Efficacy and safety of Sijunzi Decoction for chronic fatigue syndrome: study protocol for an RCT, Dai et al., 2019

[Deleted member 3253]

(Not a recommendation.)

Abstract: http://atm.amegroups.com/article/view/30395
PDF: http://atm.amegroups.com/article/download/30395/pdf

 

[Marky]

This seems like an typical "lets research something so that we get paid"-study

 

[rvallee]

Completely useless.

The primary outcome will be the change of Chalder fatigue scoring after treatment. Secondary outcomes include the short form-36 physical function subscale (SF36-PF), spleen deficiency rating scale, quality of life and self-rated clinical global impression (CGI) scales.

 

[Hutan]

I'm not following why it's completely useless.

It's double blind and placebo controlled. The sample size is good - 212 participants, allocated to either the treatment or placebo. The treatment time of 2 months, with a one month followup, is perhaps a bit short but it's not terrible.

The Chalder fatigue scale is a bad instrument, but some of the secondary outcomes might be better (eg SF36 Physical function, clinical global impression, quality of life). Subjective outcomes are ok with a double blind trial. The spleen deficiency rating scale is a bit random of course, but still, with questions such as 'easy getting cold', 'insomnia', I've seen sillier things.

The selection criteria (CDC? - 6 months of fatigue and 4 of a list that includes but does not require PEM) is not good. The protocol is written up clearly with things that you'd expect from a well run trial. The placebo and treatment will be packaged identically, with only a randomly allocated serial number allowing traceability. Adverse events will be monitored.

Yes, there's probably a fair degree of bias in the researchers. But that only makes the trial completely useless if you are confident that the researchers are going to be fraudulent.

 

[Cheshire]

I'm not following why it's completely useless.

The Chalder fatigue scale is a bad instrument, but some of the secondary outcomes might be better (eg SF36 Physical function, clinical global impression, quality of life).

I'm afraid the use of the Chalder fatigue scale could muddle the results. It seems so unreliable and ineffective to evaluate anything seriously.

 

[Jonathan Edwards]

I think we can be reasonably confident that anything based on ginseng root will be completely useless. I suspect nine hundred and seventy three different diseases have been treated with ginseng and as far as I know there is no evidence that it helps any of them.

I agree that if we really want to know that things that do not work do not work then this might be an OK design but it is pretty easy for blinding to be incomplete - which gets you back to the usual problem.

 

[Hutan]

I agree that if we really want to know that things that do not work do not work

There's the vast country of China with TCM practitioners diagnosing spleen deficiency and prescribing Sijunzi decoction to people with CFS-ish illnesses. And plenty of people in other countries spending lots of money on TCM consultations and products.

I think it's very useful to know if it works or not.

The design is pretty robust with respect to blinding - there are third parties preparing the packages. Like I said, fraudulent behaviour could subvert it, but that doesn't mean the protocol is completely useless.

 

[Jonathan Edwards]

But if ginseng tastes, all is lost!

 

[Hutan]

The placebo is the simulant granule, which the main components are starch and dextrin. In order to present comparable color, smell, taste and texture with SJZD, food colorants and flavoring agents are added. Both SJZD and placebo are prepared as soluble granule by a third company, which is qualified in CHM processing and preparation. The third pharmaceutical company will equally divide daily dose (SJZD or placebo) into two individual pack. Participants from either group will be required to dissolve one pack with 200 mL hot water and take the decoction orally twice daily for 2 months. Both SJZD and placebo will be packed in sealed medicine box. Only the serial number will be printed outside the package to ensure successful blinding.

The protocol suggests decent efforts are being made. Yes, the food colourants and flavouring agents might have impacts of their own (good or bad).

I just think 'totally useless' goes further than is justified. I'd rather keep that assessment for the likes of the shopping bag studies e.g. studies with no blinding and subjective outcomes.

I'd prefer that these researchers felt able to come on this forum so that we could talk about things like selection criteria and the Chalder scale (and their idea that the BPS theories have some validity) rather than felt completely dismissed.

 

[rvallee]

The protocol suggests decent efforts are being made. Yes, the food colourants and flavouring agents might have impacts of their own (good or bad).

I just think 'totally useless' goes further than is justified. I'd rather keep that assessment for the likes of the shopping bag studies e.g. studies with no blinding and subjective outcomes.

I'd prefer that these researchers felt able to come on this forum so that we could talk about things like selection criteria and the Chalder scale (and their idea that the BPS theories have some validity) rather than felt completely dismissed.

It really should not be necessary to explain to researchers why primary subjective outcomes are bad. Even if the rest is moderately competent this just makes it all too suspicious to take seriously.

It is better than the typical BPS stuff, but understanding science 101 is still a requirement to be taken seriously. Even when done correctly science is hard to interpret, to separate noise from signal. The CFQ is simply uninterpretable and lacking any merit, this is too basic a failure to bother any further.