The table is adapted from the PACE participant data analysed according to the Primary Outcome Measure for ‘Overall Improvers’. It has taken over 5 years for QMUL to publish these figures. [i]
Many more participants in all groups reached the threshold of ‘Improved’ for Physical Functioning (PF), than reached the threshold of ‘Improved’ for Fatigue. E.g., at 52 weeks in the CBT group 79 met the PF criteria, nearly double the number (42) that reached the Fatigue criteria. This is a problem.
‘Improved’ on the Physical Function scale needed either a PF score of 75, or an improvement of 50% over the individual participant’s baseline (pre-treatment) score.
For many participants meeting the PF threshold for ‘improved’ required a score of around 60. This is because the average baseline score of all participants was 38ish. A baseline score of 40, plus 50% improvement in that score would equal 60 and be counted as ‘improved’ on that scale, even though with a score of 60 they would clearly remain severely restricted in physical ability.
The fixed threshold for ‘improved’ was 75. As the researchers changed the PF eligibility for inclusion from 60 to 65 – a mere 10 points could make the least disabled participants ‘improved’. Those participants who just about qualified for inclusion, with a PF score of 65, started off over 50% BETTER than the average for all participants.
Whilst a 50% improvement might be of interest, it must be distinguished from the 75 score which might only be 10 points or 13% improvement for the least ill participants. A criteria which encompasses a range of 13% to 50% improvement is useless at best and highly misleading at worst. The 50% over baseline improvement is only interesting and useful as a separate set of data correlated to individual baseline scores.
These points reasonably explain why 3 to 4 times more participants reached PF ‘Improved’ criteria than met the Fatigue scale threshold. The PF criteria must be dismissed as excessively inclusive and not a measure of anything meaningful.
It is concerning that a substantial proportion that reached Fatigue criteria for ‘Improved’ did not also reach the very modest improvement to meet the PF criteria. After all, the researchers claim that the trial was for patients ‘whose primary complaint was fatigue’.
The chart below shows 2 theoretical but perfectly feasible participants. Participant 1 had a PF score of 30 before treatment and at 52 weeks had a score of 45, which means they remained severely physically restricted, but still counted as ‘Improved’. Participant 2 started with a score of 65 and got just 10 points better to qualify as ‘Improved’. Combining these two in a single outcome measure is highly misleading.
Improved on the PF criteria might be interesting and potentially useful, but only when the criteria are separated and correlated independently with Fatigue and individual baseline scores.
The evidence suggests that QMUL are withholding the data at their disposal because they do not want independent analysts to extract practical and relevant information. The PF criteria they belatedly released confounds proper analysis. Perhaps one should be grateful that it is possible to at least start the debate which should have begun 5 years ago. I am not. The release of dribs and drabs of data will never resolve the serious questions about the conduct of the PACE Trial, nor provide participants, patients and professionals with the information they need in order to make rational choices.