The PACE trial in chronic fatigue syndrome.

In their randomised trial of treatments for patients with chonic fatigue syndrome, Peter White and colleagues (March 5, p 823) defi ne a clinically useful diff erence between the means of the primary outcomes as “0·5 of the SD of these measures at baseline, equating to 2 points for Chalder fatigue questionnaire and 8 points for short form-36”. They cite achieving a mean clinically useful difference in the graded exercise therapy or cognitive behaviour therapy groups, compared with specialist medical care alone, as evidence that these inter ventions are “moderately eff ective treatments”. The source for this defi nition of clinically useful diff erence states that such a method has a “fundamental limitation”: “estimates of variability will diff er from study to study...if one chooses the between-patient standard deviation, one has to confront its dependence on the heterogeneity of the population under study”. In White and colleagues’ study, we do not have heterogeneous samples on the Chalder fatigue questionnaire and short-form 36 physical function subscale, since both are used as entry criteria. Patients had to have scores of 65 or less on short-form 36 to be eligible for the study. However, most, in practice, would probably need to have scores of 30 or more to be able to participate in this clinic-based study. Indeed, only four of 43 participants in a previous trial of graded exercise therapy scored less than 30. Guyatt and colleagues suggest that “an alternative is to choose the standard deviation for a sample of the general population”, which White and colleagues have given as 24. An SD of 24 gives a clinically useful diff erence of 12; both graded exercise therapy and cognitive behaviour therapy fail to reach this threshold. Whether they “moderately improve outcomes”, as claimed, is therefore questionable. I am chair of a myalgic encephalomyelitis support and advice group—an unpaid voluntary position.