Chronic fatigue syndrome in children: a cross sectional
M X Patel, D G Smith, T Chalder, S Wessely
Arch Dis Child 2003;88:894–898
Background: Chronic fatigue syndrome (CFS) in children is a controversial diagnosis with unclear
aetiology, ill defined but likely increasing incidence, and debatable clinical management options.
However these children experience real and considerable suffering. Appropriate research in this clini-
cal population is sparse and usually occurs in tertiary referral units.
Methods: Cross sectional survey of 36 children attending a GP specialist interest clinic in southeast
Results: Patient sociodemographics and clinical morbidity were largely comparable to the literature
from tertiary referral research centres. Some prognostic indicators for adults did not readily transfer to
this younger age group, although several children had a positive family psychiatric history. Receiving
treatment was associated with increased school attendance, but one third of subjects obtained no
qualifications. Return to normal health or significant overall improvement was reported by 29/36 sub-
Conclusions: The outcomes in this setting are favourable and comparable to those seen in a control-
led setting; this study supports the concept that the prognosis for CFS in children and adolescents is
generally good. However, the impact of the illness is significant and this is perhaps most evident in
terms of education. Current methods of reporting educational outcomes in the literature are varied and
merit development of standardised tools.
considered by some as too long for children.3 4Prolonged
school absenteeism is also occasionally included in the defini-
tion as a proxy measure for functional impairment and
severity.5Epidemiological studies for CFS in children are
sparse,6although in the Netherlands, the prevalence in
inhabitants.5In Australia, eight of 42 CFS cases were found to
be under the age of 15.7More recently, in a community survey
of nearly 4000 children in the USA,2% were found to have had
CFS-like illness.8Existing epidemiological data for children in
Great Britain is mainly founded on small scale studies, and
thus it has been concluded that the exact dimensions of the
problem remain unclear.9 10
In CFS, symptom occurrence in children is similar to that
seen in adults,11and most commonly includes: muscle ache,
sore throat,headache,abdominal pain,increased somnolence,
concentration difficulties,12–16and depression.17Most children
with CFS also display an impaired school performance and a
decrease in social activities.13 16However, children seemingly
differ from adults with earlier presentation and a more
optimistic outcome. In a systematic review,18four studies on
children with CFS showed that 54–94% of children recovered.
In this age group, good prognostic indicators include specific
physical triggers to the illness, start of illness in the autumn
school term, and higher socioeconomic status.19Alternatively,
important factors in the persistence of fatigue include somatic
attributions5or attributions to biological processes,20as well as
illness enhancing cognitions and behaviour of parents as well
as physical inactivity.5
Children with CFS have been successfully treated in
secondary and tertiary referral units.19The literature suggests
that these children should be treated with a multifactorial
approach; this usually includes a graded behavioural pro-
gramme, with an additional graded approach for return to
he definition of chronic fatigue syndrome (CFS) in
children is usually taken to be the same as that for
adults,1 2although the criterion of a six month duration is
as 10–20per 100 000
school, which is sometimes combined with family therapy.6 21
the family to encourage more activity, decreasing attention to
somatic complaints, and reducing illness behaviour. Disrup-
tion to education, social, recreational, and peer group integra-
tion should be minimised.9The use of antidepressants in chil-
dren has been advocated,6where depression is a significant
feature(asoften seen in
Unfortunately, CFS in children has received much adverse
publicity. The very essence of diagnosis and the clinical label-
ling of such children has been questioned.23–25Even effective
management options have been ridiculed. However, there has
been recent widespread concern about an apparent increase in
CFS in children.9The impact of this illness is profound as one
survey suggested that CFS is responsible for 50% of long term
absences from school.26There is also the frequently repeated
assertion that the mean length of illness in childhood is
approximately five years.27It is against this undesirable back-
drop that we detail a cross sectional study reporting on degree
of disability and symptomatology in a GP specialist interest
clinic for CFS. Further, prognostic indictors in children with
CFS are considered.
CFS), based onevidence
Using a cross sectional design, all eligible patients were sent a
battery of standardised questionnaires by post. Further data
were later obtained from the clinical notes and a follow up
Abbreviations: CFS, chronic fatigue syndrome; GHS, General Health
Questionnaire; GP, general practitioner; HAD, Hospital Anxiety and
See end of article for
Dr D G Smith, Malagay
Barn, Church Road, West
Tilbury Village, Essex
RM18 8TU, UK;
16 January 2003
DGS runs a general practitioner (GP) specialist interest clinic
in chronic fatigue syndrome for assessment and treatment of
patients of all ages in southeast England. On average, he sees
one hundred new National Health Service patients per year.
Two thirds of these patients reside in the areas of Sussex,Kent,
East London, and Essex. The majority are referred directly by
the patient’s GP and approximately one third of patients are
referred directly by consultants.No referred patients are too ill
to travel to the clinic and none are private referrals.
Initial assessment clinical interviews, including screening
for anxiety and depression, are followed by a standard battery
of clinical investigations. Treatment entails engagement, edu-
cation, and medication (antidepressants of low to average
dose). Further psychological aspects of treatment, which pri-
marily involve an individualised modified activity programme
for both mental and physical activity, address the following
• Discussion regarding stressors
• Normalisation of sleep pattern (sleep hygiene)
• Structured graduated daytime activity programme with self
• Graduated stepwise reintroduction to previous main
daytime activity (for example, schooling)
• Development of non-maladaptive coping strategies for
For the first 6–9 months, patients are reviewed on a monthly
basis followed by two-monthly appointments. The average
length of treatment duration is two and a half to three years.
All children aged 18 or less, who were patients currently
receiving treatment in summer 1998 or who had completed
treatment in the previous 12 months were included in this
study. The only exclusion criterion was a primary diagnosis
other than chronic fatigue syndrome.
The battery of standardised questionnaires included the
following outcome measures:
• Fatigue Questionnaire.28Eleven fatigue items are scored on
a four-option continuum from “less than usual” to “much
more than usual”. Ratings are coded using a bimodal
method,with caseness being defined as a minimum score of
4 (range 0–11).
• General Health Questionnaire (GHQ), 12 item.29Twelve
depression and anxiety items are rated on the same four
options as used in the fatigue questionnaire. Ratings are
coded using a bimodal method with a cut off of 4 indicating
psychological caseness (range 0–12).
• Social adjustment scale.30Impairment in work, home man-
agement, social activities, and private leisure are rated on
scales from 0 (no impairment) to 8 (maximal impairment).
The subscales of work and home management were
adapted to school and home life measures, deemed more
suitable for children. Scores were summated to give a total
score of function of “work” and social ability.
• Hospital Anxiety and Depression scale (HAD).31Fourteen
depression and anxiety items are rated on by three
individual options per item relating to severity of the symp-
tom. Subscale scores are calculated for depression and
anxiety respectively (subscale ranges 0–17 each). Subscale
scores greater than 11 are defined as reaching caseness.
• Fear questionnaire.32Twenty items are rated on scales from
0 (no avoidance) to 8 (total avoidance). Scores are
calculated for five-item subscales of agoraphobia, blood,
social,and dysphoria (subscale ranges 0–40 each) as well as
a total aggregate score (range 0–100).
• Symptom checklist.33Forty common symptom items scored
as being present or not present within the previous week.
An aggregate total score of all the affirmative answers is
calculated (range 0–40).
• Attribution scale.34Nine attribution items scored as “not a
factor”, “might be a factor”, or “definitely a factor” and are
• Patient satisfaction.35Sixteen items concerning satisfaction
with various aspects of treatment, including the clinician-
patient relationship, are scored as occurring or not
occurring,and also on a five-point Likert scale for degree of
usefulness. Items are coded on an individual item basis.
• Global outcome measure.34Five items measuring global
improvement, satisfaction with treatment outcome, degree
of handicap (restriction), improvement in fatigue, and use-
fulness of treatment are measured with the first four factors
on a seven-point scale and the latter on a five-point Likert
scale with individual options per item. Low scores indicate
significant improvement. Ratings are collapsed into di-
chotomous categories with scores of 1–3 (representing sig-
nificant improvement/usefulness/satisfaction), and scores
of 3 or more (representing deterioration/uselessness/
Additionally, questions on demographic data, school attend-
ance, benefits, and alcohol intake were included.
Subjects were identified by the clinic database and were sent
the battery of standardised questionnaires by post. Patient
consent was taken as implicit on return of completed
questionnaires. Further data were obtained from the clinical
notes including diagnosis, medication used, and duration of
treatment. Eighteen months later, a follow up telephone call
was made to further identify educational outcome data, for
example, number of GCSEs achieved. Data were analysed
using SPSS software.
Thirty six of 40 children and their parents responded to our
questionnaires (90%). Table 1 shows the sociodemographic
Standard occupational classification of parents36
Mean SD Range
No. of months pre-presentation
Total no. of months in treatment
Occupations are classified in terms of their skill as defined by the
nature and duration of the qualifications, training, and work
experience required for competency in tasks of a particular job. Level
4 relates to professional occupations and managerial positions; level
3 includes technical occupations and a variety of trade occupations;
level 2 includes machine operation, caring, and clerical occupations;
level 1 includes cleaners, porters, and catering assistants.
Chronic fatigue syndrome in children 895
The onset was reported to be uncertain by 27 children (75%).
At the time of this cross sectional survey, 12 children (33.3%)
reported fatigue lasting less than six months, and nine
children (25%) reported fatigue of six months or more, with
15 children (41.7%) not reporting duration of fatigue. Table 2
details the findings for the main measures of fatigue, GHQ,
social adjustment, HAD scale, and fear. Co-morbid syndromes
of anxiety (n = 11, 30.6%), stress related (n = 6, 16.7%),
social phobia (n = 2, 5.6%), depression (n = 1, 2.8%), and
physical disease (n = 3,8.3%) were detected,with 15 children
(41.7%) being free of co-morbid syndromes.
The most commonly reported symptoms were: headaches
(n = 28, 77.8%), daytime drowsiness (n = 22, 61.1%), sore
throat (n = 17, 47.2%), eyestrain (n = 17, 47.2%), stomach
pain (n = 15, 41.7%), dry mouth (n = 14, 38.9%), back pain
(n = 14, 38.9%), joint pain (n = 14, 38.9%), muscle pain
(n = 14, 38.9%), and stiffness in limbs (n = 12, 33.3%).
Five (13.9%) children had a positive family history of chronic
fatigue syndrome in a first degree relative.Nine (25.0%) had a
positive psychiatric family history in a first degree relative:five
parents, two siblings, and two (of the same family) were all
first degree family members.Only four (11.1%) of the children
received benefits, and only six (16.7%) belonged to a self-help
group. Of the 14 (38.9%) who indicated that they drank alco-
hol, seven were drinking normal amounts, four had reduced
their intake, and three were abstaining.
Twenty two (61.1%) felt that a viral onset was definitely a fac-
(11.1%) considered a current or ongoing infection was
definitely a factor,with a further nine (25.0%) thinking that it
might be a factor. Only two (5.6%) thought that food or other
allergies were definitely the cause. In terms of psychological
attributions, 13 (36.1%) indicated that emotional upset was
definitely a factor and 14 (38.8%) thought that it might be a
factor. Twelve (33.3%) highlighted that school stresses were
definitely a factor and eight (22.2%) believed that the stress of
relationships was definitely involved.
Thirty four of 36 children (94.4%) were prescribed at least one
antidepressant, of which eight received a tricyclic antidepres-
sant only and 26 received both a tricyclic antidepressant and a
selective serotonin reuptake inhibitor. Thirty one (86.1%) of
the children and their families felt that they could tolerate the
medication and 30 of them (83.3%) believed that it helped
them. Twenty nine (80.6%) thought that the modified activity
programme for mental activity was useful and 27 (75.0%)
believed that the modified activity programme for physical
activity was useful.
Patient satisfaction scale
Highest scoring items included “having my illness taken seri-
ously”(97.2%),“changing how I manage the illness”(94.4%),
“getting support and encouragement” (94.4%), “learning
greater control over my symptoms” (86.1%), and “adopting a
step by step approach” (86.1%). However, only 16 children
(44.4%) felt that they had a one-to-one relationship with their
Global outcome measure
For this measure, lower scores indicate good results (see table
In the week prior to first presentation at the clinic, the mean
percentage of time spent at school was 29.9% (SD 33.9%,
range 0–100%). Fourteen children (43.8%) informed us that
they had not attended school at all in the week prior to pres-
entation and only 12 (37.5%) stated that they had attended for
more than 50% of the full school week. Sixteen children and
their families applied for home tuition, of which 13 received
some. Worrying about returning to school was highlighted by
26 (72.2%) children, and 22 (61.1%) felt that their school had
At the time of the survey, 32 of 36 of the children (88.9%)
were registered for full time education, of which 28 (77.8%)
MeanSD Range Caseness n (%)
Total fatigue score (bimodal)
GHQ total (bimodal)
Summated scores of 4 subscales
Social leisure impairment
Private leisure impairment
“Home management” impairment
(Relationship impairment2.682.86 0–8)
Symptom checklist 107.46 1–32
896Patel, Smith, Chalder, et al
did actually attend. Receiving treatment was therefore associ-
ated with increased school attendance, but in the absence of
randomisation, this cannot be considered as causal.
At the follow up telephone call GCSE results were discussed
(pass grades A–E).The median number of passes was 3 (range
0–10).Sixteen (44.4%) achieved five or more passes and a fur-
ther nine children (25%) achieved 1–4 passes,although not all
of these were at a single sitting. Unfortunately, 11 children
(30.6%) did not achieve any passes or did not even sit for their
examinations. Twenty (55.6%) remained in full time educa-
tion and five (13.9%) were in full time employment with a
further one (2.8%) in part time employment. Three (8.3%) of
the children were not working and were not in education,and
there was missing data for the remaining seven (19.4%) chil-
This study was conducted in a GP specialist interest clinic in
southeast England, and therefore differs from most other
studies which report findings from tertiary centres.37With a
sample size of 36 children aged 11–18 years,this study is large
in comparison to most in the current literature.
Sociodemographic factors and the extent of general
morbidity and functioning in the clinic were predominantly
comparable to the literature.13 16 19 38–40As a cross sectional
study, subjects were captured at different stages of treatment
and this may explain why higher fatigue scores have been
reported elsewhere.38However, it has been further suggested
that children and their parents might have unrealistic views of
normative fatigue levels.20Further our data showed that 75%
reported the illness onset to be slow and uncertain in nature.
Similarly another study41has reported that 77% of their 35
patients had a gradual illness onset, and together these find-
ings challenge the more commonly reported acute onset.12
For prognostic factors, we endorse that CFS in children be
considered as an entity separate to that found in adults,6as
several common adult prognostic factors1 18occur less fre-
quently in children. However, a familial influence was seen in
five subjects and others had a positive family psychiatric his-
tory. Furthermore, we found a high attribution to viral onset
and lower attribution to psychological processes.
Overall the subjects found the treatment to be useful and
were satisfied with it. However, it is of concern that over half
of the children felt that they did not have a one-to-one
relationship with their doctor and this is perhaps indicative of
the presence of parents during all clinical appointments.
While this is understandable, it may adversely impact on per-
ceived patient autonomy for the adolescent.
Subjects also indicated that they were much better than
they were before treatment as overall improvement or return
to normal health was reported by 29/36 (80.6%) subjects and
no change or worsening health was reported by the remaining
seven (19.4%). These findings are favourably comparable to
those reported in the literature (see table 4). In general,
outcome for children with CFS is favourable,13 15 16 41and
several children resume normal activities.19
Notably, school attendance prior to treatment was low and
nearly half the children were not attending school at all.These
findings are worse than those previously reported.12 13 16 19One
of the main difficulties here is that there is a lack of
consistency across the literature in style and time parameters
used for reporting school attendance and academic outcomes.
In our study, home tutelage was received by 13 children; this
in itself remains controversial,6as most advocate a return to
school as soon as possible.12The issue of school refusal in chil-
dren with CFS has been previously highlighted,3and with 26
children worrying about returning to school, it is of concern
that the return to school was possibly further prolonged.It has
even been suggested that CFS could be conceptualised as a
form of school refusal with the two conditions having essen-
tially the same causes.6In this study receiving treatment was
associated with increased school attendance. However, from
the follow up telephone call we identified that a third of the
subjects either obtained no passes or did not even sit for their
Very or much better Little, same, or worseMean (SD)Range
Very or moderately
Slightly satisfied or
dissatisfied Mean (SD) Range
Satisfaction30 (83.3%) 6 (16.7%)1.85 (1.52) 1–7
Very or moderately
Useful or not
Useful30 (83.3%)6 (16.7%) 1.56 (1.05)1–5
Reported global improvement outcomes18
No change or
worse Follow up duration
Krilov et al (1998)16
Feder et al (1994)12
Chalder et al (2002)21
Bell et al (2001)41
Carter et al (1995)13
Marshall et al (1991)15
Garralda et al (1999)40
Smith et al (1991)14
Range 1–4 years
Mean 3.8 years
Mean 6 months
Average 13 years
Median 16.9 months
Median 26 months
Mean 45.5 months
Mean 18.4 months
Chronic fatigue syndrome in children897
Conclusion Download full-text
This study shows that in children and adolescents with CFS,
patient characteristics in a GP specialist interest clinic setting
are comparable to those seen in the more controlled setting of
a tertiary referral research unit. Similarly, treatment methods
used can be transferred into the real world with largely
favourable outcomes, which are again comparable to the con-
trolled setting. While there is clearly a need for better, more
focused multidisciplinary rehabilitation, this study suggests a
far more favourable prognosis for CFS than some of the popu-
lar literature. Given that the poor prognostic speculation27can
become a self-fulfilling prophecy by increasing helplessness,
avoidance,and depression,we feel it is important to bring this
more favourable prognosis to light.
In our study the most striking finding was perhaps that less
than half of the children were attending school in the week
prior to presentation in the clinic. The possible impact of
treatment on school attendance was favourable but still with
varying degrees of time spent in school. Comparison with the
literature was in general only possible with smaller scale stud-
ies and the method of recording data on education varied con-
siderably. In real life, the outcome of education is generally
seen as the degree of success at national examinations
(GCSEs) and entry to further education. These are rarely
reported in the clinical literature.It is therefore suggested that
consistency be obtained by development of standardised
scales for school attendance and educational outcomes. Only
then can the true impact of CFS on education be properly
The main criticism of this study is that it used a cross sec-
tional design and therefore relied on the responses to
questionnaires as certain parameters were not detailed at first
presentation. Despite this, this study supports the existing
evidence that the prognosis for CFS in children and
adolescents is generally good. However, the impact of the
illness is significant,and this is perhaps most evident in terms
Grateful thanks are extended to all participants and their families,
Isama Smith for her administrative skills, Lakshika Tennakoon and
Sophia Rabe-Hesketh for assistance with statistical analysis, Carmine
Pariante and Paola Dazzan for their guidance,and particular thanks to
M X Patel, T Chalder, S Wessely, Institute of Psychiatry and GKT
School of Medicine, De Crespigny Park, London SE5 8AF, UK
D G Smith, Malagay Barn, Church Road, West Tilbury Village, Essex
RM18 8TU, UK
1 Sharpe M, Archard L, Banatvala J, et al. Chronic fatigue syndrome:
guidelines for research. J R. Soc Med 1991;84:118–21.
2 Royal Colleges of Physicians, Psychiatrists & General Practitioners.
Chronic fatigue syndrome. Report of a joint working group. London:
Royal Colleges of Physicians, Psychiatrists & General Practitioners, 1996.
3 Vereker M. Chronic fatigue syndrome: a joint paediatric-psychiatric
approach. Arch Dis Child 1992;67:550–5.
4 Richards J, Smith F. Chronic fatigue syndrome in children and
adolescents: general practitioners’ experience of the problem and their
views about its treatment. Psychiatr Bull 1998;22:203–6.
5 de Jong LW, Prins JB, Fiselier TJ, et al. [Chronic fatigue syndrome in
young persons]. Nederlands Tijdschrift voor Geneeskunde
6 Richards J. Chronic fatigue syndrome in children and adolescents: a
review article. Clin Child Psychol Psychiatry 2000;5:31–51.
7 Lloyd A, Hickie I, Boughton R, et al. Prevalence of chronic fatigue
syndrome in an Australian population. Med J Aust 1990;153:522–8.
8 Jordan KM, Ayers PM, Jahn SC, et al. Prevalence of fatigue and chronic
fatigue syndrome-like illness in children and adolescents. Journal of
Chronic Fatigue Syndrome 2000;6:3–21.
9 Wessely S, Hotopf M, Sharpe M. CFS in children. Chronic fatigue and
its syndromes. Oxford: Oxford University Press, 1998:284–97.
10 Straus SE. Caring for patients with chronic fatigue syndrome:
conclusions in CMO’s report are shaped by anecdote not evidence. BMJ
11 Whiting P, Bagnall AM, Sowden AJ, et al. Interventions for the treatment
and management of chronic fatigue syndrome: a systematic review.
12 Feder H, Dworkin P, Orkin C. Outcome of 48 paediatric patients with
chronic fatigue; a clinical experience. Arch Fam Med 1994;3:1049–55.
13 Carter B, Edwards J, Kronenberger W, et al. Case control study of
chronic fatigue in paediatric patients. Pediatrics 1995;95:179–86.
14 Smith M, Mitchell J, Corey L, et al. Chronic fatigue in adolescents.
15 Marshall GS, Gesser RM, Yamanishi K, et al. Chronic fatigue in
children: clinical features, Epstein-Barr virus and human herpes virus 6
serology and long term follow-up. Pediatr Infect Dis J 1991;10:287–90.
16 Krilov LR, Fisher M, Friedman SB, et al. Course and outcome of chronic
fatigue in children and adolescents. Pediatrics 1998;102(2 pt 1):360–6.
17 Bell D. Chronic fatigue syndrome in children. Journal of Chronic Fatigue
18 Joyce J, Hotopf M, Wessely S. The prognosis of chronic fatigue and
chronic fatigue syndrome: a systematic review. Q J Med
19 Rangel L, Garralda ME, Levin M, et al. The course of severe chronic
fatigue syndrome in childhood. J R. Soc Med 2000;93:129–34.
20 Garralda ME, Rangel L. Childhood chronic fatigue syndrome. Am J
21 Chalder T, Tong J, Deary V. Family cognitive therapy for chronic fatigue
syndrome: an uncontrolled study. Arch Dis Child 2002;86:95–7.
22 Emslie GJ, Rush AJ, Weinberg WA, et al A double-blind, randomised,
placebo-controlled trial of fluoxetine in children and adolescents with
depression. Arch Gen Psychiatry 1997;54:1031–7.
23 Moutarde O. Myalgic encephalopmyelitis in children. Lancet
24 Van Houdenhove, B. Does myalgic encephalopmyelitis exist? Lancet
25 Speight N, Franklin A. Does myalgic encephalopmyelitis exist? Lancet
26 Colby J. Study finds 12000 pupils may be suffering from ME. Sunday
Times 5 June 1994.
27 Franklin A. Children with ME: guidelines for school doctors and general
practitioners. Stanford-le-Hope, Essex: ME Association, 1995.
28 Chalder T, Berelowitz G, Hirsch S, et al The development of a fatigue
scale. J Psychosom Res1993;37:147–53.
29 Goldberg D The detection of psychiatric illness by questionnaire.
London: Oxford University Press, 1972.
30 Marks I. Behavioural psychotherapy. Maudsley pocket book of clinical
management. Bristol: Wright, 1986:37–9, 45–6.
31 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta
Psychiatr Scand 1983;67:361–70.
32 Marks IM, Mathews AM. Brief standard self-rating for phobic patients.
Behav Res Treatment 1979;17:263–7.
33 Wittenborn JR, Buhler R. Somatic discomforts among depressed women.
Arch Gen Psychiatry 1979;36:465–71.
34 Deale A, Chalder T, Marks I, et al. Cognitive behaviour therapy for
chronic fatigue syndrome: a randomised control trial. Am J Psychiatry
35 Llewelyn SP, Hume WI. The patient’s view of therapy. Br J Med Psychol
36 Office for National Statistics. Standard occupational classification,
Vols 1 and 2. London: The Stationery Office, 2000.
37 Euba R, Chalder T, Deale A, et al. Comparison of the characteristics of
chronic fatigue syndrome in primary and tertiary care. Br J Psychiatry
38 Walford GA, Nelson WM, McCluskey DR. Fatigue, depression, and
social adjustment in chronic fatigue syndrome. Arch Dis Child
39 Rowe KS. Double-blind randomised controlled trial to assess the efficacy
of intravenous gammaglobulin for the management of chronic fatigue
syndrome in adolescents. J Psychiatr Res 1997;31:133–47.
40 Garralda E, Rangel L, Levin M, et al. Psychiatric adjustment in
adolescents with a history of chronic fatigue syndrome. J Am Acad Child
Adolesc Psychiatry 1999;38:1515–21.
41 Bell DS, Jordan K, Robinson M. Thirteen-year follow-up of children and
adolescents with chronic fatigue syndrome. Pediatrics 2001;107:994–8.
898 Patel, Smith, Chalder, et al