has become involved in advising on research directed at ME/CFS because of specific relevant expertise in therapeutic trials. This condition presents particularly severe methodological issues in trials because the desired primary endpoint is subjective and even supportive secondary endpoints are not as objective as one would wish. In this context, an unblinded trial such as PACE is simply uninterpretable. It does not provide useful information on which to base clinical protocols.

There appear to have been a number of additional methodological problems with the trial, but those aside, the basic design does not allow of sufficient scientific rigour. A large number of patients are concerned that the trial should be used as a basis for recommending treatment and I think they are right to be concerned.

As far as I can see all that the PACE trial can tell us about is apparent changes in patients' beliefs about their illness. The trial appears to be based on the supposition that these may be unrealistic beliefs, so it is hard to see why a shift in these beliefs should be considered even relevant unless there is some way of establishing that they become more realistic - which there is not. (Other measures suggest that any change was unwarranted.) We cannot even be sure that these are patients' beliefs since an important part of the 'placebo' phenomenon (that demands blinded controls which are here impossible) is likely to be a response designed to please a therapist, perhaps because of fear of being discharged from care if not perceived as grateful. The fact that perceptions can modify behaviour seems to be acknowledged by the aims and design of the present study.

I do not treat people with this illness and have no other personal interest other than feeling that I may be of use in encouraging useful research. Apart from anything else I am saddened to see poor data of this sort being used in a way that will perpetuate the lack of trust between patients and their carers. The patients are very aware of the weaknesses of the study and I am surprised that those designing the study are not equally aware. Competing interests: No competing interests

18 January 2015 Jonathan CW Edwards Professor of Medicine, Emeritus University College London Gower St, London WC1