-
10.1192/bjp.108.456.642Access the most recent version at DOI:
1962, 108:642-654.BJP
Andrew McGhie and S. M. RussellThe Subjective Assessment of
Normal Sleep Patterns
Referenceshttp://bjp.rcpsych.org/content/108/456/642#BIBLThis
article cites 0 articles, 0 of which you can access for free
at:
permissionsReprints/
[email protected] To obtain reprints or permission to
reproduce material from this paper, please write
to this article atYou can respond
http://bjp.rcpsych.org/letters/submit/bjprcpsych;108/456/642
from Downloaded
The Royal College of PsychiatristsPublished by on October 24,
2014http://bjp.rcpsych.org/
http://bjp.rcpsych.org/site/subscriptions/ go to: The British
Journal of PsychiatryTo subscribe to
-
THE SUBJECTIVE ASSESSMENT OFNORMAL SLEEP PATTERNS
By
ANDREW McGHIEand
S. M. RUSSELL
INTRODUCTIONWhENwe consider the immense human significance of
sleep, the absolutenecessity for us to spend a considerable part of
our lives in abject mentalannihilation, it is remarkable how little
we know about it. . . .In thesewords, Sir John Eccles introduced
the recently published Ciba FoundationSymposium on TheNature of
Sleep(3), in which are described many of theadvances which have
been made in recent years in our attempts to understandthe
phenomena which are involved in the state of sleep. In one of the
papersincluded in this Symposium, Bremer (2) reviews the work which
has beenreported since 1954 on the neurophysiological mechanisms
involved in sleep.The relationship of dreaming to the depth of
sleep has been investigated bysuch workers as Kleitman (5), Dement
(4) and Wolpert (11). These studieshave been made possible by the
application of modern techniques of measuringmental activity. The
use of electroencephalographic methods has allowed anapproximate
classification of the different levels of sleep which enables
differentworkers to adopt a standard frame of reference.
Measurements of rapid eyemovements and of changes in muscle
potential during sleep have provided moreobjective methods of
assessing dream activity. Other recent studies haveattempted to
make an objective assessment of the physical and mental effectsof
sleep deprivation (Morris et a!., 7, Bliss eta!., 1, Murray eta!.,
8, Wilkinson, 9,Williams et a!., 10).
Many of the findings arising from this recent interest in the
state of sleepare of interest to psychiatrists who have long
recognized the correspondencebetween disturbance of the normal
sleep pattern and mental illness. Sleep disturbance has been noted
in such varied psychiatric conditions as depression,anxiety states,
delirium and catatonia. In depressive illness the disturbance ofthe
sleep pattern may be regarded as one of the most important and
characteristic of the clinical features. The prevalence of early
morning wakening indepressive psychosis has caused this aspect of
sleep disturbance to be regardedas a symptom of particular
importance in establishing the diagnosis and inassessing the
patient's response to treatment.
In view of the number of studies which have been made on
different aspectsof sleep in recent years, it is perhaps somewhat
surprising to find that verylittle appears to be known about the
normal pattern of sleep and the way inwhich this varies with such
factors as age, sex and social class. Although severalstudies have
been made of the variations in the sleep rhythm in infancy,
littleattention has been paid to the sleep cycle of the adult
population. It tends to beassumed that the average healthy adult
sleeps around 8 hours a night and afew reported studies of small
groups who were at liberty to sleep as long as they
642
-
THE SUBJECTIVEASSESSMENTOF NORMAL SLEEP PA1TERNS 643
wished (e.g., Lewis, 6), substantiate our impression that this
period representsthe normal human requirement. Although we may
recognize that the length ofthe sleep period decreases with age,
there have been no critical studies of thevariation of sleep with
age. The present study attempts to study the sleeppatterns of a
group of normal subjects and to assess the variations in the
sleeprhythm with age.
METHODIn spite of the obvious deficiencies of subjective
evaluation, the question
naire method seemed the only feasible method of surveying a
population largeenough to allow a statistical comparison of the
various age groups. In order tolimit this subjective influence as
much as possible, our questionnaire was constructed so that all
possible responses were already noted after each question,the
subjects being required to underline the response appropriate to
them. Thefirst section of the questionnaire included items relating
directly to the quantitative aspects of the sleep pattern. It thus
contained questions on the habitualtime of retiring to bed, the
time at which the subject normally falls asleep,normal wakening
time, rising time, etc. The second part of the
questionnaireenquired into the more qualitative aspects of
sleepe.g. did the subject describehimself as a light, moderate, or
deep sleeper? was his sleep frequently brokenby night wakening? did
he tend to feel tired in the morning or during the day?The last
section contained items related only indirectly to sleep. The
subjects wereasked if they were in the habit of taking hypnotics,
if they would describe themselves as being of a nervous
disposition, and finally they were asked a numberof questions on
the state of their physical health. The subjects were not asked
tonote their names on the questionnaire, but they were asked to
give details of theirage, sex and occupation, the latter being used
to obtain a rough grading of thesubject's social class according to
the Registrar General's Classification ofOccupations.
SELECTIONOF SUBJECTSThe subjects who took part in the enquiry
were made available through the
kind co-operation of a variety of agencies in the Dundee and
Glasgow areasincluding Further Education authorities, community
centres, Territorial Armyunits and old people's clubs. Each of the
bodies involved supported the enquiryby distributing the
questionnaires among its members and collecting thecompleted forms
for subsequent dispatch to us. It will be readily seen that
ourmethod of selection was unlikely to provide us with a random
sample of thegeneral population, most of our subjects belonging as
they did to some educational or social organization. The enquiry
was designed to include subjectsfrom both the Dundee and Glasgow
areas, to allow not only an extension in thenumber of subjects
taking part, but also a comparison between the two communities. It
proved impossible to obtain a wholly accurate assessment of
thenumber of subjects who refused to co-operate in this enquiry,
but it appearedthat approximately 80 per cent. of the subjects who
were approached completeda questionnaire. In this manner a total of
2,446 completed questionnaires weremade available for subsequent
analysis.
The age distribution of the survey population is shown in Table
I. Fourteenper cent. of the subjects in the total sample omitted to
note their age on thequestionnaire and consequently could not be
considered in the analysis. Forpurposes of comparison the
comparative figures for the Scottish populationbased on the 1961
census figures are also noted. We have shown all subjects
-
644 THE SUBJECTIVEASSESSMENTOF NORMALSLEEPPATTERNS [Sept.over
the age of 65 years as constituting a single age group to allow
ready comparison with the census figures. In analysing our data,
however,we examinedthese subjects in two separate age groups of
6574years and over 75 years inorder to observe any further
noticeable changes occurring some time after themale retiral age.
It may be seen that the survey sample deviates most markedlyfrom
the national sample at the age groups 1524,4554and 5564years.
Ourpopulation over-represents the younger group and
under-represents the middleaged sector of the community. This is
particularly marked in the case of the malesubjects, the deviation
being much less in the case of the females. It may alsobe seen that
our male sample under-represents the 65 year and above group(7 per
cent. as compared with 12 per cent. in the national population). In
thecase of the female subjects of 65 years and over, the reverse is
true (18 per cent.as compared with 14 per cent. in the national
population). The sample as a wholecarries an over-weighted sex
ratio (160 @2F: 100 M) as compared with the sexratio in the total
population of Scotland in 1961 (108 @5F: 100 M).
TABLE IAge Distribution
Not1524 2534 35-44 4554 55-64 65+ Known
Survey Sample .. 23% 15% 17% 12% 7% 12% 14%National Population
19% 19% 19% 17% 13% 13% 0%Survey Sample
(Male) .. .. 31% 23% 18% 11% 5% 7% 5%National Population
(Male) .. .. 19% 19% 20% 18% 12% 12% 0%Survey Sample
(Female) .. .. 21% 13% 19% 14% 10% 18% 5%National Population
(Female) .. .. 19% 18% 19% 17% 13% 14% 0%
On the basis of the Registrar General's Classffication of
Occupations, eachsubject was placed in a social class category. The
percentage distribution of thesocial class categories is shown in
Table II for our male subjects, together withthe comparative
figures of the male social class distribution for Scotland as
awhole. It was not possible to make a similar comparison for our
female subjects,as the national figures are calculated separately
for single women and workingwives. It can be seen that the social
class distribution of the males in our surveysample is almost
identical to that of the national population.
TABLE IISocial Class Distribution (Male)
I II ifi iv VSurvey Sample .. .. 3% 14% 50% 19% 14%National
Population .. 3% 13% 51 % 18% 15%
METHOD OF ANALYSISThe information on the completed questionnaire
was coded, transferred to
punch-cards and subsequently collated by a Powers-Samas machine.
The resultant data were analysedby applyingthe@ techniqueor, where
multiplecorn..parisons were involved, Ryan's Significance Test of
Proportions. The mainvariables considered in the analysis were age,
sex, area and social class and the
-
1962] BY ANDREWMCGHIEAND S. M. RUSSELLrelation between each of
these factors and the various items of informationgiven by the
questionnaire responses was examined. Age groups were consideredin
decades beginning at 1524years and ending with the group 75 years
and over.Although we were mainly concerned with the effect of age
on the normal sleeppattern, we also considered any variations
related to sex and social class differences. The preponderance of
females in our sample was particularly evidentin the upper age
levels. In order to ensure that apparent age differences werenot
due to the unequal sex ratio, we have analysed the male and female
responsesseparately for each item which showed a significant sex
difference. Throughoutthe analysis of the data we have accepted as
significant only these items whichare beyond the 1 per cent. level.
(In most cases the differences involved were infact beyond the
O@lper cent. level.)
Our treatment of the data has been selective in that we were
mainly interested in finding the answers to a number of specific
questions which appearedto be of relevance to the clinical field.
We have been less interested in the averagepatterns of response
than in the responses which deviate appreciably from theaverage. In
analysing the data we have accordingly chosen an arbitrary
cut-offpoint in the distribution of the responses given to each
item of the questionnaire,and in presenting our results, it is the
responses beyond this point which will beconsidered. The analysis
of the main variable, that of age, is most easily repre-.sented in
the form of histograms.
RESULTSBefore considering the responses which deviate from the
average we might
first consider the distribution of the number of hours slept as
reported by thesurvey population as a whole (Fig. 1). It may be
seen that this approaches closelyto a normal distribution curve, 62
per cent. of the subjects reporting an averagesleep of between 7
and 8 hours per night.
We next examine the variation with age, sex, and social class of
the responsesgiven to each item of the questionnaire. Any
differences between the two urbanareas covered by the survey are
also indicated, but there were very few caseswhere such differences
reached a significant level.
1. Average Hours Sleep (5 hours)We have restricted our analysis
of this item to those subjects who reported
that they habitually slept for a maximum of 5 hours per night. A
comparison
645
300
200
005% 2%
S65es. 7$hrs. 9lOPws. [email protected] S lirs.
FIG. 1
- (p
-
1962] 647BY ANDREW MCGHIE AND S. M. RUSSELLlittle variation with
age in males. Pre-sleep difficulty is particularly evident inwomen
after the age of 55 years, reaching a peak between the ages of 65
to 74years when 30 per cent. of the women gave a positive response.
It would appearthat the increased difficulty in getting off to
sleep with advancing age is more orless specific to women.
3. Time of Wakening (before 5 a.m.)In dealing with this item we
have concentrated on subjects who reported
that they usually waken before 5 a.m. In this case no
significant relationshipwas found between early morning wakening
and the three variables, social class,sex and area. Indeed, early
morning wakening, as here defined, constituted theonly variable in
the sleep pattern which did not show a clear sex difference.
Thevariation in early morning wakening with age is indicated in
Figure 4. Although
55.
FIG. 4
thefrequencyof earlymorning wakening
increasesslightlywithage,theincreases are insignificant until the
age group 6574years. Of the subjects in thisage group, 18 per cent.
report early morning wakening although this figure fallsto 12 per
cent. after the age of 75 years. Although early morning wakening
asrepresented in this study appears to have a very low incidence in
the normalpopulation, the significant increase (from 5 per cent. to
18 per cent.) at the ageof 65 years suggests that, in assessing
early morning wakening as a clinicalsymptom, the age of the patient
should be taken into consideration.
4. Depth of Sleep (Light)Subjects were asked to describe their
usual sleep level as being deep,
moderate or light. In our analysis we have considered only those
subjects whodescribe themselves as being light sleepers. Once again
there was no significantdifference between the two areas. With
social class the differences were significant, the tendency towards
light sleep increasing as the social class grading movesdownwards
from Class I to Class V (x@=i5 85;p
-
648 THE SUBJECTIVEASSESSMENTOF NORMALSLEEPPATTERNS [Sept.item is
presented in Figure 5 for the next two sexes separately. In the
case ofthe men, it can be seen that light sleeping occurs most
frequently in the age range3544years (22 per cent.) and after the
age of 65 (25 per cent.). There is aninteresting drop off in the
incidence of light sleeping in men between the ages of5564years at
which time the reported incidence is at its lowest. In women
theincrease in light sleep with age is much more marked. By the age
of 35 years,25 per cent. of the women described themselves as being
light sleepers and theincidence accelerates significantly after the
age of 45 years to reach a peak of40 per cent. Apart from a small
but insignificant decrease between the ages of5564years, the level
remains constant at 40 per cent. thereafter. It would appearthat
light sleeping is more pronounced in women, particularly from
middle ageonwards.
so454035Iz 3025205I05
FIG. 5
5. DisturbedSleep(FrequentNightWakening)In this item we asked
subjects to describe to what extent their sleep was
usually disturbed by night wakening and we have examined only
those subjectswho reported that this occurred with great frequency.
Although there was noappreciable social class or area difference,
there was again on this item a markedsex difference (x2=4l @07;
p
-
1962] BY ANDREWMCGHIEAND S. M. RUSSELL 649women. At the age of
45 years, 21 per cent. of the women reported disturbedsleep, 30 per
cent. at the age of 55 years and, between the ages of 6574years,43
per cent. of the women described themselves as wakening frequently
duringthe night. Here again we might conclude that this change in
the sleep pattern isnot evident in men until after the age of 65
years, whereas in women, disturbedsleep may be apparent from middle
age onwards.
6. Morning Tiredness (Frequent)The subjects were asked if they
usually felt tired in the morning and here
we have considered only those subjects who described frequent
morning tiredness. Examination of the data showed a relationship
between morning tirednessand social class (x2==ll P65; p
-
650 THE SUBJECTIVEASSESSMENTOF NORMALSLEEPPATFERNS [Sept.two
sexes separately in Figure 8. Both sexes show very little variation
in thefrequency of day tiredness until the age of 75 years where
there is a significantincrease in the incidence reported by both
sexes, reaching a maximum incidenceof 20 per cent. for men and 29
per cent. for women.
5045
8. Hypnotics
so45(MALE) (FEMALE)4035
1-z 30
@25205l0
S
40
@ 34@2520$510
5
$
15@25+ 35+ 4555' 65+ 75@ $54' 25' 35+ 45* 55+ 65' 75'
FiG. 8
Subjects were asked if they were in the habit of taking a
hypnotic and asubsequent question attempted to establish the source
of the sedative which wasbeing taken. In analysing this item we
have taken into consideration only thosesubjects who describe
themselves as taking a regular hypnotic issued eitherdirectly by a
doctor or through a chemist. No variation between social class
orarea was apparent, but again the sex difference was significant
(@=4l 40;p
-
9. Nervous DispositionThe subjects were asked ifthey would
describe themselves as being by nature
a nervous person. The incidence of positive responses to this
question wasapproximately the same in each social class group, but
there was a significantdifference between the two areas (A@=3713;
p
- 652 THE SUBJECTIVEASSESSMENTOF NORMALSLEEPPATFERNS [Sept.tend
to waken frequently during the night (x2=4S 73 ; p
-
1962] BY ANDREWMCGHIEAND S. M. RUSSELL 653a reduction in the
number of hours slept, an increase in the time taken ingetting off
to sleep and a greater tendency for sleep to be light and easily
disturbed by frequent night wakening. At the same time we have
noted that as ageadvances people tend to complain less of morning
tiredness. It is possible that,as age advances, people do not
subjectively experience any great harm from thechange in their
sleep pattern, and that, as we grow older, we require less
sleep.The variation of the sleep pattern with age is much more
marked in the case ofthe female subjects of our sample. Although
the two sexes do not differ significantly in the average number of
hours slept, there are a number of significantvariations in the
form which their sleep takes. In general the sleep pattern ofthe
female tends to change earlier than that of the male, usually
during middleage. The change in the male sleep pattern with age is
not only less marked buttends to be fully established latersome
time after the retiral age of 65 years.We have also seen that the
incidencein the use of hypnoticsappears to increaserapidly with age
but our findings suggest that this increase is almost
entirelyspecific to women, 45 per cent. of whom admitted to taking
a regular hypnoticin their 70's. The incidence of reported
nervousness is again much higher in thecase of women and reaches
its peak between the ages of 25 and 44 years. Thesefindings might
be, of course, interpreted as merely indicating that men are
lessinclined to admit to taking a regular hypnotic or to being of a
nervous disposition. A positive relationship has been demonstrated
between the subjectiveestimation of nervousness and specific
changes in the pattern of sleep asevaluated by the subjects. Those
people who describe themselves as beingnervous tend to report more
difficulty in getting off to sleep and a more disturbednight's
sleep. They are also more likely to complain of physical symptoms
suchas heart conditions, high blood pressure and severe headaches.
It would thusseem that our general physical and mental state is
reflected in the form and qualityof our sleep.
Investigations of this type seem worthwhile in providing some
sort ofstandard upon which an individual's sleep pattern may be
more objectivelyevaluated. It is our intention to follow this
investigation with a similar studyapplied to the patterns of sleep
typical of patients receiving psychiatric treatment.
ACKNOWLEDGMENTSWe should like to express our thanks to the many
people whose co-operation and advice
made this surveypossible.We are indebted to Dr.!. C. R.
Batchelor, Physician
Superintendent,DundeeRoyalMentalHospital,forhisconstructivecommentsintheinterpretationofthedata.
We should also like to acknowledge the help of Dr. J. A. Chapman
and Miss S. Dick,of Dundee Royal Mental Hospital, and Miss Audrey
Brinkler and Miss Margaret McLaggan,of the Department of
PublicHealth,Queen'sCollege,in assistingin the laborious task of
collatingthe results obtained in the survey. Finally, we are most
indebted to the many agencies in theDundee and Glasgow areas who
kindly distributed the questionnaires among their members.
REFERENCES
1. BLISS,E. L., CLARK,L. D., and WesT, C. D. (1959). Studiesof
sleep deprivation: relationship to schizophrenia,A.M.A. Arch.
Neurol. and Psychiat., 81, 348.
2. BREMER,F. (1954). Theneurophysiological problem of sleep.In
J. F. Delafresnaye (Ed.)Brain Mechanisms and Consciousness. Oxford:
Blackwell Scientific Publications.
3. CIBAFOUNDATION(1961). Symposium, The Nature of Sleep. London:
J. & A. ChurchillLimited.
4. DEMENT, W., and KLErrMAN, N. (1957).CyclicvariationsinEEG
duringsleepand theirrelationto eye movements, body motilityand
dreaming,Electroencepha/og.and C/in.Neurophysiol., 9, 673.
5. KLEITMAN,N. (1939).Sleep and Wakefulness. Chicago: University
of Chicago Press.6. LewIs, H. E. (1961).Sleeppatterns on polar
expeditions.In Ciba Foundation Symposium
on The Nature of Sleep.London: J.& A. ChurchillLimited.
-
654 THESUBJECTWEASSESSMENTOFNORMALSLEEPPATTERNS7. Moiuus, G. 0.,
WLLUAMS,H. L., and LUBIN,A. (1960). Misperceptionand
disorientation
during sleep deprivation,A.M.A. Arch. Gen.Psychiat., 2, 247.8.
MURRAy,E. J., SCsEIN, E. H., ERJKSON,K., HILL, W. F., and C0REN, M.
(1959). The
effects of sleep deprivation on social behaviour,J. Social
Psycho!., 49, 229.9. Wiuw@so@,R. T. (1957). Effectsof lack of
sleep,FPRC Rep. No. 961. Cambridge:
Medical Research Council.10.Wn..u@is,H.
L.,LUBIN,A.,and000DNOW,J.J.(1959).Impairedperformancewith
acutesleeploss,PsychologicalMonographs,73,14.II. WOLPERT,E. A.
(1960). Studiesin psychophysiology of dreams,Arch. Gen. Psychiat.,
2,
231.
Andrew McGhie, M.A., Ph.D., Principal Psychologist and Lecturer
in ClinicalPsychology
S. M. Russell, M.B., Ch.B., D.C.H., D.P.M., Registrar (now with
Departmentof Child Psychiatry, Glasgow)
From the Dundee Royal Mental Hospital and the Department of
Psychiatry,University of St. Andrews