SF-36 and Chalder Fatigue scores for healthy age & sex matched healthy controls to CFS patients
There's been some discussion about how appropriate were the 'population norms' used by PACE for determining how many CFS patients were within the normal range. So I was interested to see some SF36/Chalder data from a
2005 study co-authored by Peter White that- quite appropriately - chose to compare CFS patients to age, sex & BMI matched healthy but sedentary controls. Here's what they found:
SF-36 Physical Function score:
CFS patients= 47.5 (27.5-65)
Controls= 95 (90-100)
Chalder Fatigue Scale score
CFS= 11 (9-11)
Controls=0 (0-2) nb Bimodal scoring, 0 in bimodal is
<=11 in Likert scoring
Note that scores are given as
Median (Interquartile range), not means as the distributions are highly skewed so non-paramateric stats are appropriate (median, IQR) rather than parametric ones (mean, SD). The PACE trial inappropriately used parametric stats - not sure how that got past the two MRC biostatisticians on the paper. Here's a comparison between the medians from the 2005 (Healthy, age-, sex- and bmi-matched to CFS patients with a similar profile to those in the PACE Trial) and the PACE means (using populations that made their data look good).
SF-36
PACE reference population: 84
2005 Study matched controls: 95
Chalder (likert scoring)
PACE reference population: 14.2
2005 Study matched controls: <=11
spot the difference...