probably been the acceptability of myalgic encephalo-
myelitis as a diagnosis. Fatigue scores in our study
were strongly correlated with depression scores, and
a recent prospective study in general practice has
confirmed this relation.18
Depression as a cause of fibromyalgia is a much
debated issue. Persistent pain
perience. Our finding that the depression score rises
with tender point count irrespective of pain status,
however, suggests that depression and fatigue may
play a part in the genesis oftender points. This fits with
the theory ofthe central modulation ofpain experience.
is a depressing ex-
We conclude that tender points are a measure
of general distress. Although they are related to
complaints ofpain, they seem to be linked separately to
depression, fatigue, and poor sleep. Our data support
the hypothesis that sleep disturbances may be a factor
in the development of tender points, but prospective
studies are required to investigate this further. Tender
point counts may prove a useful measure in epidemio-
logical studies ofthe causes of chronic musculoskeletal
pain, but high counts do not define a distinct disease
entity in the general population.
We thank Dr John Sandars, Dr Gerald Coope, and their
partners and staff of the general practices in the study; Dr Jeff
Marks of Stepping Hill Hospital for help with the pilot
studies; and Lesley Jordan for typing the manuscript. The
study was supported by a grant from the North West Regional
Health Authority and by the Arthritis and Rheumatism
1 Smythe HA, Moldofsky H. Two contributions to understanding of the
"fibrositis" syndrome. BullRheum Dis 1977;28:928-31.
2 Yunus MB, Masi AT, Calabro JJ, Miller KA, Feigenbaum SL. Primary
fibromyalgia (fibrositis): a clinical study of 50 patients with matched normal
controls. SeminArthrins Rheum 1981;1l1:151-71.
3 Yunus MB, Ahles TA, Aldag JC, Masi AT. Relationship of clinical features
with psychological status in primary fibromyalgia. Arthritis Rheum 1991;34:
4 Wolfe F. Fibromyalgia. Rheum Dis Clin NorthAm 1990;16:681-98.
5 Reilly PA. Fibromyalgia in the workplace: a "management" problem. Ann
Rheum Dis 1993;52:249-51.
6 Cohen ML, Quintner JL. Fibromyalgia syndrome, a problem of tautology.
7 Croft P, Rigby AS, Boswell R, Schollum J, Silman A. The prevalence of
chronic widespread pain in the general population. j? Rheumatol 1993;20:
8 Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg
DL, et al. The American College of Rheumatology 1990 criteria for the
classification of fibromyalgia. Report of the multicenter criteria committee.
Arthritis Rheum 1990;33:160-72.
9 Goldberg D, Williams P. A user's guide to the general health questionnaire.
Windsor: NEFR-Nelson, 1988.
10Jenkins CD, Stanton B-A, Niemcryk SJ, Rose RM. A scale for the
estimation of sleep problems in clinical research. J Clin Epidemniol 1988;41:
11 Chalder T, Berelowitz G, Pawlikowska T, Watts L, Wessely S, Wright D,
etaL Development ofa fatigue scale.JyPsychosom Res 1993;37:147-53.
12 Yunus M. Towards a model of pathophysiology of fibromyalgia: aberrant
central pain mechanisms with peripheral modulation. YRheumatol 1992;19:
13 Wall PD. The mechanisms of fibromyalgia: a critical essay. In: Voeroy H,
Merskey H, eds. Progress in fibronnyalgia and myofascial pain. Amsterdam:
14 Campbell SM, Clark S, Tindall EA, Forehand ME, Bennett RM. Clinical
characteristics offibrositis. Arthritis Rheum 1983;26:817-24.
15 Moldofsky H, Scarisbrick P. Induction of neurasthenic musculoskeletal pain
syndrome by selective sleep stage deprivation. Psychosom Med 1976;38:
16 Leigh TJ. Sleep in rheumatic patients. ScandJRheumatol 1990;19:5-9.
17 Goldenberg DL, Simms RW, Geiger A, Komaroff AL. High frequency of
fibromyalgia in patients with chronic fatigue seen in a primary care practice.
Arthritis Rheum 1990;33:381-7.
18 Ridsdale L, Evans A, Jerrett W, Mandlia S, Osler K, Vora H. Patients with
fatigue in general practice: a prospective study. BMJ 1993;307:103-6.
(Accepted 12NJy 1994)
Hertfordshire ALIO 9AB
DJohn Done, senior lecturer
Clinical Research Centre,
TimothyJ Crow, headof
Amanda Sacker, research
Royal Edinburgh Hospital,
Edinburgh I1110 SHF
Eve C Johnstone, professor
Childhood antecedents ofschizophrenia and affective illness: social
adjustment at ages 7 and 11
DJohn Done, TimothyJ Crow, Eve C Johnstone, Amanda Sacker
Objective-To investigate the social adjustment
in childhood of people who as adults have psy-
Design-Subjects in a prospectively followed up
cohort (the national child development study) who
had been admitted as adults to psychiatric hospitals
were compared with the rest ofthe cohort on ratings
of social behaviour made by teachers at the ages of
7 and 11 years.
Subjects-40 adult patients with schizophrenic
with neurotic illness who had been admitted for
psychiatric reasons by the age of 28. 1914 randomly
selected members ofthe cohort who had never been
admitted for psychiatric treatment.
Main outcome measures-Overall scores and
scores for overreaction (externalising behaviour)
and underreaction (internalising behaviour) with the
Bristol social adjustment guide at ages 7 and 11.
Results-At the age of 7 children.who developed
schizophrenia were rated by their teachers as mani-
festing more social maladjustment than. controls
(overall score 4 3'(SD 24) Av 3- 1 (2*0; P<O.O1). Ths
was more apparent inmt eboys (5(246)) than the girls
(3 4 (1.8)) and related to6overreactive .rather than
underreactive behaviour. At both ages prepsychotic
controls. By the age of 11 preneurotic children,
particularly the girls, had an increased rating of
little from notmal
maladjustment (including overreactions and under-
Conclusion-Abnormalities of social adjustment
are detectable in childhood in some people who
develop psychotic illness. Sex and the rate of
development ofdifferent components ofthe capacity
for social interaction are important determinants of
the risk ofpsychosis and other psychiatric disorders
Schizophrenic and affective psychoses-the major
psychiatric illnesses in adulthood-have a lifetime
prevalence (probably similar in different societies) of
2-3%. Onsets are rare before puberty but then rise
steeply (earlier in males than females), with a predomi-
nant impact in early and middle adult life. Twin,
family, and adoption studies establish a genetic role in
aetiology, and there is no strong evidence for an
Abnormalities have been reported in patients with
and. affective disorders. long before
the onset''of psychotic symptoms. People who were
admitted '-to. hospital with schiz'ophrenia had' lower
intelligence quotients (IQ) 'than 'their siblings and
classmates2 and were also reported by their teachers as
exhibiting deviant 'behaviour.3 4 These abnormalities
(apparently confined to the boys) have' been a'ttributed
to neurodevelopmental impairment, susceptibility to
BMJ VOLUME 309
17 SEPTEMBER 1994
psychosocial stressors, or an interaction of these two
We previously identified the subjects in the 1958
cohort of the national child development study who as
adults were admitted to psychiatric hospitals, and by
examining their case notes we identified those that had
schizophrenia, affective psychoses, and non-psychotic
disorders.6 We investigated the social behaviour in
childhood ofthese patients in relation to diagnosis and
sex by analysing the results of psychometric assess-
ments performed at the age of7 and 11 years.
Subjects and methods
The British perinatal mortality survey of 1958
included some 98% of all births in England, Scotland,
and Wales registered during the week ofthe 3-9 March
1958. Four subsequent attempts (in 1965, 1969, 1974,
and 1981) to trace members of the cohort to monitor
physical, educational, and social development became
known as the national child development study. The
numbers of subjects followed up at each stage were
15 398, 15 303, 14 761, and 12 537.
Using the mental health enquiry we identified all
subjects in the national study who had been treated as
adults in hospital for psychiatric reasons between 1974
and 1986. Such a sampling procedure is unlikely to
miss cases of schizophrenia as only a small proportion
of schizophrenic patients do not have contact with
patients with neuroses that are presumably severe.
We derived diagnoses of schizophrenia,
psychosis, and neurosis from the casenote histories as
previously described6 using the present state examina-
tion by application of the CATEGO program.8 From
hereon we refer to these groups as preschizophrenia,
respectively. The control group was a randomly
selected 10% sample of subjects in the national cohort
who had never been admitted to hospital for psychia-
tric treatment and had been followed up at least once.
Table I shows the numbers ofsubjects in each group at
it will select an atypical group of
TABLE i-Numbers of subjects from national child development study who had psychometric test results at
ages 7and 11 according togroup andpsychiatric diagnosis in adulthood6
No (%) of
sample at age
at age 11
No withBSAG data at age 7
No with BSAG data at age 11
Group (adult diagnosis)
Normal control (n- 1914)
Preaffective psychotic disorders
(mania, depressive psychosis,
retarded depression) (n-35)
Preneurosis (neurosis) (n-79)
16 14 25 (75)
BSAG-Bristol social adjustment guide.
MEASURES OF SOCIALAND EMOTIONALADJUSTMENT
In 1965 and 1969 teachers were asked to complete
the Bristol social adjustment guide,9 a standardised
psychometric test of social maladjustment, by under-
lining which of some 150 descriptions of behaviour at
school were relevant to each child in the national child
development study. The scoring system allows assess-
ment of either global or more specific aspects of
maladjustment. An overall score has been found to be
useful in broad epidemiological surveys, although
subsequent revisions have acknowledged two separate
types of maladjustment referred to as underreaction
and overreaction (see appendix).9 This dichotomy is
relevant to studies of psychosis. Social isolation (one
aspect of underreaction) and difficulty in forming
reported in the premorbid personality ofschizophrenic
(aspects of overreaction) leading to weak peer affilia-
tions has been reported in studies of the childhood
behaviours of adult schizophrenic patients."2'3 Thus
we compared groups in terms ofboth overreaction and
The Bristol social adjustment guide (standardised in
the United Kingdom, the 1987 manual providing
normative data by social class and sex) also recognises
more focused dysfunctions, which are referred to as
core syndromes (appendix). High scores indicate social
maladaption. Subjects in social classes 4 and 5 (General
Register Office's classification'4) obtain higher scores
than those in social classes
however, that the social class of those with preschizo-
phrenia was significantly higher than that of the
controls (P=0-02). This would tend to lower scores,
but it should be noted that the father's occupation was
more often unavailable in the preschizophrenic group.
Our main concern was to establish whether the
four groups of children-that is, normal controls,
those with preschizophrenia, those with preaffective
psychotic disorders, and those with preneurosis-
differed in terms of their social adjustment scores. We
therefore used analysis of variance. Scores with the
Bristol social adjustment guide have a skewed distribu-
tion, so we used a square root transformation of the
scores to reduce the skewness and make the variances
comparable between groups. A four factor analysis of
variance with repeated measures on one factor was
calculated using spssx, release 2.1.'5 Missing data were
replaced for each case by the mean value for the same
sex, age, and group except for cases with data missing
at both ages.
have often been
1 and 2.9 We found,
Table II presents the scores for overreaction and
underreaction for each group at each age. There was
an overall difference between the groups (F- 14-0,
effect of sex
TABLEn-Mean (SD) square root transformed scores with Bristolsocialadjustmentguide in differentgroups at ages of7and 11
Overreaction Underreaction Total
5 (2 6)
4-55 (2 0)
3 0 (2 7)
3-5 (2 2)
2-7 (1 9)
4 0(2 4)
4 9(2 2)
1 7 (1-3)
2-2 (0 9)
2 1 (1 0)
2-0 (1 2)
2-7 (2 9)
4-2 (2 0)
17 SEPTEMBER 1994
FIG 1-Mean (square root transformed) scores with Bristol social
adjustment guide for overreaction and underreaction in boys and girls
according to mental health in adulthood. Ratings at ages of 7 and 11
(F= 19-05, P< 0001), boys scoring higher than girls in
three of the four groups (fig 1). Interpretation of these
main effects is, however, difficult until the significant
two way interaction between group and age (F=2.9,
P=003) and the significant three way interaction
between group, sex, and type of subscale (F=4-36,
P=0 005) have been clarified.
Figure 1 shows the three way interaction between
group, sex, and overreaction and underreaction. In
boys differences between groups were more obvious
for overreaction (F= 12.1, P<0001) than for under-
attributable to significantly higher scores in the pre-
schizophrenic boys than in each of the other groups
(minimum t-2-96, K=3 (adjustment for multiple
comparisons), P-=0 1). Significant group differences
(F=8- 6, P< 0 0O1) were not so obviously confined to
overreaction in girls (the simple interaction effect of
group by subscale of the Bristol social adjustment
guide just failed to reach the critical F ratio: F= 1-93,
P=0 12). Group differences resulted from higher
scores in preneurotic as well as preschizophrenic girls
compared with normal controls (minimum t-2-5,
K=3, P=003) or girls with preaffective psychotic
disorders. Figure 1 also shows that social maladjust-
ment in preschizophrenia was more apparent in the
boys than in the girls, particularly in relation to
(including both overreaction and underreaction) was
obvious in the preneurotic girls when they were
compared with their same sex peers and was unremark-
able in the boys (fig 1).
All of these differences between groups should,
however, be considered in the light of the significant
two way interaction of group and age, indicating that
the size of the differences between groups changed
with time. The overall scores in table II show that
the normal controls and the group with preaffective
psychotic disorders obtained similar scores at the ages
of 7 and 11, whereas those with preschizophrenia and
preneurosis had increased scores by the age of 11
(F=2.61, P=0 1l and F=4-88, P-003, respectively).
In addition, whereas those with preschizophrenia
scored significantly higher than the normal controls at
both ages (at 7 t=336, K=3, P<0'003 and at 11 t=5-3,
K=3, P<0-003), those with preneurosis differed little
from the normal controls at the age of 7 (t= 18, K=3,
P=0-21) but had significantly raised scores by the age
of 11 (t=4-5, K=3, P<0-003). Since no other inter-
action with age was found this progression between 7
and 11 seems to take place in both boys and girls.
PROFILE ANALYSIS OF CORE SYNDROMES BY GROUP
Although those with preschizophrenia had higher
social adjustment scores at the ages of 7 and 11
and those with preneurosis had higher scores at 11,
the degree and type of social maladjustment varied
between boys and girls. When the scores for over-
reaction and underreaction were aggregated in the
preschizophrenic group the boys had high overreaction
scores whereas the girls differed from controls across
the Bristol social adjustment guide as a whole; there
was, however, a trend between sex and overreaction
To investigate whether the profile across the core
syndromes for boys and girls differed, multivariate
profile analysis'6 was carried out using standardised
scores by sex for each core syndrome after the scores
for the two anxiety and the two hostility syndromes
were combined to reduce the number of dependent
variables. The core syndromes were grouped together
as overreaction-that is, acceptance anxiety to restless-
ness-and underreaction-that is, unforthcomingness
to dismissing adult values.
In preschizophrenia the profiles for boys and girls
were similar-that is, parallel-at the age of 7 (F= 1-5,
P=0-15). At the age of 11, however, the profiles
were significantly different (F=4-2, P<0 01), this
difference arising at the transition from overreaction to
underreaction-that is, the girls showed fewer features
of overreaction and more of underreaction than the
boys (fig 2).
As the significant differences between the pre-
neurotic group and normal controls largely occurred at
the age of 11, profile analysis was restricted to this age.
The profile scores as a whole for the preneurotic group
were higher than those of the controls except for the
syndromes of unforthcomingness and miscellaneous
nervous symptoms (fig 3). The increase was more
obvious in the girls but the profile was not different
from that ofthe boys.
FIG 2-Mean standardised scores for core syndromes of Bristol social
Age I I
FIG 3-Mean standardised scores for core syndromes of Bristol social
adjustmentguide in preneurosis
Even at the age of 7 subjects who will later develop
schizophrenic illness differ from schoolmates in the
eyes of their teachers. Preschizophrenic boys were
more likely to be rated as overreactive than boys who
later developed an affective psychosis or neurosis or
who were psychologically normal
tended to be distinguished by being "anxious for
acceptance," being hostile to other children and adults,
and engaging in inconsequential behaviours. Half of
these boys were considered to be as deviant as the
most deviant 10% of the school population, and we
found little evidence of the social withdrawal that has
sometimes been regarded
schizophrenic children. Both preschizophrenic and
preneurotic girls were more maladjusted socially than
normal controls, especially by the age of 1 1.
In the preschizophrenic group the aggregated scores
for overreaction and underreaction showed
increase between the ages of 7 and 11, but the profiles
of the core syndromes, which were similar for both
sexes at the age of 7, differed significantly by the age of
11. Preschizophrenic boys showed a similar profile of
significantly more likely to be rated as underreactive,
particularly as withdrawn but also as unforthcoming
and depressed at the age of 11. This could correspond
to the asociality which is often a precursor as well as a
sequel to a first episode ofschizophrenic psychosis.
Subjects with affective psychotic disorders were less
clearly abnormal than those with schizophrenia
either age. Subjects who later became neurotic also
showed few abnormalities at the age of 7 but became
steadily more maladjusted between the ages of 7 and
11. Social maladjustment (including both overreaction
and underreaction) was significant in girls but not in
children who later develop schizophrenia and those
who later develop an affective psychosis or neurosis.
The increase in abnormality in the preneurotic group
between the ages of 7 and 11 may reflect an early stage
of a depressive or anxiety disorder, a finding consistent
with the suggestion that the jump in the rate of
depression between child and adulthood occurs in
as adults. They
as characteristic of pre-
ages, whereas the girls were
early adolescence and is substantially higher in girls.'7
Interpretation of this observation is hazardous because
the nature ofthe disorder in these patients is uncertain.
They were all admitted for non-psychotic disorder by
the age of 28, but they all had reports of depression
in their case notes and many also had experienced
COMPARISON WITH OTHERSTUDIES
In several respects our findings are in agreement
with those ofWatt et al, who studied school records of
preschizophrenic patients and normal controls.3 They
found considerable sex differences, the boys showing a
pattern of irritability, disagreeableness, and defiance of
authority while the girls were insecure, shy, and
participated less in groups. As in our study, the
changes were more striking in the boys; our study
supports the conclusion of Watt et al that overinhibi-
tion is characteristic of preschizophrenia in girls rather
than boys. But the two studies differ with respect to the
sequence of changes in the two sexes. Watt
concluded that there was greater evidence of progres-
sion of deviance in boys than in girls between kinder-
garten and the age of 18. Our findings, resting on
at two specific
maladjustment is certainly present (mainly as over-
reaction) in boys as a group as early as 7 and the profile
changes little by the age of 11, whereas the abnormali-
ties (which include withdrawal, unforthcomingness,
and depression) seen in girls at the age of 11 are not
significant at the age of 7. "Progressive deviance," as
defined by Watt et al, could describe the changes in
girls, but in boys the deviance
probably occurring before the age of 7 and progressing
in a limited way duringmuch ofchildhood.
ages, suggest that social
is more obvious,
The onset of psychotic illness in adult life, through-
out the period ofmaximum fertility, is one of its most
characteristic but unexplained
abnormalities in social behaviour can be detected 15 or
more years before the onset of the more characteristic
illnesses should be seen in the context of a lifetime
course of development. Moreover, our data suggest
that the major forms of psychotic illness differ in the
pattern and timing oftheir precursors in childhood and
that sex differences are prominent. Changes over time
that differ between the two sexes may be relevant in
further understanding the origins ofpsychosis.
17 SEPTEMBER 1994
children could be an early sign of psychotic
illness in adulthood
hostility, and inconsequential behaviour-is un-
common in 7 and 11 year old boys in general
but was common at these ages in those who
developed schizophrenia in adult life
* Girls who developed schizophrenia did not
behave abnormally at the age of 7 but had
become noticeably withdrawn (rather than over-
reactive) by the age of 11
* The rate of development of the capacity
for social interaction may be important in
determining the risk of psychosis and other
psychiatric disorder in adulthood
This work was supported by the Medical Research Council Download full-text
and in part by a grant from the Theodore and Vada Stanley
Foundation. We thank John Bynner, Peter Shepherd, Kevin
Dodwell, and colleagues at the Social Statistics Research
Unit, City University, London, for their help with the data
from the national child development study; the National
Birthday Trust and the National Children's Bureau for
sponsoring the study; Andre Charlett, Diana Kornbrot, Josie
Pearson, Heinz Hafner, Peter Jones, and an anonymous
statistical referee from the BMY for their valuable statistical
advice; and Malcolm Weller for useful discussions.
The following is a fuller description of the aggregated scales
of the Bristol social adjustment guide with behaviours that
define each core syndrome.
Overreaction is an aggregate ofsix core syndromes.
Anxiety for acceptance by children comprises behaviours
including buffoonery, being overly brave, showing off, and
Anxiety for acceptance by adults is separated into two parts:
(a) being overly friendly-that is, bringing gifts or other
objects to the teacher-and talking excessively to the teacher
and (b) seeking to engage excessively or monopolise the
teacher and requiring sympathy.
Hostility towards other children includes various criteria for
being unpleasant to other children.
Hostility towards adults has three components (a) showing
lability of mood when asked to do something or when
actually doing it and having variable standards of perform-
ance; (b) damaging personal property, lying and using bad
language, or being obscene; and
Inconsequential behaviour comprises (a) poor concentration
or lack of perseverance and (b) carelessness and untidiness,
lolling about, and being mischievous.
Restlessness includes being too restless in individual games
orwhen working alone and lacking persistence.
(c) stealing and being
Underreaction is an aggregate offour core syndromes.
Unforthcomingness-Timidity-for example, not initiating
conversation and being reluctant to approach the teacher-
Withdrawal-being distant and cut off from people and
Depression has two parts: (a) variation in mood, the child
sometimes being alert and sometimes lethargic and lackdng in
interest and (b) apathy and lifelessness, or miserableness, and
Dismissing adult values has two parts: (a) unwillingness to
work except when compelled to and (b) suspiciousness,
selfishness, and untrustworthiness.
Miscellaneous symptoms has two parts
playing only with younger children, or being bullied and (b)
Miscellaneous nervous symptoms include stuttering, twitch-
ing, and biting nails badly.
1 Lane EA, Albee GW. Childhood intellectual differences between schizo-
phrenic adults and their siblings. AmJOrthopsychiatry 1965;35:747-53.
2 Offord DR. School performance of adult schizophrenics, their siblings and age
mates. BrJPsychiatry 1974;125:12-9.
3 Watt NF, Stolorow RD, Lubensky AW, McClelland DC. School adjustment
and behaviour of children hospitalised for schizophrenia as adults. Am J
4 Watt NF. Patterns of childhood social development in adult schizophrenics.
Arch Gen Psychiatry 1978;35:160-5.
5 Rutter ML. Psychosocial resilience and protective mechanisms. In: Rolf JE,
Master AS, Cicchetti D, Neuchterlein KH, Weintraub S, eds. Risk and
protective factors in the development ofpsychopathology. New York: Cambridge
University Press, 181-214.
6 Done DJ, Johnstone EC, Frith CD, Golding J, Shepherd PM, Crow TJ.
Complications of pregnancy and delivery in relation to psychosis in adult
data from the British perinatal
7 McCreadie R. The Nithsdale schizophrenia survey. I. Psychiatric and social
handicaps. BrJPsychiatry 1982;140:582-6.
8 Wing J, Cooper J, Sartorius N. The description and classification of
an instruction manual for the PSE and
CATEGO system. London: Cambridge University Press, 1974.
9 Stott GH. The social adjustment of children. Manual to the Bristol social
adjustmentguides. London: Hodder and Stoughton, 1987.
10 Clausen JA, Kohn ML. Relation of schizophrenia to the social structure of a
Washington, DC: American Association for the Advancement of Science,
11 Cannon-Spoor HE, Potkin SG, Wyatt RJ. Measurement of premorbid
adjustment in chronic schizophrenia. SchizophrBull 1982;8:470-84.
12 Hartman E, Milofsky E, Vaillant G, Oldfield M, Falke R, Ducey C.
Vulnerability to schizophrenia: prediction of adult schizophrenia using
childhood information. Arch Gen Psychiatry 1984;41:1050-6.
13 Amnbelas A. Preschizophrenics: adding to the evidence, sharpening the focus.
14 General Register Office. Classification ofoccupations. London: HMSO, 1960.
15 SPSS. Users'guideforSPSSX release 2.1. 2nded. Chicago: SPSS, 1986.
16 Tabachnick BG, Fidell LS. Using multivariate statistics. 2nd ed. New York:
Harper and Row, 1989.
17 Gotlib IH, Hammen CL. Psychological aspects ofdepression: towards a cognitive-
intrpersonal integration. Chichester: Wiley, 1992.
mortality survey sample. BMJ
In: Pasamanick B,
ed. Epidemiology of mental disorder.
Box 116, Addenbrooke's
ColinJ Morley, honorary
campaign to reduce risk ofcot
ChristineMH Hiley, Colin J Morley
In December 1991 the Department of Health's "Back
to Sleep" campaign advised that babies should not
sleep on their front, be exposed to cigarette smoke, or
be overheated.' Since then cot deaths have halved,
from 912 in 1991 to 456 in 1992.2 This has been
attributed to the campaign,3 with little evidence that
child care practice has in fact changed. This study
compared the way mothers cared for their infants
before and after the campaign.
Subjects, methods, and results
Questionnaires were sent to two randomly selected
groups of 450 mothers of normal term babies born in
Cambridge, Huntingdon, or Bury St Edmunds. The
first group gave birth at least eight months before the
campaign. The second group gave birth after the
campaign and received the questionnaire when their
baby was six months old. Tog values were calculated
using similar methods to those of Fleming et al.4
Approximately equal numbers of babies were chosen
from each month and hospital.
The first group returned 385 (86%) completed
questionnaires and the second 399 (89%). The high
response rate adds robustness to this study. There were
no social or demographic differences between the
groups and results were consistent from all three
Babies' sleeping positions changed after the cam-
paign at the three ages studied (table). Sleeping
position also changed with age, older babies being
more likely to sleep on their backs than on their sides.
There was no influence of mothers' education, social
class, marital status, or parity.
Although duvets or quilts were used less after the
campaign, their use increased with the infant's age.
The use of more than three blankets increased,
younger babiesusing more
babies. The tog value for the babies' coverings were
unchanged. After the campaign mothers having their
first baby were less likely to use a duvet than were
mothers with other children (for newborn babies
51/182 (28%) v 156/217 (72%), P< 00002; for infants
of3 months 62/182 (34%) v 143/217 (66%), P<0001;
and for infants of 6 months 62/182 (34%) v 143/217
BMJ VOLUME 309
17 SEPT-EMBER 1994