BACK HOME to Canberra FM/CFS page

SUBMISSION ON DRAFT CFS GUIDELINES

by Dr. Andriya Martinovic

RACP
Dr Graham Stewart,

Re: your request for comments on the "Draft Clinical Practice Guidelines on the evaluation of prolonged fatigue and the diagnosis and management of chronic fatigue syndrome produced by the Working Group convened by the Royal Australasian College of Physicians" (hereafter draft guidelines).

In summary - 5 published controlled trials supporting sub-components of the lipid/EFAM management model, 4 of these after the presentation for publication, plus now the confirming basic science work of the NewCastle group - what more does a GP have to do to get the attention of a predominantly group of specialists supposedly generating guidelines to help GPs ?

1. I refer you to the attached paper to be presented at the Sydney Conference on CFS on the successful predictions, and other corroborating evidence supporting the lipid/EFAM model of CFS management. That paper offers no new evidence beyond what was already forwarded to you/ at your disposal, but summarizes the literature evidence your group failed to mention and/or referenced inaccurately.

2. I refer you also to the various findings of the Newcastle group that will be presented at that conference. These validate various lipid/EFAM model predictions published in 1994, with respect to the role of lipids and EFAMs in the pathogenesis and treatment CFS e.g. the finding of subgroups within the CFS population related lipids/EFAMs.

3. Putting aside for a moment your failure to even mention the lipid/EFAM model of CFS management or its published hypothesis, in my opinion the draft document is seriously flawed. Most tangible of those flaws are various instances of serious literature misrepresentation. In short the conclusions/inferences drawn by the draft guidelines in various instances are unsupportable and/or contradicted by the literature that the draft guidelines claims supports those conclusions. This is clearly unscientific and contrary to the interests of patients, not to mention the "evidence based" claim made by the draft guidelines.

I am of the opinion that these serious scientific flaws mean the draft guidelines cannot serve the stated purpose. Considering the contract offering $200,000 to the RACP and MJA by the Medicare Branch of the Health Department to resource the guideline project, can you explain why there are so many scientific flaws ?? I have structured my comments as questions to ask you. Perhaps the working group can shed some light on how or why the misrepresentation of the literature occurred.

4. The scientific flaws of the draft guidelines compound a string of procedural irregularities by the guideline process. I have only mentioned a sample in the attached and not duplicated the ones I have raised with you already. I find it totally unacceptable, that after Dr Loblay replied to my pointed questions (via email ), he claiming your group were using the NHMRC quality of evidence ratings; that your group then changed the scale of evidence to the effect that it excluded my level III-3 dramatic results in an uncontrolled experiment, and my level IV opinions as an respected authority amongst my peers, in respect of the primary care management/ rehabilitation of CFS patients. Combined with using Fulcher and White to support CBT, but not the science based lipid/EFAM model is further concern.

I also find it unacceptable that my other questions of the working group remain unanswered after almost 12 months.

If your process is to be viewed subsequently as credible science based, you will need to do a more thorough job of tabling all the evidence and considering it objectively.

Sincerely,

Andriya Martinovic
13-2-98

(attachment to introductory letter)

GROSS MISREPRESENTATION OF LITERATURE

Note - "misrepresent" (for whatever reason) = "represent wrongly" Australian Pocket Oxford.

1. - Why do the draft guidelines make a fanciful claim of 50% recovery for CFS in the first year, grossly misrepresenting the cited literature ?

1.(a) Why does page 14 of the draft guidelines seriously misrepresent the literature cited in table 3.2 by claiming :

"The likelihood of spontaneous resolution of CFS of short duration (i.e., 6-12 months) is about 50% over the following year ........ (Box 3.2)"

when :

• of the 7 references in that table 3.2 claiming recovery rates for CFS, only the first reference Calder et al ( 45% recovery at 12 months follow up) is claimed to have more than a 20% recovery at any phase of the illness,
• Calder et al is not a study of Chronic (i.e. long term) Fatigue Syndrome as defined by the working group, but is an analysis of Acute (i.e. short term) viral like episodes after as little as 24 hours of illness.

1.(b) Is it correct that there are no studies cited by the draft report that show anywhere vaguely near the 80% recovery rate at 5 yrs for CFS as claimed by figure 2 on page 2 ?

1.(c) Why does table 3.2 grossly misrepresent so much of the literature cited ?

More Obvious Literature Misrepresentations in Sample First Four Citations Used by "3.2 Natural history of chronic fatigue syndrome"

PRIMARY CARE

• Onset to Presentation/ Study time zero : 28% <24hrs, 40% <1wk, 65%<1mth
  
• Skewed Population : Only Coxsackie B Pos selected out of 140 similar presentations

TERTIARY CARE

• Onset to Presentation/ Study time zero : Minimum 6 wks (not 6 mths)
  
• (Sample size only 144, not 177 & Median follow up 12, not 24 months)

• While this sample size correctly reported, because the other 3 studies were inflated above actual, this appears to downplay this more pessimistic recovery i.e. Calder 63 / 65, Sharpe 144 / 177, Wilson 103 / 149, Hinds 291 / 393
  
• Bombardier & Buchwald 1995 with <5% of 226 recovering, was not even mentioned!

• Sample Size only 291, not 393
• 32% of the 19% did not "recover" i.e. could not return to premorbid physical activities
  
• No "follow-up" period reported - 19% "recovery" was at mean 5.2 yrs after ONSET

1.(d) Would you agree that after Calder et al should be removed from 3.2 (as it is not a study of CFS), there is subsequently no study cited by table 3.2 that finds a 50% recovery rate 12 months after an initial 6 to 12 months incapacity ?

As the 12 month recovery rate for these initially short term viral like presentations are certainly not recovery rate after "unexplained, persistent or relapsing fatigue persistent for six months or more"; Calder et al should be removed from table "3.2 Natural history of chronic fatigue syndrome" and is not even eligible for "3.1 Natural History of prolonged fatigue in primary care".

DRAFT GUIDELINES REVIEW OF EVIDENCE INCOMPLETE

2. Why was the published Lipid/EFAM management model not even mentioned by the draft guidelines, when it has a larger amount of validation through successful predictions and controlled trials than enjoyed by CBT ?

2.(a) Why did the draft guidelines not even mention, let alone describe, the EFAM model, hypothesis or case series despite :

• 5 independent case controlled trials corroborating predictions / individual component treatments of the EFAM treatment model
• voluminous basic science literature supporting EFAM mediated CFS like symptom development; as well as exacerbation and/or relief by exercise and psychogenic factors
• the claim by the draft guidelines preface that they are "following procedures recommended by the NHMRC"
• a published open case series on adjuvant EFAM modulation showed dramatic results qualifying as level III-3 evidence on the NHMRC quality of evidence ratings as published in "NHMRC guidelines for development of clinical practice guidelines" (owing to procedural irregularities such as clandestine modifications to the evidence ratings being made by the working group after submissions closed, the EFAM model failed to attain any evidence rating in the those guidelines)
• the draft guideline preface :
  • acknowledged the 1993 Ministerial CFS Review Committee
  • claimed to have read its report
• EFAM hypothesis and treatment model was a key instigator of the CFS Review Committee, was a key focus of the CFS Review Committee 1993 report, and was acknowledged as CFS research in that report
• p 20 of "NHMRC guidelines for development of clinical practice guidelines" require guidelines to "include a description of each of the alternative treatments".

2.(b) Given the above, plus the additional corroboration of the EFAM hypothesis via further of its previously published predictions being validated by lipid related studies presented at this conference, will the final guidelines address the EFAM treatment model, and lipid related pathogenesis of CFS ?

PROCEDURAL FLAWS

3. Why do the draft guidelines claim to have followed NHMRC guidelines, but do not e.g. in contravention of NHMRC guidelines, they fail to declare the Medicare Branch funding and do not document the contractual obligation of cost effectiveness applications of the CFS guidelines ?

Why do the draft guidelines never once measure the word "Medicare", nor mention "cost effectiveness" given :

• The RACP has a contract with the Medicare Branch of the Health Department offering them $200,000 to develop guidelines on : "cost effective methods of diagnosing and treating Chronic Fatigue Syndrome" (emphasis added)
• a Medicare branch representative attended working group meetings to oversee that pecuniary interest
• The Minister has confirmed in parliament that cost effectiveness application of the guidelines.
• The Draft guidelines claim they were developed by : "following the procedures recommended by the NHMRC"
• The "NHMRC guidelines for development and implementation of clinical practice guidelines" state that clinical practice guidelines should :
  • p 20 "document the specific purpose for which they were developed"
  • p 21 "identify ... who ... funded the guidelines"

BIAS TOWARDS COGNITIVE BEHAVIOUR THERAPY AND LACK OF EVIDENCE FOR GENERAL PRACTICE GUIDELINES

4. If the graded exercise trial is, in accordance with the opinion of its authors, removed from the present table 4.2 regards to CBT, is it not correct that CBT would then share with IV gammaglobulin 2 studies for and 2 studies against and therefore level III-4 conflicting results and not recommended ?

5. However, is it not correct, that none of the authors of the controlled trials cited by the draft guidelines make claims of suitability of CBT for carte-blanch recommendation of CBT to general primary care populations of CFS ?

6. Why do the draft guidelines claim to have evidence based guidelines for general practitioners, when they lack data from unfiltered primary care CFS patient populations on either natural history, or treatment e.g.:

• table 3.2 misrepresents Calder et al to have data on the natural history of CFS in Primary care ?
• table 4.1 misrepresents all the CBT papers as code "5" "Subject group enrolled likely to be representative of the general CFS population" when it is commonly accepted, and elsewhere the draft guidelines is at pains to argue, that tertiary referral centre populations of CFS are biased and unrepresentative of the general primary care CFS population and this document admits its recommendations are primarily intended for GPs ?

7. Why did the draft guidelines misrepresent Friedberg and Krupp as finding no benefit for CBT in CFS, when they found it was of benefit, but only of benefit in more depressed CFS sub populations; thereby contradicting the draft guidelines carte-blanche recommendation of CFS for general CFS populations ?

4/5/6/7.(a) Why was Fulcher and White 1997 seriously misrepresented in text on page 20 :

"The positive studies show a continuing benefit for cognitive behaviour therapy at long-term follow-up (Sharpe et al. 1996, Deale et al 1997; Fulcher and White 1997)"

and also misrepresented in table 4.2 :

"Cognitive behaviour therapy .... beneficial effect three of five studies"

when Fulcher and White:

• did not apply Cognitive Therapy in their study ?
• stated of their Graded Exercise study vs separate CBT studies : "The treatments cannot fairly be compared" ?

4/5/6/7.(b) Why do table 4.1, 4.2 and the text together misrepresent literature references so as to favour the use of CBT unsupported by and/or contrary to the original source evidence?

• Cited as support for WG hypothesis contrary to author's express statements
• No Cognitive Therapy Applied - stated of their Graded Exercise study vs separate CBT studies : "The treatments cannot fairly be compared"
• Filtered sample unrepresentative of General CFS e.g. excluded patients with insomnia before enrolling in Exercise Study
• On Sharpe et al and Deale et al's positive findings for CBT in CFS : "Whether cognitive behaviour therapy is equally effective in the absence of psychiatric disorders is uncertain"

Both above studies on tertiary referral populations that included major depressive disorder & other psychiatric conditions

*- Code 5 in table 4.1 :

"Subject group enrolled likely to be representative of the general CFS population"

But elsewhere the draft guidelines state of tertiary referral settings :

"Such patient samples are biased"

Both Level I and II evidence of scale used by draft guidelines requires evidence to be from "representative patient samples" and these trials fail that test for unfiltered GP CFS populations. The prescriptive extrapolation from unrepresentative tertiary patient populations to: "These guidelines are primarily aimed at assisting general practitioners", is also scientifically untenable given the admission by the draft guidelines

"key confounding variables in studies of CFS include the likely heterogeneity"

and lack of a discriminatory case definition to allow appropriate sub group selection from GP populations based on available biased tertiary referral data samples.

4/5/6/7.(c) Is it correct that when the present table 4.2 is revised to exclude only Fulcher and White in regards CBT, the raw numbers of studies for and against CBT would be the same as found for :

"Intravenous immunoglobulin G .... two of four studies reported benefit" ?

4/5/6/7.(d) What controlled evidence on CBT in CFS justifies anything different to the "level IV" evidence rating and "Not recommended Further studies indicated" given to IV immunoglobulin G by table 4.2. given :

• CBT and IV gammaglobulin share 2 studies for and 2 studies against in table 4.1
• level IV evidence is defined by the working group as "Conflicting evidence from two or more well-designed and controlled studies" while both level I and II would require "consistent evidence" from "randomized controlled studies" ?

4/5/6/7.(e) Why does table 4.1 misrepresent the references under "Behavioural treatments" as the tables code 5* i.e. "Subject group enrolled likely to be representative of the general CFS population" when :

• such claims are contrary to the expressed opinions of the authors cited
• the draft guidelines
  • p12 stress "the diagnosis of CFS currently identifies a heterogeneous group of people"; and under the heading "gaps in our knowledge" it states "key confounding variables in studies of CFS include the likely heterogeneity" and "Future studies must seek to identify potentially homogenous patient groups">.
  • specifically on p15 state of tertiary referral setting data : "Such patient samples are biased" and hence unrepresentative of the general CFS population ?

Note the opinions of the authors whose evidence is claimed to support the draft guidelines hypothesis of benefit of unqualified carte-blanche CBT prescription for general CFS populations :

Sharpe et al: "findings show that patients referred to hospital for chronic fatigue syndrome have a better outcome if they are given a course of cognitive behaviour therapy"

Deale et al: "As chronic fatigue syndrome is heterogenous ..... issues of who benefits from such treatment and how the response rate can be maximized merit further attention."

Fulcher and White : "Whether cognitive behaviour therapy is equally effective in the absence of psychiatric disorders is uncertain"

Note the findings of Friedberg and Krupp that table 4.1 incorrectly claims showed no benefit in CFS:

"those who reported more depressive symptoms showed significant reductions in all measures, while those with lower depression scored did not show significant changes in any outcome measure."

4/5/6/7.(f) As the working group's own quality of evidence scale requires Level I or Level II to have "consistent evidence" from "representative patient samples", how then can either a level 1 or level II recommendation for CBT's use in general CFS populations be made until data from "representative patient samples" is available ?

4/5/6/7.(g) Why then do both the Level II recommendation in table 4.2 of CBT in CFS, and the unexplained promotion to level I recommendation for CBT in CFS found on page 17, fail to carry the acknowledgement of the lack of general practice data for CBT's use in general practice, especially given that the guideline preface states:

"These guidelines are primarily aimed at assisting general practitioners".

4/5/6/7.(h) Is it correct that there is no study referenced by the draft guidelines that offers analysis of effects of Cognitive Behaviour Therapy in primary care CFS patients, rather only tertiary referral centres ?

4/5/6/7.(i) Would you agree that after Calder et al is removed from table 3.2 ( as it is not a study of CFS but a study of acute viral like presentations as early as 24 hrs after onset of symptoms) there is subsequently no study cited by table 3.2 that is upon the natural history of CFS in primary care patients ?

4/5/6/7.(j) Should not the guidelines state : "there is no Level I or II evidence that shows :

• most primary care CFS patients as defined recover spontaneously without treatment
• Cognitive Behaviour Therapy is effective in unqualified general populations of primary care CFS patients "

because :

• the absence of data on the natural history evidence of CFS in general populations,
• the specific statements of the authors of the referenced Cognitive Behaviour trials limiting their findings to specific populations,
• the draft guidelines preface state the draft guidelines are "evidence based guidelines, following procedures recommended by the NHMRC" and p 20 of NHMRC guidelines for development of clinical practice guidelines states:

  "the guidelines should : ......... identify the specific patient population to which they apply, keeping in mind the possible differences in expected outcomes for different populations "

• the draft guidelines also state :
  • preface : "These guidelines are primarily aimed at assisting general practitioners"
  • p 19 "level I or II evidence is regarded as adequate to guide routine clinical practice"
  • p 26 : level I or II recommendations require "consistent evidence" derived from "representative patient samples"
  • variously re CFS being made up of aetiologically heterogenous groups and "key confounding variables in studies of CFS include the likely heterogeneity"
  • on p15 : regarding studies on tertiary referral centre patients, "Such patients samples are biased", and hence unrepresentative of primary care patients generally.

4/5/6/7.(k) Is it correct that according to NHMRC standards, there is insufficient data for evidence based guidelines in general practice management of CFS.

On page 16 of the NHMRC guidelines, the distinction between evidence based guidelines, consensus based guidelines and non-consensus statements is spelt out :

Where there was direct evidence of the outcome of alternative interventions from well-conducted trials, the former classification ( i.e. evidence based guidelines ) would be appropriate. In the absence of direct evidence on outcomes the panel would have to use its judgement to interpret how the non-outcomes based evidence was likely to translate into outcomes in practice. No matter how strong the consensus on this interpretation the guidelines should be classified as consensus-based... Where no good evidence on outcomes is available and where no consensus exists on the preferred method(s) of treatment, the panel should develop non-consensus statements to inform consumers and clinicians.

Continued ...

[ Top ] [ Guidelines page ]

BACK HOME to Moira's Canberra Fibromyalgia and Chronic Fatigue Syndrome Page at http://www.masmith.inspired.net.au/

Moira A Smith - Canberra, Australia
last revised 14 March, 1998