Concerned Citizen wrote:
Part 1 of 2-
PACE was not good news for the majority of CFS patients who don't benefit much from these therapies but will be now be goaded into it because of a flawed trial. Results were relatively unimpressive and don't support the hype we saw in the news media following the Lancet publication. Reactions from the patient community only seem extraordinary to those who are unaware of the struggle we have faced due to psychosocial interpretations of this condition. This is not about automatically rejecting psychological factors or behavioural interventions because we simply don't like them, as misguided stereotypes of CFS patients suggest, but rather it is criticism of the exaggeration, oversimplification, and overgeneralisation we have endured from the principle authors of the PACE trial and colleagues throughout their careers.
Cognitive behaviour therapy (CBT) is based on the hypothesis that abnormal illness beliefs are the primary factor in perpetuation of CFS symptoms, graded exercise therapy (GET) is based on the hypothesis that deconditioning is the primary factor. These hypotheses have not been confirmed in theory or practice. Cochrane systematic reviews into for CBT (Price et al 2007) and GET (Edmonds et al 2004/8) suggest there is a small temporary effect on fatigue for a minority of ambulatory CFS patients but all other measures are in doubt, not to mention severely affected patients remain excluded from this research.
Multiple patient surveys have consistently shown adverse effects due to CBT and GET, including people who underwent professional rehabilitation. The authors dispute this and took great care to avoid symptom exacerbations in the trial. The handling of adverse effects is open to interpretation, and although the patients were carefully guided to gradually increase activity levels, we are not told what proportion of patients actually managed to increase their activity due to therapy or by how much.
It is ironic that the authors view the patients' irrational beliefs and unreliable perceptions as central to CFS but deem them dependable enough for subjective questionnaires. In comments on their published protocol (White et al 2007), the authors resisted the use of objective tests to confirm the alleged improvements, actigraphy in particular with the excuse it would have been too much of a burden to patients, despite using it at the beginning of the trial. Their colleagues in Europe (Wiborg et al 2010) in a meta-analysis have confirmed with the use of actigraphy that CFS patients are not increasing their activity with CBT/GET, which wouldn't have looked too good in a study that cost 5 million pounds and expected to deliver. PACE did use the 6 minute walking test, GET had a non-significant improvement over the other groups (under 0.5SD), but all groups were still on average well under 400m, much lower than an average healthy score of 650m and comparable to patients with serious medical diseases.
Part 2 of 2-
PACE used the Oxford criteria (notoriously heterogeneous with higher rates of psychiatric comorbidity) and had a strict process for excluding possible medical conditions, yet a massive 80% of candidates from fatigue clinics in the UK were excluded from the trial. For sub-group analyses, PACE also used modified and unrepresentative versions of the London ME criteria and the CDC criteria. Much of the patient community endorse the Canadian criteria (Carruthers et al 2003, updated by Jason et al 2009) which White et al have avoided, many patients who fit this criteria probably would have been excluded from the trial.
PACE compared CBT and GET to a modified form of pacing which does not represent what the CFS community knows as pacing. We do not avoid all exacerbations or stick to a 70% of perceived energy rule. Can the authors be trusted to fairly test a rival therapy?
In the original protocol, PACE counted 75/100 points on the physical subscale of the SF-36 healthy survey towards a positive outcome, with 85+ counting towards recovery. These values were derived from calculating the mean (average) score minus 1 standard deviation from a working age population, a questionable definition when even the majority of a general population are scoring at or close to the upper limit ie 95-100/100 points (Bowling et al 1999 cited by PACE for normative data). Jenkinson et al 1993, also cited in the original protocol, reveal a mean(SD) score for a *healthy* working age population of 92.5(13.4), or a lower threshold of about 80/100 points according to the PACE definition.
Yet in the Lancet paper, PACE dropped that and now claim 60 points is the lower threshold for normal (as one of two criterion, the other being based on fatigue). We are not given a reason for this goalpost shift. Calculations reveal the authors did not use normative data from a working age population as claimed, but instead a general population which included the elderly and diseased (84-24=60). To some extent, similar problems also exist for their measurement of fatigue using the Chalder Fatigue Questionnaire. This invalidates the figures given for participants returning to normal and a correction should be published. A related mistake was repeated in the editorial which claimed PACE used a strict definition of recovery (which didn't exist in the final paper). Oddly, it was possible for a patient to classified as normal despite being impaired enough to qualify for trial entry. Please ask Richard Horton (the Editor in Chief) to explain how these blunders managed to pass the prestigious peer-review process of the Lancet?
There are other identified problems in this trial which we as a community of patients and carers have had to discover on our own because apparently the media and wider scientific community are dropping the ball. We are sick and tired of being subjected to spin, peoples' lives and health are at stake here while other people pontificate.
