The Relation Among Body Consciousness, Somatic Symptom Report, and Information Processing Speed in Chronic Fatigue Syndrome
May 10, 2002
Sieberen P. van der Werf, M.Sc.*; *Berna de Vree, M.Sc.; **Jos W.M. van der Meer, M.D.; *Gijs Bleijenberg, Ph.D.
Departments of *Medical Psychology and **General Internal Medicine, University Medical Center Nijmegen, the Netherlands.
Address correspondence and reprint requests to Sieberen P. van der Werf, University Hospital Nijmegen, Department of Medical Psychology, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail: S.vanderWerf@cksmps.AZN.NL.
Received April 2, 2001;
revised June 28, 2001;
accepted August 27, 2001.
Objective: The aim of this study was to assess the potential influence of body consciousness and levels of somatic symptom report upon information processing speed in patients with chronic fatigue syndrome (CFS).
Background: According to a model of a fixed information processing capacity, it was predicted that in a group of patients with CFS, high body consciousness in combination with a high report of somatic symptoms would affect information-processing speed negatively.
Methods: Information- and motor-processing speed were simultaneously measured with a simple- and a choice-reaction time task, whereas cognitive complaints were rated with two questionnaires. The hypothesized influence of private body consciousness and somatic symptom report upon information-processing speed was tested in a model. A symptom-validity test was used to screen for possible illness behavior.
Results: Private body consciousness was directly related to information-processing speed and somatic symptom report. Somatic symptom report was related to both test performance and memory and concentration complaints.
Conclusions: Levels of private body consciousness directly affected somatic symptom report and information-processing speed. This finding supports the role of attentive processes in CFS, and offers, besides possible cerebral dysfunction, an alternative explanation for slowing of information processing in CFS.
Chronic Fatigue Syndrome (CFS) is a medically unexplained clinically defined condition, characterized by long-lasting (at least 6 months) severe disabling fatigue, and a combination of at least four out of eight symptoms (1). One of these eight major symptoms is self-reported impairment in concentration and short-term memory. Despite the fact that many patients with CFS report problems with concentration and memory, neuropsychological findings are not consistent (2-10). Recent reviews suggested that these varying results were caused by differences in diagnostic criteria for CFS, the heterogeneity of the CFS population, small sample sizes or methodological diversity (11,12). In a recently published study we also found evidence that the neuropsychological test performance in comparative studies could have been biased by illness behavior (13). Despite these possible confounders, a possible reduction in information processing speed has been reported most consistently (11,12).
In a tested model for CFS, focusing on bodily symptoms, as measured by high somatic symptom report, turned out to be strongly related to both fatigue severity and experienced disability (14). Several studies showed that a wide variety and high frequency of somatic symptoms characterized patients with CFS. Compared with patients with multiple sclerosis or depression, patients with CFS reported significantly more somatic symptoms on the somatization subscale of the Symptom Checklist-90 (14-16). These findings suggested a heightened bodily self-awareness in CFS.
In the literature, the following two constructs of self-focused attention have been distinguished: private self-consciousness and private body consciousness (17,18). Private self-consciousness relates to the tendency to attend to inner psychological or emotional processes, whereas private body consciousness reflects the person's sensitivity to perceive internal bodily sensations. Several studies showed that private body consciousness was related to symptom report and symptom severity, problem solving and attentional processing (19-21).
A recent study, using 57 nonclinical subjects, demonstrated that a heightened selective attentional focus on the body and the degree of symptom report predicted levels of illness anxiety (23). A small, though significant, relation was found between performance on a continuous performance test, measuring visual sustained attention, and the extent subjects reported private body consciousness. Higher levels of private body consciousness were associated with a lower sensitivity index of the continuous performance test. Based on their findings the authors hypothesized that bodily preoccupation seemed to interfere with external task demands, making less attentional resources available for processing of external information.
Martin et al (24) studied the relation between private body consciousness and state anxiety in the report of somatic symptoms during a clinical magnetic resonance imaging (MRI) investigation. They found that state anxiety and private body consciousness interacted to predict the report of symptoms during the MRI procedure. These findings were explained using the information-processing model of Ahles et al (25). According to this model, anxiety interacts with the predisposition to focus on bodily sensations by increasing the likelihood that these sensations will be processed affectively. This type of affective processing amplifies the somatic sensations, which are then experienced and reported as symptoms.
Klein et al (22) assessed in 45 students whether a fixed-capacity information-processing model could be used to explain the relationship between problem-solving errors and life stress as moderated by anxiety and private body consciousness. The data showed that high state anxiety in combination with high life stress affected the performance on an analogical reasoning task negatively, especially in subjects with high levels of private body consciousness. It was concluded that stressful life events, and the task-irrelevant thinking that characterizes state anxiety, occupied some portion of a finite processing capacity. Furthermore, the authors reasoned that anxiety is often accompanied with somatic sensations and that subjects who have high levels of private body consciousness will be more attentive to these sensations, leaving them with even less recourse for cognitive demanding tasks.
The aforementioned studies suggested that private body consciousness was associated with increased focusing on bodily sensations and higher somatic symptom report, and might affect decision-making and information processing capacity negatively. Therefore, the objective of the current study was to test a model that incorporated both body consciousness and the extent patients report somatic symptoms. Body consciousness was considered an independent predisposing factor for the detection and subsequent report of somatic symptoms, whereas it was assumed that the report of somatic symptoms could also represent underlying psychosocial processes. According to the reports in the literature, body consciousness was hypothesized to directly influence information-processing speed, especially in subjects that are characterized by high somatic symptom report. Because the somatic symptom report could involve different causal processes, such as active focusing on bodily sensations, anxiety, or perhaps even some form of illness behavior, the influence of somatic symptom report upon neuropsychological test-performance was expected to be less specific and not only to be limited to information processing speed. Therefore, the model was also tested with motor-speed measures and measures of self-reported cognitive complaints. To control for the possibility of response bias influencing both neuropsychological performance and interrelations between variables, the proposed models were tested for the sample that showed valid performance on a symptom validity task.
MATERIALS AND METHODS
This study was conducted at the Department of Medical Psychology of the University Medical Center Nijmegen, the Netherlands. Complete data were collected in 57 (44 females and 13 males) patients with a Center for Disease Control and Prevention diagnosis of CFS, aged between 16 and 56 years, and fulfilling operational criteria for fatigue severity and disability. Informed consent was obtained before the start of the study (1). The mean age of this group was 34.7 years and the mean illness duration was 1.9 years (range, 1 to 4 years).
Private Body Consciousness
The subscale private body consciousness of the Body Consciousness Questionnaire was used to measure this concept (26). This subscale has 5 items that can be answered on a scale that ranged from 0 (extremely
uncharacteristic) to 4 (extremely characteristic). The private body consciousness scale had the following 5 items: (1) I am sensitive to internal bodily tensions; (2) I know immediately when my mouth or throat gets dry; (3) I can often feel my heart beating; (4) I am quick to sense the hunger contractions of my stomach; and (5) I'm very aware of changes in my body temperature.
A validation study showed that patients who scored higher on the private body consciousness scale were stimulated by caffeine (18). The instrument has been translated and psychometrically evaluated in the Dutch population. The evaluation of the translation resulted also in a three-factor structure and Cronbach alpha comparable to the original test (ranging between 0.63 and
Degree of Somatic Symptoms
>From the Symptom Checklist (SCL-90-R) the subscale somatization
(SCL-somatization) was used (28). This subscale rated the burden patients experienced from a number of psychosomatic symptoms.
Information Processing Speed and Motor Speed
A simple- and a choice-reaction time task were administered to measure motor and mental speed under two conditions. This test has been used in an aging study of a normal population and a previous neuropsychological CFS study (2). The aging study showed an age effect for both the motor speed and the information processing speed measures (29).
Five target buttons were situated on a response board at equal distance around a start button. Each target button contained a stimulus light. During the two tasks, the subject kept the start button pressed, until a stimulus lit up. In the first task only one stimulus button (the middle) could light up. In the second task, three different target buttons could light up in random order. In both tasks subjects were asked to respond as fast as possible to the button that lit up by releasing the start button and moving to and subsequently pressing the target button. In both conditions a distinction could be made between speed of information processing (stimulus selection, choice selection and response initiation) and motor speed (movement time between releasing start button and pressing the target button). The neuropsychological test resulted in two movement times (MT1 and MT2) and two reaction times (RT1 and RT2).
Subjective Memory and Concentration Problems
The subscale alertness behavior of the Sickness Impact Scale (SIP alertness
behavior) (30) and the subscale Concentration of the Checklist Individual Strength (CIS-concentration) were used to measure the degree of subjective memory and concentration problems (15).
The Amsterdam Short Term Memory Test (ASTMT) is a forced choice verbal recognition task developed to detect under performance during neuropsychological testing (31). This test has been described in detail in a previous study. The maximum score is 90 points (3-30 items). Validation studies showed high average scores (> 89) for healthy subjects and patients with mild Closed Head Injury (CHI). Based on different validation studies a cut-off value of 86 points has been proposed (32).
Structural equation-modeling techniques were used to test the model (33). Calculations were performed with the computer program AMOS 4.0 (34). The maximum-likelihood method of estimation was used to estimate unknown regression parameters.
In the proposed model (Fig. 1), information-processing speed was defined as a latent construct with the two reaction times being the measurement variables. Information processing speed was thought to be directly influenced by body consciousness, somatic symptom report, and age, whereas body consciousness also had a direct effect upon somatic symptom report. For cross-validation purposes this model was also tested with the motor speed measures and the self-report measures as separate latent constructs.
The chi^2 statistic and the Adjusted Goodness of Fit Index were used to test data fit. According to guidelines for structural modeling, the data were considered to fit the model when chi^2 statistic was not significant and Adjusted Goodness of Fit Index was high (preferably above 0.90). Furthermore, the population error of approximation (Root Mean Square Error of
Approximation) was taken into account as measure of discrepancy per degree of freedom. Values up to 0.05 indicate a close fit and value up to 0.08 represent reasonable errors of approximation in the population. Preferably the 90% confidence interval (90% CI) of the RMSEA index lies between 0.00 and 0.10 (35). When the data did not fit the model well, the modification indices provided by the AMOS program were used to investigate whether the model could be adjusted according theoretically valid propositions
Thirteen (23%) of the 57 patients had ASTMT scores possibly indicative of under-performance. Hence, model testing took place in a sample of 44 subjects (32 females and 12 males).
Table 1 depicts mean age and illness duration of this sample together with the descriptives of the model variables. Table 2 shows the interrelations of the model variables in this sample of 44 subjects.
Both the chi^2 index and AGFI index indicated good fit (chi^2=2.6; df=8; p=0.96; AGFI=0.94). The 90% CI of the Root Mean Square Error of Approximation also indicated a close fit of the model in relation to the degrees of freedom (RMSEA=0.00; 90% CI, 0.00-0.00; p=0.97).
The direct relation between Body Consciousness and Information Processing Speed did reach significance (beta=0.32; p=0.01). Level of somatic symptom report had also a direct significant effect upon information processing speed (beta=0.43; p<0.01), whereas Body Consciousness was significantly related to level of somatic symptoms report (beta=0.30; p=0.04). Education had significant relation with information processing speed (beta=-0.25; p=0.04). However, age was not significantly related to information processing speed (beta=0.19; p=0.12). In total, 46% of the variance in information processing speed was explained by the model.
Testing the Model With the Motor Speed and Self Report Measures Motor Speed
The chi^2 and AGFI indices indicated good fit (chi^2=4.2; df=8; p=0.54; AGFI=0.92). The 90% CI of the Root Mean Square Error of Approximation indicated a less close fit of the model in relation to the degrees of freedom (RMSEA=0.00; 90% CI, 0.00-0.11; p=0.87).
In this model, the direct relation between Body Consciousness and Motor speed did not reach significance (beta=0.20; p=0.14), but level of somatic symptom report did have a significant effect upon Motor Speed (beta=0.30; p=0.03). There was a direct effect of Body Consciousness upon level of somatic symptom report (beta=0.30; p=0.04). Age was also significantly related to motor speed (beta=0.33; p=0.01), but education was not (beta=0.00; p=0.96). In total 27% of the variance in motor speed was explained by the model.
Experienced Memory and Concentration Problems
In this sample, the model showed reasonable fit of the data (chi^2=6.2; df=8; p=0.77; AGFI=0.88), but the RMSEA 90% CI indicated a less close fit in relation to the degree of freedom (RMSEA=0.00; CI, 0.00-0.15; p=0.70). The direct relation between Body Consciousness and level of somatic symptom report did reach significance (beta=0.32; p=0.03), and level of somatic symptom report had a significant direct effect upon experienced memory and concentration problems (beta=0.43; p<0.01). However, there was no direct significant effect of Body Consciousness upon experienced memory and concentration problems (beta=0.23; p=0.11). Age was significantly related to experienced memory and concentration problems (beta=0.29; p=0.04), whereas education was not (beta=0.18; p=0.20). In total 42% of the variance in experienced memory and concentration problems was explained by the model.
In the introduction, two studies were described that reported private body consciousness was associated with increased focusing on bodily sensations and higher somatic symptom report, and affected information processing negatively in nonclinical populations (19,21). Similarly, Martin et al (20) reported that in a clinical population, body consciousness and state anxiety levels predicted the report of symptoms. In all three studies, attentive processes were thought to play an important role in the report of symptoms and the processing of external information because patients with CFS report high levels of somatically unexplained (physical) symptoms, frequently show slowing of information processing during formal testing, and generally believe that their illness has a physical cause (2,15,16). We hypothesized that such information-processing models might be used to explain possible relations between somatic symptom report and processing of external information in CFS.
The findings of this study indeed lend some support to the proposition that in CFS attentive processes, such as private body consciousness, might affect somatic symptom report and either directly or indirectly interfere with external task demands. Therefore, similar mechanisms as described in the information-processing model of Ahles et al (25) could possibly be applied to CFS. Previous CFS research has stressed the importance of illness cognitions. Physical causal attributions and self-efficacy were found to predict persistence of symptoms in CFS. One could hypothesize that the conviction of the patient that there is physically something wrong, and that one cannot influence their complaints, does increase body consciousness and state anxiety concerning symptoms and exertion. Instead of labeling slowing of information processing as a direct result of cerebral impairment or deficit, it could also be interpreted as a consequence of affective processing of symptoms or a too strong attention for bodily sensations.
Because motor output is independent of stimuli processing, one would expect that body consciousness or active focusing on bodily symptoms would to a less extent associated with motor speed. The results of the model testing were in agreement with this proposition. A previous study of our research group, using similar tests and self-report measures, showed that there were no significant associations between neuropsychological test performance and experienced neuropsychological problems, but experienced neuropsychological problems were related to both fatigue severity and depression (2). The findings of the present study also suggest that the expression of cognitive complaints is associated with somatic symptom report and not merely reflected in information processing speed.
The motor planning task has been used in a cohort aging study of healthy subjects and subjects who had biologic life event possibly affecting brain function. In contrast to this study, age was not associated with information processing speed. However, the aging study in which a relation between age and information processing speed was reported, pertained to a large sample of healthy controls who's age ranged between 17 and 84 years (29). In comparison, our sample was relatively young, and most patients had short illness duration; thus, there was less variation in age. It is possible that other factors than age dominated test performances in this clinical sample. Furthermore, the sample that provided data to test the hypothesized models was small; therefore, the power to detect significant relationships (standardized regression coefficients) was restricted to medium effect sizes. Our results would gain validity when they could be replicated in other CFS samples with different age distributions. In addition, one could verify whether similar mechanisms as found in the current study could apply to other somatically unexplained chronic conditions.
To what extent attentive processes play a role in CFS does need further research. It would be interesting to compare early and late components of somatosensory event related potentials during an experimental task in which attention will be manipulated by a distraction task. Similarly, one could test the effects of somatosensory stimulation upon a continuous performance task. The clinical implications could be a therapeutic focus upon relabeling bodily sensations, diverting attention to bodily sensations by distracting tasks, or exposing to physical exertion to increase bodily sensations and, as such, possibly blunt the affective response to less intense bodily stimuli. These treatment strategies could be part of either a cognitive behavior therapy or a graded exercise protocol, both of which have been proven effective in CFS (36,37).
TABLE 1. Demographics and average test scores of model variables in the
sample with valid ASTMT scores
Minimum Maximum Mean SD
Age 16.0 56.0 34.1 12.0
Illness duration 1.0 4.0 1.9 0.7
Body Consciousness Scale 7.00 20.00 15.8 3.1
SCL-somatization 16.00 52.00 29.0 8.6
RT1* 0.22 0.73 0.32 0.10
RT2 0.25 0.80 0.37 0.12
MT1 0.14 0.42 0.23 0.07
MT2 0.15 0.48 0.24 0.07
CIS-concentration 6.0 35.0 27.1 6.9
SIP-alertness behavior 0.0 702.0 338.2 204.7
*Reaction (RT) and movement (MT) times in seconds. Sample total (n=44).
Valid ASTMT scores (>85). ASTMT=Amsterdam Short Term Memory Test, CIS=Checklist Individual Strength, SD=standard deviation, SIP=Sickness Impact Scale.
TABLE 2. Pearson correlations between model variables in sample with valid
Age Education SCL-S BCS RT1 RT2 MT1 MT2 CIS-C SIP-AB
Education -0.14 1.00
SCL-S -0.08 0.02 1.00
BCS -0.04 -0.01 0.32 1.00
RT1 0.14 -0.22 0.43* 0.43* 1.00
RT2 0.12 -0.24 0.52* 0.42* 0.86* 1.00
MT1 0.30* -0.01 0.34* 0.29 0.54* 0.54* 1.00
MT2 0.31* -0.03 0.22 0.13 0.51* 0.57* 0.87* 1.00
CIS-C 0.27 0.09 0.30* 0.08 0.04 0.09 0.08 0.09 1.00
SIP-AB 0.17 0.16 0.41* 0.35* 0.35* 0.40* 0.31* 0.27 0.58* 1.00
*Correlations are significant at the 0.05 level (2-tailed). The RT values and MT values have been logtransformed to approach a normal distribution.
ASTMT=Amsterdam Short Term Memory Test, BCS=Body Consciousness Scale, CIS= Checklist Individual Strength, CIS-C=CIS concentration, MT=movement times, RT=reaction times, SCL=Symptom Checklist, SCL-S=SCL somatization, SIP= Sickness Impact Scale, SIP-AB=SIP alertness behavior.
FIG. 1. Model information-processing speed. Error terms have been omitted in the figure. RT=simple reaction time, RT2=choice reaction time, beta==betas in the model. *p<0.05
1. Fukuda K, Straus SE, Hickie I, et al. The chronic fatigue syndrome: a
comprehensive approach to its definition and study. International Chronic
Fatigue Syndrome Study Group [see comments]. Ann Intern Med
2. Vercoulen JH, Bazelmans E, Swanink CM, et al. Evaluating
neuropsychological impairment in chronic fatigue syndrome. J Clin Exp
3. Michiels V, de Gucht V, Cluydts R, et al. Attention and information
processing efficiency in patients with Chronic Fatigue Syndrome. J Clin
Exp Neuropsychol 1999;21:709-29.
4. Joyce E, Blumenthal S, Wessely S. Memory, attention, and executive
function in chronic fatigue syndrome. J Neurol Neurosurg Psychiatry
5. Moss MR, Petrie KJ, Large RG, et al. Neuropsychological deficits in
chronic fatigue syndrome: artifact or reality? J Neurol Neurosurg
6. DeLuca J, Johnson SK, Beldowicz D, et al. Neuropsychological impairments
in chronic fatigue syndrome, multiple sclerosis, and depression. J Neurol
Neurosurg Psychiatry 1995;58:38-43.
7. Krupp LB, Sliwinski M, Masur DM, et al. Cognitive functioning and
depression in patients with chronic fatigue syndrome and multiple
sclerosis. Arch Neurol 1994;51:705-10.
8. DeLuca J, Johnson SK, Natelson BH. Information processing efficiency in
chronic fatigue syndrome and multiple sclerosis. Arch Neurol
9. Johnson SK, DeLuca J, Diamond BJ, et al. Selective impairment of auditory
processing in chronic fatigue syndrome: a comparison with multiple
sclerosis and healthy controls. Percept Mot Skills 1996;83:51-62.
10. Scheffers MK, Johnson R, Grafman J, et al. Attention and short-term memory
in chronic fatigue syndrome patients: an event-related potential analysis.
11. Tiersky LA, Johnson SK, Lange G, et al. Neuropsychology of chronic fatigue
syndrome: a critical review. J Clin Exp Neuropsychol 1997;19:560-86.
12. DiPino RK, Kane RL. Neurocognitive functioning in chronic fatigue
syndrome. Neuropsychol Rev 1996;6:47-60.
13. van-der-Werf SP, Prins JB, Jongen PJH, et al. Abnormal neuropsychological
findings are not necessarily a sign of cerebral impairment: a matched
comparison between chronic fatigue syndrome and multiple sclerosis.
Neuropsychiatry Neuropsychol Behav Neurol 2000;13:199-203.
14. Vercoulen JH, Swanink CM, Galama JM, et al. The persistence of fatigue in
chronic fatigue syndrome and multiple sclerosis: development of a model. J
Psychosom Res 1998;45:507-17.
15. Vercoulen JH, Swanink CM, Fennis JF, et al. Dimensional assessment of
chronic fatigue syndrome. J Psychosom Res 1994;38:383-92.
16. Vercoulen JH, Hommes OR, Swanink CM, et al. The measurement of fatigue in
patients with multiple sclerosis. A multidimensional comparison with
patients with chronic fatigue syndrome and healthy subjects. Arch Neurol
17. Fenigstein A, Scheier MF, Buss AH. Public and private self-consciousness:
Assessment and theory. J Cons Clin Psychol 1975;43:522-7.
18. Miller LC, Murphy R, Buss AH. Consciousness of body: Private and public. J
Pers Soc Psychol 1981;41:397-406.
19. Vervaeke GAC, Bouman TK, Valmaggia LR. Attentional correlates of illness
anxiety in a non-clinical sample. Psychother Psychosom 1999;68:22-5.
20. Martin JB, Ahles TA, Jeffery R. The role of private body consciousness and
anxiety in the report of somatic symptoms during magnetic resonance
imaging. J Beh Ther Exp Psychiatry 1991;22:3-7
21. Klein K, Barnes D. The relationship of life stress to problem solving:
Task complexity and individual differences. Social Cognition.
22. Baradell JG, Klein K. Relationship of life stress and body consciousness
to hypervigilant decision making. J Pers Social Psychol. 1993;64:267-73.
23. Vervaeke GAC, Bouman TK, Valmaggia LR. Attentional correlates of illness
anxiety in a non-clinical sample. Psychother Psychosom 1999;68:22-5.
24. Martin JB, Ahles TA, Jeffery R. The role of private body consciousness and
anxiety in the report of somatic symptoms during magnetic resonance
imaging. J Beh Ther Exp Psychiatry 1991;22:3-7.
25. Ahles TA, Cassens HL, Stalling RB. Private body consciousness, anxiety and
the perception of pain. J Beh Ther Exp Psychiatry 1987;18:215-22.
26. Miller LC, Murphy R, Buss AH. Consciousness of body: Private and public. J
Pers Soc Psychol 1981;41:397-406.
27. Vervaeke GAC, Bouman TK, Sanderman R, et al. De
lichaamsbewustzijnsvragenlijst: vertaling van de Body Consciousness
Questionnaire. Groningen: Department of Clinical Psychology, University of
28. Arrindell WA, Ettema H: Dimensional structure, reliability and validity of
the Dutch version of the Symptom Checklist (SCL-90):Data based on a phobic
and a 'normal' population. Nederlands Tijdschrift voor de Psychologie en
haar Grensgebieden. 1981;36:77-108.
29. Jolles J, Houx PJ: Age-related decline of psychomotor speed: effects of
age, brain health, sex, and education. Percept Mot Skills 1993;49:195-211.
30. Bergner M, Bobbit RA, Carter WB, et al. The Sickness Impact Scale:
development and final revision of a health status measure. Med Care
31. Schagen S, Schmand B, de-Sterke S, et al. Amsterdam Short-Term Memory
test: a new procedure for the detection of feigned memory deficits. J Clin
Exp Neuropsychol 1997;19:43-51.
32. Schmand B, Lindeboom J, Schagen S, et al. Cognitive complaints in patients
after whiplash injury: the impact of malingering. J Neurol Neurosurg
33. Joreskog K., Sorbom D. LISREL 8: Structural equation modelling with the
SIMPLIS command language. Hillsdale: Lawrence Erlbaum associate
34. Arbuckle J.L., Wothke W. Amos 4.0 User's guide. Smallwaters Corporation,
35. Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen
KA, Long JS, eds. Testing structural equation models. London: Sage
36. Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy
for chronic fatigue syndrome: a multicenter randomised controlled trial.
37. Reid S, Chalder T, Cleare A, et al. Extracts from 'Clinical Evidence'
Chronic fatigue Syndrome. BMJ 2000;320:292-6.
(c) 2002 Lippincott Williams & Wilkins
Be the first to comment on this article!