There are 11 questions in the Chalder Fatigue Scale, as used in the PACE trial:-

  • twotalkingDo you have problems with tiredness?
  • Do you need to rest more?
  • Do you feel sleepy or drowsy?
  • Do you have problems starting things?
  • Do you lack energy?
  • Do your muscles have less strength?
  • Do you feel weak?
  • Do you have difficulty concentrating?
  • Do you make slips of the tongue when speaking?
  • Do you find it more difficult to find the correct word?
  • How is your memory?

and asks if the factor is

  • less (of a problem) than usual
  • no more than usual
  • more than usual
  • much more than usual

In the PACE trial patients were asked to respond comparing their present state with when they were last well.

Bimodal scoring allocates 0, 0, 1 and 1 to each of these answers, which would give a score of 0 to someone who was perfectly healthy, down to a maximum of 11. In effect, Bimodal scoring simply counts how many of these factors are affected by the illness.

Continuous, or Likert scoring allocates 0, 1, 2, and 3 to these answers, with scores from 0 to 33.

In our survey (more details in section 6) many people commented on the difficulty of quantifying the difference between more and much more. A factor that would be considered by any normal healthy person to be severe can get even worse in a bad spell. They considered that at least one more level, extreme, scoring 4, would have been appropriate. There is a study by Goudsmit et al. looking at the limitations of both scoring methods.

The PACE Trial

Originally, as shown in the protocol, the trial was to have used bimodal scoring, specifying that a positive outcome would be a 50% reduction in fatigue score, or a score of 3 or less. A score of 4 or more would show abnormal fatigue.

Later, after the trial had started, the scoring was changed from Bimodal to Likert/continuous, and a positive outcome was re-defined as a score of 18 or less. This change has led to unease among some scientists who regard a change of an agreed protocol as being acceptable only under very special circumstances.

Difficulties with the Continuous Scoring

It puzzles us how any patient could say that their symptom had become less of a problem than it had been when they were fully healthy. That would mean that a score of zero for any component is very unlikely (although from other studies there are occasional scores which suggest that patients must have scored zero on some components, and we have found subsequently on our own survey that there are rare circumstances that permit this to happen). It would make more sense if patients had been asked to compare themselves with healthy friends of the same age, but this is not the case.

One of the entry conditions was that patients had to score 6 or more under the bimodal system - i.e. they had 6 or more factors that gave them difficulties. Subsequently 18 was used on the Likert scale as a boundary for returning to normal health.

In a relatively small study, looking at cortisol levels in CFS, the Likert scale was used to assess fatigue levels in 40 patients and 40 controls. The patients scored 244 29, and the healthy controls 76 23 (which of course means that the healthy controls generally marked 3 or 4 of the items as being less of a problem than usual). A much larger Norwegian study gave a value of 11.2 as a score for healthy adults. Clearly, the use of a score of 18 to mark "normal function" is, at best, overgenerous, but the differences between the results of different studies prompts the question "Is the Likert scale fit for purpose?"

with the patient below:-


Below is a graph linking the Bimodal scoring system (which simply counts how many items gives problems) to the Likert scoring. Click on each button to see the appropriate graph with an appropriate explanation at the side.


A study published in 2008 by Goudsmit on 24 patients did measure both bimodal and Likert scores as well as whether they rated mild, moderate or severe with ME/CFS. They kindly let us have those scores: they are plotted on the graph below, and show just how complex the situation can be. Notice that one patient (the top right) must have scored zero on two questions.


The overlap between mild and moderate, and between moderate and severe is surprising, as Goudsmit has noted before.

It is clear that there may be severe cases with scores equal to mild cases (here we have an example at 25). The lowest two scores, one mild and one moderate, would not have qualified for the trial, and the moderately affected one would have met the original criteria for normal health.

Click here to add the results of our own survey (described in more detail in section 6).

Notice that five of the people in our survey would qualify for the original trial, as they are to left of the green line, but at the same time would be described as "recovered" - within the normal health scores - as they are above the horizontal black line.

How did that possibility escape the notice of the authors and reviewers? It wasn't missed by Sarah Feehan who subsequently wrote online to The Lancet on behalf of the Liverpool ME Support Group.


more          details (anim)         further graphs           further details          chalder scale

summary          more          details (stills)          further graphs          further details          chalder scale



pdf version  




item1b contact/comments > Home < Return to Summary > 3. It has not further details further graphs details (anim) more here