Defining the minimally clinically important difference of the SF-36 physical function subscale for paediatric CFS/ME - Crawley et al

Defining the minimally clinically important difference of the SF-36 physical function subscale for paediatric CFS/ME: triangulation using three different methods

• Amberly Brigden
• Roxanne M Parslow
• Daisy Gaunt
• Simon M Collin
• Andy Jones
• Esther Crawley

https://link.springer.com/article/10.1186/s12955-018-1028-2

 

NOTE: I'm unusually sleep deprived, so this may be nonsense.

"Ethics approval and consent to participate
Qualitative data: Full ethical approval was obtained from the NRES Committee North West (08/04/2014, ref 14/NW/0170). An amendment to use one-to-one interviews with children, parents and clinicians who are unable to attend focus groups and to include exploration of important outcomes was approved (21/10/2014, ref 14/NW/0770). Relevant R&D approval was obtained from the RNHRD (20/06/2014, ref-RBB 427).

Quantitative data: The quantitative data was collected routinely as part of service evaluation. The North Somerset & South Bristol Research Ethics Committee determined that collection and analysis of these CFS/ME patient data did not require ethical review by an NHS Research Ethics Committee or approval by NHS Research and Development offices (REC ref. 07/Q2006/48)."

Looks like they did also get ethics approval for interviews this time, so maybe this doesn't matter, but re the Quantitative data, isn't that the same code @dave30th has written about being misused elsewhere?

re this study: Looks like they're just relying on subjective self-report outcomes, rather than assessing what level of change in SF36-PF was reliably associated with a significant improvement in some objective outcome, like actometers. I also couldn't see any reference to them having explained potential problems with thigs like response bias to the participants who were giving their views on what a MCID would be. So this study is of no value for deciding what should be viewed as a MCID in the sort of nonblinded trials Crawley has been conducting.

Can Crawley's team get patients to validate her past spin!

Wouldn't this depend on what the treatment was, and how much effort it was?

I guess it's obvious that they'd value different items very differently, but doesn't that make the idea they they've endorsed a MCID of 10 difficult to justify? It sounds like it really depends on which two items improved. The way they phrased things in the quote before last would mean that participants will have been encouraged to choose the two items that they valued more than any other, whereas an MCID of 10 would mean that improvement in any two items would be seen as a MCID.

This is a tiny bit interesting, but would be so much more interesting if they'd looked at correlations with an objective outcome:



I don't think this paper contributes much. An MCID of 10 in a rigorously conducted double-blind trial sounds reasonable (although this paper actually makes me think that many children with CFS would only agree if the change came from their prioritised items). This paper provides no reason to think that a change of 10 indicates a clinically significant difference in nonblinded trials where participants in one arm were encouraged to adopt positive views about their treatment or health.

 

When will we get some decent objective paediatric research?
Why is this such a big ask?

Unfortunately, i don' t think it will be in the UK

 

I just got re-annoyed about this study, but that's possibly because I'm sleep deprived.

Is it just me, or does it look as if they just wanted to do a science-looking experiment, rather than get more useful information by just asking kids what they thought should be the MCID?

As they note, kids value changes in different items very differently, so their experiment is likely to lead to a lower MCID than if they'd just explained to kids what an MCID was used for, and then asked them what they thought a good MCID would be.

They set up this 'experiment' so that every child will have been using the items that they most valued change on to set the MCID. It could be that if the kids involved had been asked what change should be used to estimate a MCID in clinical trials or observational studies, when an MCID of 10 would mean improvement for any two items and it should not be assumed that the first two to change will be the two most valued by the patient, that they'd have thought an MCID of greater than 10 would be appropriate.

I don't understand why they'd do things the way they did, unless they just wanted to make that bit look a bit more 'sciencey' than a poll of 21 kids.

 

But the numbers ticked made it "quantitative" so it is proper science!

 

I am worried that Crawley (and others like her) are training up PhD students to do cargo-cult science that lets them get a career but contributes nothing to society.

 

it's even worse than that:
https://www.s4me.info/threads/nihr-fellowships-2017-round-11.5095/

 

Sorry for the necro. Do we have any reliable studies showing MCID/MID for the major symptom questionnaires currently in use? For instance, for the De Paul Symptom Questionnaire.

 

Correction: Defining the minimally clinically important difference of the SF-36 physical function subscale for paediatric CFS/ME: triangulation using three different methods

The original article [1] required an amendment to the Quantitative Data statement in the Ethics Declarations section. The statement has since been updated.

https://hqlo.biomedcentral.com/articles/10.1186/s12955-022-02078-7