BACK
HOME to Canberra FM/CFS page
by Rose-Marie Johnson
I was a Medical Researcher until I was unable to continue in that career due to my illness, now diagnosed as CFS. I then became a Systems Analyst/Programmer and my training included the analysis of Human Activity Systems. As a past Medical Researcher with CFS, I am a member of the CPAR discussion list for CFS Professionals and Researchers on the internet. This group discusses the wide variety of issues related to CFS research and clinical practice and many of the medical professionals have the additional insights of also being CFS patients. I believe these factors enhance my ability to comment on the Guidelines.
On the other hand, my illness imposes limitations. I was unable to comment on all that I would have liked to. The following was written in little bits over a long period, so is not as neat or concise as I used to be able to write.
It is good how the preface starts by emphasising the importance of clinical expertise and patient preference, in addition to research, in evidence-based clinical practice. It is very unfortunate that the remainder of the Guidelines fails to incorporate these aspects.
It is good that GP's opinions and practices were surveyed and their submissions invited. It is very unfortunate that the Guidelines do not refer to these at all, unless that is where the information for the list of treatments table 4.3 (p20) came from. It seems a shame to sum up so much important clinical expertise with the phrase "scientific evidence is lacking".
It is good that "The Working Group conducted an exhaustive review". It is a pity that a great deal of recent research, especially into physical aspects of CFS, barely rates a mention or not at all.
The goal of "evidence-based" guidelines is a worthy one, but with a condition like CFS, which is made up of a heterogeneous group of patients, there is naturally a lot of conflicting evidence due to different studies having different proportions of sub-types. A great deal of valuable evidence is overlooked by too great an emphasis on "consistent" evidence. The Guidelines did not turn out to be a good example of evidence based medicine - because the list of references seems selective and misses some areas of research and under-emphasises others; because poor statistical understanding resulted in erroneous interpretation of research; because there are gaps and inconsistencies in the presentation of data; and because there are errors in assigning quality of evidence ratings.
GOOD -
"Symptomatic Drug treatment" p3
Statement about increased susceptibility to drugs and starting with small doses (I
experience this).
Please CHANGE -
"Physical activity" -p3
Comments about "mistake of putting off chores..." and dangers of "avoidance
of physical activity" could result in people being encouraged to push too hard on
their bad days. This could result in "worsening of symptoms and delayed
recovery" just as easily as described for pushing too hard on good days. It would be
helpful here to instead discuss the day-to-day (and even within a day) variability of
symptoms and the need to adjust activity level accordingly - neither resting too much or
over- doing it whether it is a good or a bad day. To do this patients need to be educated
about ways of assessing their variable functional capacity. For example what helps me to
push the boundaries of my functional capacity without relapsing too much is noticing when
I am becoming breathless or my heart rate goes up as if exercising hard when active.
Sleep -p3 (see also discussion under section 4 below)
Recommendations are made without supporting evidence, are contrary to my experience and
could cause problems for other patients too.
In general, this section contains many statements without supporting references with evidence. This would be okay in an overview IF they were later backed up by references in the main document but the 2 examples discussed above are not.
GOOD -
"Spectrum" section p5
- about illness, disability and the need to "relieve" suffering. (I don't
actually agree with Jennings 1986 definition of disease, since diseases have existed for
thousands of years before 1986 whether or not the "underlying cause or
pathophysiology" was known.)
Please CHANGE -
2nd column p5, paragraph beginning "In primary care..."
- should not be in this part labelled "What is CFS?" because it is NOT about CFS
at all - it's about chronic fatigue. It should be in the earlier section on "The
spectrum of fatigue states". This intermingling of statements about CFS, chronic
fatigue and prolonged fatigue contributes to confusion throughout the Guidelines. Since
the main title of the Guidelines is "Chronic Fatigue Syndrome", it would be much
clearer to confine statements that don't specifically apply to CFS to a separate
preliminary section about FATIGUE STATES, which would also include the first part about
"The spectrum of fatigue states" as well as the part about other names used for
fatigue states (like neurasthenia) on p6.
Box 1.1 p6
Only lists ONE CRITERION - fatigue - for diagnosis of CFS even though the reference to the
box claims it "outlines the CRITERIA". Since the title is
"Definitions", ALL the symptoms in the Fukuda definition should be repeated
here. This over-emphasis on fatigue and failure to mention other symptoms is continued
throughout the Guidelines and I have discussed it further under the heading GENERAL
COMMENTS.
"Other names commonly used for CFS" P6
Should make it clear that Neurasthenia is NOT another name for CFS, since it's definition
and prevalence are very different - more like chronic fatigue. It would be more accurately
considered as another one of the overlapping fatigue states - another petal in diagram
2.5. Any research using the Neurasthenia label should not be included as evidence in
relation to CFS. Listing Neurasthenia as a name "commonly used for CFS" is
another example of how the Guidelines contribute to a confusing blurring between CFS and
chronic fatigue.
GOOD -
Statements about how CFS differs from depression and somatization. Addressing the issue of "permanent disability" with a reasonably definite statement (p14).
Please CHANGE -
P12 "What laboratory tests are appropriate?"
This section overly emphasises "specific diagnostic test". It should include
other reasons for testing and discuss other tests which could give useful information.
Note that the diagnosis for CFS itself does not include any symptoms that are specific
(any one could occur in other illnesses) - it's the information in the COMBINATION of
symptoms that is important. In the same way laboratory tests that are not specific for
diagnosis could give a doctor important information about the condition of individual
patients, what sub-type of CFS they may have (twice the Guidelines describe the CFS
patient group as heterogeneous) and clues as to useful directions for treatment.
The Guidelines recommend symptomatic treatment and this should be supported by a list of tests for TREATABLE SYMPTOMS associated with CFS. For example, there is replicated research showing that CFS patients may suffer from NMH or postural hypotension and treatment for low blood pressure can help some symptoms (as increased salt and water helped my postural hypotension after it was checked for about a year ago). There is also recent research by more than one group showing cardiovascular abnormalities.
In particular, where there is dispute about whether a patient is physically ill or simply suffering a psychological disorder (eg somatoform or somatization disorders mentioned twice- p11 &p13, or depression), ANY tests showing physical abnormalities not typical of the psychological disorders would be useful (eg. the U. Newcastle urine test, tests which Wakefield et al. showed different results for CFS and depression). There are also possible CONSEQUENCES of the illness that may be important to check for, such as the increased risk of osteoporosis due to low activity levels. This is the sort of information that would be very helpful for GPs treating CFS patients and therefore should form a MAJOR part of guidelines for managing CFS. Instead, throughout the Guidelines there seems to be too much emphasis on checking for associated psychiatric disorder and NONE AT ALL regarding checking for associated PHYSICAL disorders.
GOOD -
Statements about "long term outcome for CFS" and complete recovery being uncommon after 5 years.
Please CHANGE -
p15 "What factors may delay recovery?"
I have studied MANY statistics subjects at university and was a Medical Researcher until a
few months after I became ill. I could understand the statistical errors made by the
Working Group members if they have not studied much statistics, but suggest they correct
their errors in use of statistics before the final document. Also, if evidence based
medicine is a serious goal, the guidelines should be written by somebody with a thorough
understanding of statistics at the university level, so medical statistics can be
effectively used rather than being misapplied as in the Draft Guidelines.
I recommend the book: "How to Lie with Statistics" by Darrell Huff (Pelican Books, ISBN 0 14 02.1300 7). Of particular interest is chapter 8, about correlation.
"The fallacy is an ancient one which, however, has a powerful tendency to crop up in statistical material, where it is disguised by a welter of impressive figures. It is the one that says that if B follows A, then A has caused B" (p85).
B may have caused A or they may both have a common cause or it might even be a chance correlation.
"...All too many medical statistics that are quoted without the qualification that although the relationship has been shown to be real, the cause-and-effect nature of it is only a matter of speculation" (pp86- 87). "Keep in mind that a correlation may be real and based on real cause and effect - and still be almost worthless in determining action in any single case" (p88).
This means that a correlation is NOT useful as a prediction - with any correlation less than a complete correspondence, there are, by definition, exceptions. The book has lots of interesting and amusing examples illustrate these points. These statements are relevant to misuse of statistics in the Guidelines.
I shall address in detail "What factors delay recovery" (p15) as an example of this. DELAY????? This implies cause and effect that has not been demonstrated! This paragraph title is wrong and misleading. The paragraph title should be "What factors ARE ASSOCIATED WITH poorer outcome for PROLONGED FATIGUE". In the context of evidence based medicine, there should be no implication of causality or prediction since there was no evidence for these. The studies referenced showed no more than a relationship (including papers that mistakenly use the words "predictive" or "predictor").
The paragraph talks about studies that show "factors associated" or "relationship". In particular "poorer outcome" being associated with "belief that the illness is purely physical in origin" could be the result of the common factor of physical origin causing both the poorer outcome and the belief.
The studies listed in table 3.2 as showing "belief in viral cause" or "belief in purely physical cause" have not been accurately presented. In particular, Vercoulen et al. (1996) wrote about "RELATIVE absence of physical attributions" not "Belief in PURELY physical cause". These papers also included other related factors and these should have been included in the same paragraph and/or the table. Not to include them gives a false impression of the strength of the belief factor as well as the misrepresentation of causality implied in the word "delay". To include the other factors - eg. avoiding alcohol and changing or leaving employment (Sharpe et al. 1992), less fatigue and shorter duration of complaints (Vercoulen et al. 1996) - would clearly show how inaccurate it is to use the word "delay", with its implication of causality, for associated factors. Nobody would even attempt to suggest that avoiding alcohol would cause a delay in recovery, or that patients should start consuming alcohol along with undergoing CBT!
The Cope et al 1994 paper referenced here is also mis-titled: "Predictors of chronic "postviral" fatigue". It falls for the fallacy that a relationship can be used as a predictor. As quoted above, correlations are almost worthless in predicting single cases. The only exception is when the relationship holds true in ALL cases, which was not the case here - they had to use regression to show the relationship. It is also an example of a selective sample - it studied "postviral" fatigue, whereas the guidelines list other causal factors apart from viruses (Box 5.4). In fact, the Cope paper doesn't apply to CFS at all, since out of the hundreds of patients followed up, none of the patients had a formal CFS diagnosis, and they suggest that only 5 may have had CFS as defined by the looser Sharpe (1991) definition on the basis that they reported being tired "all the time" (a far cry from the actual CFS definition!).
The place for reference to this Cope paper would be in table 3.1 showing "physical beliefs" in the column for "Correlates of persistent disability" in prolonged fatigue. This would make it more clear that the paper is not about CFS. The paper even states that they were only speculating about how the belief could actually be the cause of the poorer outcome and indicates that their research had numerous shortcomings so that further studies are necessary. This seems like very inconclusive research being used to support an erroneous conclusion about "delay" of recovery.
Last sentence in "doctor-patient relationship" p15
The idea of "increased hostility" seems to be no more than a purely speculative
opinion and would probably be best deleted.
The abstract of the Wilson paper also referenced in the above mentioned paragraph stresses the importance of subtypes in research: "Future treatment studies should examine differential response rates for possible subtypes of the disorder (eg, documented viral onset, concurrent clinical depression)...". This principle should have also been applied to the Guidelines' evaluation of treatment studies, but was not. The treatments listed in table 4.1 and discussed in section 4 should have been evaluated according to whether depressed patients were included, whether there was viral onset and also according to the severity of illness of the patients selected.
GOOD -
"Principles of managing people with CFS" p17
Statements about carefully assessing new symptoms and not assuming they are part of CFS.
(I had a difficult year or so when this was not done and my need for surgery was not
identified until after emergency hospital attendance).
Statements about symptomatic treatment on an individual basis and combinations of
treatment.
This is later described as the "cornerstone" of management so MUCH MORE DETAIL
ABOUT SYMPTOMATIC TREATMENT IN SECTION 4 PLEASE!
Please CHANGE -
"Appropriate symptomatic treatment"
should be a LARGE section discussing a wide variety of symptoms commonly experienced with
CFS and a wide range of symptomatic treatments found to be helpful. There are many more
symptoms that significant percentages of those with CFS commonly experience in addition to
those listed in the Fukuda definition. Unfortunately, only a few of these appear in the
Guidelines, for example the lists in diagram 5.5 and some of these are not even discussed
anywhere else.
There should be much more detail about symptoms like blood pressure changes, including studies of NMH and postural hypotension in CFS and their management (for which there is replicated evidence).
In line with statements about symptomatic treatment on an individual basis, this section should include discussion of level III-2 to Level IV evidence for treatments worth trying for particular symptoms in individual patients. For example in some patients (including myself) abdominal pain and bloating (and other symptoms) are improved by a diet for food intolerances and there is published evidence to support this (including publications by Loblay himself). It is unrealistic to only recommend treatments which supposedly have consistent evidence (actually NOT the case for CBT, as discussed below) when the Guidelines recognise the heterogeneous nature of the CFS population. As a consequence, no one treatment is likely to consistently work and the opening paragraph of section 4 correctly recognises that "no single treatment is "curative"".
"Expectations for a treatment trial" p17
This part is exclusively and inappropriately about RESEARCH trials. More appropriate in
"CLINICAL practice guidelines" would be explaining to doctors (especially GPs)
what to expect when trying various symptomatic treatments. For example whether particular
treatments are only helpful for particular symptoms or whether treating certain symptoms
(eg NMH) can improve the overall condition. As suggested by Wilson (1994a)
"differential response rates for possible subtypes of the disorder (eg, documented
viral onset, concurrent clinical depression)..." should also be discussed and this
would be the proper place. In particular, this section should discuss whether certain
treatments such as CBT and graded exercise have been proven to be beneficial in the
absence of depression.
Table 4.2 (p19)is obviously incomplete - there are interventions listed in table 4.1 with "Beneficial effect reported" (eg graded exercise) and/or discussed in text (eg. sleep management, discussed in more detail below) that have been left out of the evaluation for level of evidence. It would be easier to follow if the order of treatments was the same as in table 4.1, especially keeping the CNS agents together. To mix them up results in poor data presentation making table 4.2 harder to follow.
In table 4.2 there is an obvious error in the "Level of evidence" assigned to CBT, since Level II is described as "Consistent evidence" (p26). In contrast, table 4.1 shows inconsistent evidence (2 No as well as 2 Yes) and the text discusses these "discrepant results", which would put it at III-4, "Conflicting evidence obtained from 2 or more well- designed and controlled studies". If level II is to be assigned whenever there are just 2 studies that agree, regardless of any number of discrepant studies, then the whole basis of assessing quality of evidence in the Guidelines is seriously flawed and fails to follow scientific principles. This is not consistent with the assigning of level III-4 to IV immunoglobulin G, which also has evidence for benefit from 2 well- designed and controlled studies (similarly listed in table 4.1 with 2 Yes and 2 No). Just from comparing the 2 tables it seems that if IV IG is not recommended and further studies indicated, the same should apply to CBT.
Since the note about recommendation states "Level I or II evidence is regarded as adequate to guide routine clinical practice" and actually none of the evidence is up to this level, it appears that THE CURRENT LEVEL OF RESEARCH EVIDENCE IS INADEQUATE TO SUPPORT THE DEVELOPMENT OF EVIDENCE BASED GUIDELINES FOR THE TREATMENT OF CFS.
"Role for behavioural treatment" (p20)
In the discussion of CBT there is no mention of many important factors that need to be
considered when a GP, for whom the guidelines are supposedly written, wants to apply the
research results to a particular patient. For example: What particular symptoms improved
and by how much? How were the patients selected for the studies - ie. what CFS sub-type
were they likely to have been? How severe was their initial level of illness? How hard
were the patients already pushing themselves in spite of their illness? What proportion of
patients improved? Are there indications to distinguish patients more or less likely to be
helped by CBT (eg. how well patients are already coping, what activities they are managing
despite the illness)? Since the "physical activity programs...avoid immediate
worsening of symptoms" how can it be applied to patients who have commitments at work
and/or home (eg. raising a family) that keeps them at a level of activity that regularly
causes worsening of symptoms?
In the discussion of possible causes for discrepant results in CBT research, important considerations discussed in other published papers are absent. In particular the issue of the relative effectiveness in patients with or without concurrent psychiatric diagnosis and the issue of scope for improvement, which involves comparison between current activity level versus impairment (Jason LA, Richman J, Friedberg F et al. "Politics, Science, & the Emergence of a New Disease: the Case of Chronic Fatigue Syndrome" American Psychologist, Sept 1997, 52(9):973-983).
Table 4.3 p20 - "Treatments for CFS for which scientific evidence is
lacking"
The title of this table is dismissive and seems to imply that these treatments should not
be considered because of lack of evidence. This is contrary to the first paragraph and
first diagram in the preface of the Guidelines which emphasise the importance of
"clinical judgement" or expertise and "patient preference". It would
be very helpful for patients and general practitioners to have this table re-titled (eg.
CFS treatments found by some doctors to help some patients) and supported by a discussion
of the types of symptoms helped and other possible benefits. It seems a great shame to sum
up so much important clinical expertise with the phrase "scientific evidence is
lacking" without any discussion of actual clinical experience.
p 3 "Sleep" and p21 "What is the role of sleep
management?"
The papers referenced in this section show evidence for "changes in sleep" as
stated in the opening sentence - they do not include evidence for the following
recommendations for "management". These are made twice (and the list is
different in the 2 places, indicating sloppy work), without addressing the issue of
evidence in CFS patients at all! Neither the papers referenced nor any of the papers or
abstracts about sleep and CFS that I have seen have proven that any of the suggestions
made under these sections are of significant long-term benefit for CFS patients. This is
the standard of evaluation applied by the Guidelines to other treatments and should be
similarly applied to the suggestions for sleep management.
The listed suggestions regarding sleep should not be made without inclusion in tables 4.1 under "Behavioural treatments" with an evaluation of "Quality of study design and analysis" and in table 4.2 with "Level of evidence".
Their first suggestion (p3) is "avoiding day-time naps". In my personal experience, if I don't manage to have a nap when I need it, I completely crash out and cannot continue without sleep way before my bed time. I then awaken in the middle of the night needing to urinate and take hours to go back to sleep. This is far more disruptive to my regular sleep pattern than having a nap when I need it, and then being able to last until my regular bed time and follow my usual sleep pattern. It is also more disruptive to my family, because I end up too brain fogged to prepare a proper dinner and too exhausted to clean up afterward.
As far as I can recall from studying sleep at university, extra sleep has an important role in facilitating healing in sick animals and people. To me, avoiding naps and restricting night-time sleep to a usual 8-9 hours when ill with CFS (the Guidelines do agree CFS patients are ill) makes as little sense as avoiding day time sleep and trying to limit to normal sleeping hours when ill with anything else, like the flu.
The advice about restricting night-time sleep and reducing daytime naps would cause problems for me and I expect it could cause problems for many other CFS patients. Such potentially damaging advice should not be made without a proper evaluation of any supporting and contradictory evidence.
GOOD -
Discussion of "heterogeneity within patient groups" and implications for finding treatments. Describing symptomatic therapy as "one of the cornerstones of management" BUT please leave out the word "drug" as non-drug therapies are very important too (even the CBT promoted by the Guidelines is a non-drug therapy).
Please CHANGE -
"Major research findings" p22
The first 3 points listed as "leading hypotheses" are backed up with evaluation
of evidence in tables 5.1-3, however the last 2 are missing from the boxes and there is no
evaluation of the evidence for the references given. In fact box 5.3 under
"Mood" lists physical evidence that CFS is different from depression,
conflicting with the suggested hypothesis that it is linked to depressive disorders. These
last 2 hypotheses should therefore be REMOVED from the list of "leading
hypotheses".
Box 5.4 "other factors" p24
The level of evidence is left out for "Metabolic disturbance". Since there are 2
references from one group, it looks like it should be III-1.
It would be more helpful to have a more complete list of possible other factors. For example, Dr Loblay himself has many publications about food intolerances and has written a chapter in one of Byron Hyde's books about his group's studies of its role in CFS. The more other possible factors that are listed in this section, the greater the chances that individual patients will be helped by their doctor's ability to identify the factors involved in their particular case. Even information about factors that only occur in a few percent of patients could have a major impact on management of the illness for those particular people. Again it is the heterogeneity of CFS that makes this so important.
Diagram 5.5 p25
A detailed discussion of this diagram would be helpful, especially where it suggests
observations that a general practitioner may be able to make when examining patients eg.
"Abnormal blood pressure responses to postural change" and "Abdominal pain
and bloating".
Quality of evidence ratings are good in theory, but it would be very useful if this section discussed the practicalities of applying them to research into the heterogeneous CFS population where sub-types have not been identified. The discussion on conflicting evidence in particular should include a discussion of the heterogeneous aspect. Also the issue of different definitions of CFS used by different researchers should be discussed.
It would also be useful in this section to detail the other types of guidelines discussed in the "NHMRC Guidelines for the development and implementation of clinical practice guidelines". This includes "consensus-based" guidelines, based on consensus of expert opinion, and "non-consensus practice statements" detailing the opinions of clinicians. Considering the dearth of consistent evidence relating to CFS (largely due to its heterogeneity but also contributed to by differing definitions of CFS), evidence based guidelines seem to be impracticable at this stage. Considering the lack of consensus amongst experts, particularly in physical versus psychological areas, it would seem that consensus- based guidelines are not possible at this stage either. It seems that a practice statement including more details of the opinions of clinicians would be the most useful type of guidelines at this stage.
See comments on particular boxes etc. under section headings above.
Readability
: Good.To GPs
: Poor - Too much emphasis on research and insufficient on clinical practice.Awful
- The only mention of children is that they are up to 20% of CFS cases (P15). A section
with detailed discussion of the special needs of children with CFS (eg. managing their
education) would be very helpful. - Much research (eg NMH, RNAase-L, urine metabolites)
not discussed.
- Clinical experience of surveyed GPs not evident in the Guidelines.
- Insufficient detail regarding tests for symptoms which may be present in significant
percentages of people with CFS.
- Insufficient detail about a wide range of symptomatic treatments that have helped
subsets of CFS patients.
- No details about checking for other overlapping conditions shown in diagram 2.5 - eg FM
and IBS - apart from depression.
Also see above comments about particular sections.
Awful
- Fatigue is over-emphasised.
- Physical symptoms and their treatment are under-emphasised.
Except for box 1, p1 which lists 'other symptoms' that are part of the Fukuda definition, wherever the Guidelines list CFS symptoms the list is incomplete. In particular, box 1 seems to be the only place which mentions "sore throat" and "tender cervical or axillary lymph nodes". These symptoms are especially important because they are ones that a doctor can often see for themselves and also because they help to distinguish CFS from other disorders such as depression. These 2 symptoms in particular, and preferably the whole Fukuda list, should be included wherever CFS symptoms are listed in the Guidelines eg. Box 1.1 p6, bottom of p8 and Fig 2.5 p13. To leave out these extra symptoms in the discussion erroneously promotes blurring of the distinction between CFS, chronic fatigue and psychological disorders such as depression (which can also have pain and affect sleep etc).
Only the Fukuda definition is mentioned. There needs to be discussion of which research referenced in the Guidelines actually based their patient selection on this definition and, if not, how they defined CFS for their research. In particular, the likelihood of certain research patient populations including depressed patients who would not satisfy the Fukuda definition needs to be addressed.
Rose-Marie Johnson
Patient and Medical Researcher (until a few months after onset of illness)
February 1998
[ Top ] [ Guidelines page ] [ Australian/NZ information] [ Library]
BACK HOME to Canberra Fibromyalgia and Chronic Fatigue Syndrome Page at http://www.masmith.inspired.net.au/
Moira A Smith - Canberra, Australia
last revised 15 Sept, 2000