A new depression scale designed to be sensitive to change.

S A Montgomery and M Asberg

BJP 1979, 134:382-389.
Access the most recent version at DOI: 10.1192/bjp.134.4.382

References This article cites 0 articles, 0 of which you can access for free at:
http://bjp.rcpsych.org/content/134/4/382#BIBL
Reprints/ To obtain reprints or permission to reproduce material from this paper, please write
permissions to permissions@rcpsych.ac.uk

You can respond http://bjp.rcpsych.org/letters/submit/bjprcpsych;134/4/382

to this article at
Downloaded http://bjp.rcpsych.org/ on October 18, 2014
from Published by The Royal College of Psychiatrists

To subscribe to The British Journal of Psychiatry go to:

http://bjp.rcpsych.org/site/subscriptions/
Brit. J. Psychiat. (1979), 134, 382—9

A New Depression Scale Designed to be Sensitive to Change

By STUART A. MONTGOMERY and MARIE ASBERG

SUMMARY The construction of a depression rating scale designed

to be particularly sensitive to treatment effects is described. Ratings
of 54 English and 52 Swedish patients on a 65 item comprehensive
psychopathology scale were used to identify the 17 most commonly
occurring symptoms in primary depressive illness in the combined
sample.
Ratings on these 17 items for 64 patients participating in studies of
four different antidepressant drugs were used to create a depression
scale consisting of the 10 items which showed the largest changes with
treatment and the highest correlation to overall change.
The inter-rater reliability of the new depression scale was high.
Scores on the scale correlated significantly with scores on a standard
rating scale for depression, the Hamilton Rating Scale (HRS), indicating
its validity as a general severity estimate. Its capacity to differentiate
between responders and non-responders to antidepressant treatment
was better than the HRS, indicating greater sensitivity to change. The
practical and ethical implications in terms of smaller sample sizes in
clinical trials are discussed.

Introduction for depressive illness where sensitivity and

The most common use of psychiatric rating accuracy of change estimates were to be major
scales is probably for comparison of effects of criteria for the inclusion of items. This work
new drugs to standard treatment. In most trials was made possible by the recent introduction of
of antidepressant drugs a difference can be a new, comprehensive psychopathological rating
demonstrated between pharmacologically active scale or CPRS (Asberg et al, 1978). The CPRS
compounds and placebo (Morris and Beck, is composed of 65 scaled items covering a wide
1974) but only rarely are consistent differences range of psychiatric symptoms.
found between active drugs even when they are
known to have different mechanisms of action. Patients and Ratings
One possible explanation is that the standard The selection
of itemsforthe depressionscale
rating scales are not sensitive enough to pick up was based on ratings on the CPRS of 106
such differences (Angst, 1972) which is not depressed patients, 33 men and 73 women,
surprising if the scales were not designed participating in clinical trials of antidepressant
specifically for that purpose. The result has drugs. Thirty-three were out-patients and 73 in
often been that the scales reflect diagnostic patients. In the majority of cases, two raters
features rather than being sensitive to change. participated in the interview. Only patients
We therefore decided to construct a rating scale with a primary depressive illness (Feighner et al,
382
 STUART A. MONTGOMERY AND MARIE ASBERG 383
1972) were included. Inventories (Gurney ci al, cholinergic effects (Mass, 1975) ; it is also
1972) were applied to ensure diagnostic and thought to have a stronger sedative effect than
descriptive uniformity. Within this group, a the other drugs (Silverstone and Turner, 1974).
wide variation in patient characteristics was Mianserin appears to lack uptake inhibiting
sought and the sample includes endogenous effects (Ferl ci al, 1973) as well as anticholinergic
and reactive, psychotic and non-psychotic, effects (Coppen ci al, 1976) and its mechanism of
bipolar and unipolar out-patients and in action is not clear.
patients from a wide age range (18—69 years). Of the 64 patients for whom scores were
Patients from two countries (England n = 54, available from before and after four weeks of
and Sweden n = 52) were included in order to treatment 35 were simultaneouslyrated on the
reduce cultural bias in the selection of items. Hamilton Rating Scale (HRS) (Hamilton, 1967)
CPRS scores after four weeks therapy with four and on a 7-point scale for global severity of
antidepressant drugs with different pharma illness.
cological profiles were used to study change with
treatment. The drugs were mianserin, ami Construction of the Scale
triptyline, maprotiline and clomipramine. Parametric statistical methods were used
Amitriptyline (Tuck and Punell, 1973), mapro except when dealing with ranked data as
tiline (Maitre ci al, 1971) and clomipramine recommended by Anderson (1961) and Boneau
(Hamberger and Tuck, 1973) block neuronal (1961). The difference between pre-treatment
reuptake of noradrenaline, directly or by means scores has been used as an estimate of change,
of endogenously formed metabolites. Clomi both for individual and sums of items.
pramine is also a potent serotonin uptake
blocker (Asberg ci al, 1977). Amitriptyline may Preliminary item selection
also affect serotoninergic neurones (Tuck and The frequency of scores above zero on all
Punell, 1973) and has pronounced anti of the 65 items as well as the ranking by mci

TABLE I
Frequencyof scoresabove zero and correlationbetween English and Swedish raters on the 17 most commonly scoreditems of the
CPRS

of scores above zero reliability between

in per cent (both samples English and Swedish rater
ItemFrequency
11)Reported combined, n = 106)Interrater (n =

sadness1000.96Apparent
sadness
Lassitude99
0.78Inner 960.90
tension950.90Pessimistic
thoughts940.53Inability
feel930.41Worrying
to
trifles910.78Concentration
over
difficulties880.80Observed
tension880.74Suicidal
muscular
thoughts870.71Fatiguability860.89Reduced

sleep830.76Indecision
.810.27Reported
disturbances790.72Reported
autonomic
tension740.59Reduced
muscular
appetite
Agitation73 730.95 0.62
384 A NEW DEPRESSION SCALE DESIGNED TO BE SENSITIVE TO CHANGE

dence were remarkably similar in the Swedish Selecting itemsfor sensitivity

and the English samples (r = +.88, rho = Two different estimates of sensitivity were
+.88, P for both < 0.001 (Montgomery et al, used (Table II). Firstly ofthe mean changes
l978a). Since the agreement between ratings of scores (absolute values) on each of the 17 items
English patients by English and Swedish raters after four weeks of treatment were calculated
was generally high (Table I) it was decided to and ranked. The second estimate was the
merge the two samples in the further cal correlation between change on each item and
culations. the overall change on the preliminary 17 item
An arbitrary cut-off point of 70 per cent scale over the four weeks. These estimates reflect
occurrence was used to identify the 17 most different aspects of sensitivity to change.
common items (Table I) in the total sample of Ideally, an item should both yield large changes
patients.The sum of scoreson these 17 items (that can be reliably rated) and be strongly
was used as a preliminaryestimateof severity correlated to general amelioration of depression.
of the illness. This estimate was significantlyThis is not always the case.
correlatedwith the HRS scores (r = +.94, For example, reduced sexual interest yielded
P <0.001) and also with the global scores large changes but was less well correlated to
(r = +.89, P <0.001) during the fourth general outcome. Inclusion of an item like this
treatment week. Before treatment these corre in a scale might spuriously inflate the change
lations were slightly lower as would be expected scores.
from the more restricted range of scores(HRS, The summed ranks on both estimates were
r = +.73, Global, r = +.6l, P for both used to select the 10 most sensitive items for the
<0.001). final depression rating scale, which is shown in
the Appendix.
TABLE II
Mean change (regardless of direction), and correlation Reliability and Validity of the New
between change on each item and sum of change on the Depression Scale
preliminary 17 item scale (n = 64). The items are ranked
in order of descending sensitivity on both estimates combined Interrater reliability
Data from the conjoint interviews were used
CorrelationItemMean to compute interrater correlations. Comparisons
total
changeApparentsadness changeto between two English raters, two Swedish raters
and one English and one Swedish rater, rating
English patients, are given in Table III. The
0.73Inability
Reportedsadness2.12 2.190.84 interrater reliability in the simultaneously
.980.75Innertension1.790.73Suicidal
to feel1 performed HRS ratings is shown for comparison.
The interrater correlations are satisfactory for
thoughts2.030.64Lassitude1.610.79Concentration
both scales, for raw scores as well as for differ
ence scores.
.790.69Reduceddifficulties1 To test the robustness of the instrument in the
sleep hands of untrained raters, ratings were also
Reduced appetite2.45
0.63Pessimistic 1 .780.49 performed by a trained psychiatrist and a
.690.64Worrying
thoughts1
general practitioner or a psychiatric nurse (a
overtrifles1
.880.46Fatiguability1
detailed account of the procedure is given
.730.58Muscular elsewhere; Montgomery ci al, 1978). Also in this
.410.56Muscular
tension (reported)1 setting the interrater correlations were high
tension(observed)1
.220.49Indecision1 (Table III).
.570.44Autonomic
Validity studies
disturbances1
.600.40Agitation0.990.34
To test the validity of a scale which estimates
the severity of depression a comparison must
 STUART A. MONTGOMERY AND MARIE ASBERG 385
TABLE III
Interrater reliability of the new depression scale with djfferent rater pairs. Interrater correlation between simultaneously
performed Hamilton scale ratings are shown within bracketsfor comparison.All correlationsarehighly sign(ficant (P <0.001)

pairrnTwoRater

Englishraters
beforetreatment 0.89) 25)
during treatment 0.95 (HRS: 0.98) 20 (HRS: 14)
(HRS:l2)Two
difference0.89 (HRS:0.9l)30
0.90(HRS: 13(HRS:

Swedish raters
treatment0.9522One
before

English, one Swedish rater

treatment0.9711Psychiatrist
various stages of

and general practitioner

treatment0.9717Psychiatrist
various stages of

and nurse
various stages of treatment0.9312

be made with an independent measure. An level with the HRS, compared with 19 patients
experienced clinician's global judgement as to on the 10 item scale.
whether the patient has responded or not is the
criterion against which depression scales should Discussion
be judged. As a preliminary validation of this The major requirements of a rating scale for
scale's capacity to identify responders and antidepressant treatment effects is that it should
non-respondersto treatment we compared the be short and easy to apply in a clinical setting,
scale scores with a clinician's global judgement relevant for depressiveillness, and provide a
in a sample where there was a clear cut differ sensitiveand accurate estimate of change
entiation between responders (18 patients) and (Hamilton, 1976; Carroll et al, 1973; Asberg
non-responders (17 patients). etal, 1973).
Point biserial correlations between response It would of course be possibleto use a very
category and change scores were calculated for extensive rating scale covering all aspects of
the preliminary 17-item scale, the final 10-item depressive illness. A scale of this type might be
depression scale and the Hamilton Rating more likely to pick up unexpected differences in
Scale. All correlations were highly significant. the spectrum of action of drugs. However, the
The correlations were used to determine which presence of a large number of items that were
of the three scales differentiated better between scored in only a few patients would tend to
responders and non-responders to treatment. introduce and increase the random error. More
The best correlation was achieved with the important, the ratings would be cumbersome
10-item depression scale (r = +.70), the next and time-consuming to undertake. Unskilled
best was with the 17-item preliminary scale raters might have difficulties in covering a
(r = +.67). Of the three the HRS was the large number of items in a single interview.
least able to discriminate (r = +.59). Converted Repeated asking of questions which appear
into approximate patient numbers needed to irrelevant to the patient might also be detri
achieve equivalent significance at different mental to clinical rapport and reduce the
levels, for a point biserial correlation of this validity of the information provided.
order 28 patients would have been needed to When reducing the number of items, it is
reach a significance at the 0.00 1 probability important that those included are relevant to
386 A NEW DEPRESSION SCALE DESIGNED TO BE SENSITIVE TO CHANGE

the illness and indeed occur in the majority of and the correlation of change to overall amelior
cases. An item may be both of diagnostic ation was comparatively low (r = +.57).
importance and likely to change with treatment The large number of rating scales available
but because it occurs so infrequently it might to clinical investigators is a problem in psych
diminish the overall sensitivity of the scale. iatric research (Pichot, 1972) and the com
Examples of this type which failed to meet our parability between scales is rarely known. It is
frequency criterion for consideration are Ideas therefore important that a new rating scale
of Persecution and Compulsive Thoughts. Simi should be shown to have clear advantages over
lar items were included in the initial version of existing instruments before it is accepted for
the Hamilton Rating Scale (Hamilton, 1960) research purposes. Our scale appears to have
but excluded later (Hamilton, 1967). certain advantages, even when compared to the
Sensitivity of the scale in this context refers most widely used depression rating scale, the
to its capacity to measure change. Change will Hamilton Rating Scale.
be most sensitively recorded on items which are Our depression scale has fewer items to score
not restricted in range, but when the full width than the HRS (10 vs 17) and a reduction in
of the scales are used for the ratings. Restriction reliability might be expected. However, we have
ofthe range might occur ifa positive score on an demonstrated equally high reliabilities with the
item reflects a personality trait which is unlikely HRS and the 10 item scale. The scale can be
to change with short term treatment, rather than used by trained nurses and psychologists as well
a symptom of an illness. It might also occur as a as by psychiatrists (Montgomery et al, 1978b).
result of central tendency error, in which raters It appears to be a more precise measure of
tend to avoid using the extreme ends of the change than the HRS. This means that sig
scales (Guilford, 1954). nificant differences between treatments may be
revealed with smaller numbers of patients. In
Change estimates should also be accurate clinical trials this is important for ethical
and reflecta change in general severityof reasons, since fewer patients need to be exposed
depressive illness. Some items may show a high to possibly inferior treatments.
degree of change without this being related to There is a definite need for simple, clinically
the illness as such. Hospitalization would be oriented rating scales to measure treatment
expected to have effects, for instance, on sexual effects in other psychiatric syndromes such as
interest and general levels of activity. Increased mania, schizophrenia and anxiety states. To our
severity of some symptoms (for instance auto knowledge, the strategy we employed for
nomic disturbances) may be side-effects of an arriving at an empirically founded scale has not
antidepressant drug treatment. previously been used and should prove valuable
The 10 items included in the new depression in constructing further scales.
scale are all core symptoms of depressive illness.
A few characteristic symptoms are, however, not Acknowledgement
This study was supported by the Swedish Medical
included. Motor retardation (called Slowness of
Research Council, 21P—4676; 27X—05015, the Bank of
Movement in the full CPRS) is perhaps the Sweden Tercentenary Fund, 74/110, the Gurli Wehtje
most conspicuous omission. It was excluded from Fund for Depression Research, S.E. Thames Regional
the primary selection, since it occurred in Health Authority, Grant No. LOR/76/24, and Guy's
relatively few patients (63 per cent of the Hospital Medical School.
English patients and 69 per cent of the Swedish References
patients). Clinical experience shows that motor ANDERSON, N. H. (1961) Scales and statistics: parametric
retardation only occurs in a proportion of and non-parametric. PsychologicalBulletin, 58, 305—16.
patientsand thisis indeed the background of ANGST, J. (1972) Concluding discussion. In Depressive
one of the classifications of depressive illness into Illness (ed. P. Kielholz). Bern: Huber.
ASBERG, M., KRAGH-SORENSEN, P., MINDHAM, R. H. S. &
retarded and non-retarded forms. In any case
TUCK, J. R. (1973) International reliability and
the mean change in Slowness of Movement was communicability of a rating scale for depression.
smallerthan forany item included in the scale Psychological Medicine, 3,458-65.
 STUART A. MONTGOMERY AND MARIE ASBERG 387
—¿ RINGBERGER, V. A., SJOQvI5T, F., THORN, P., —¿ —¿ JORNESTEDT, L., THOREN, P., TRASKMAN, L.,

TRASKMAN, L. & TUCK, J. R. (1977) Monoamine MCAULEY, R., MONTGOMERY, D. & SHAw, P. (l978b)
metabolites in the cerebrospinal fluid and serotonin Reliability of the CPRS between the disciplines of
uptake inhibition during treatment with clomi psychiatry, general practice, nursing and psychology.
pramine. Clinical Pharmacology and Therapeutics, 21, AcM Psychiatrica Scandinavica. Supplement 272, 29—32.
201—7. MORRIS, J. B. & BECK, A. T. (1974) The efficacy of
MONTGOMERY, S., PERRI5, C., SCHALLING, D. & antidepressant drugs—a review of research ( 1958 to
SEDVALL, G. (1978) A comprehensive Psycho 1972) Archives of General Psychiatry, 30,667—74.
pathological Rating Scale. Acta Psychiatrica Scandi PICHOT, P. (1972) The problem of quantifying the
navica, Supplement 272. symptomatology of depression. In Depressive lilness
BONEAU, C. A. (1961) A note on measurement scales and
(ed. P. Kielholz). Bern: Huber.
SILVERSTONE, T. & TURNER, P. (1974) Drug Treatment in
statistical tests. American Psychologist, 16, 160—261.
Psychiatry,p 117. London: Routledge & Kegan Paul.
CARROLL, B. J., FIELDING,J. M. & BLASHKI,T. G. (1973)
TUCK, J. R. & PUNELL, G. (1973) Uptake of 3H-5-
Depression rating scales. A critical review. Archives of
hydroxytryptamine and 3H-noradrenaline by slices of
GeneralPsychiatry, 28, 361—6.
rat brain incubated in plasma from patients treated
COPPEN, A., GUPTA, R., MONTGOMERY, S., GHOSE, K., with clomipramine, imipramine or amitriptyline.
BAILEY, J., BURNS, B. & DE RIDDER, J. J. (1976) Journal of PharmacyandPharmacology,25, 573-4.
Mianserin hydrochloride : a novel antidepressant.
BritishjournalofPsychiatry, 129, 342—5.
FEIGHNER, J. P., ROBINS, E., GUZE, S. B., WOODRUFF, R.
A., WINOKUR, G. & MUNOZ, R. (1972) Diagnostic
APPENDIX
criteria for use in psychiatric research. Archives of Montgomery and Asberg (MADRS) Depression
General Psychiatry, 26, 57—63.
Rating Scale.
FERL, P. J., QUANTOCK, D. C. & VAN DER BURG, W. J.
(1973) The human pharmacology of GB94—a new The ratingshouldbe basedon a clinical
interview
psychotropic agent. European Journal of Clinical moving from broadly phrased questions about
Phar@nacology,5, 166—73. symptoms to more detailed ones which allow a precise
GUILFORD, J. P. (1954) Psychometric methods. 2nd Ed., rating of severity. The rater must decide whether the
p278. New York: McGraw-Hill. rating lies on the defined scale steps (0, 2, 4, 6) or
GURNEY, C., ROTH, M., GAR5IDE, R. F., KERR, T. A. & betweenthem (1,3,5).
SCHAPIRA, K. (1972) Studies in the classification of It is important to remember that it is only on rare
affective disorders. The relationship between anxiety occasions that a depressed patient is encountered who
states and depressive illnesses—I!. British Journal of
Psychiatry, 121, 162—6.
cannot be rated on the items in the scale. If definite
answers cannot be elicited from the patientall
HAMBERGER, B. & TUCK, J. R. (1973) Effect of tricyclic
antidepressants on the uptake of noradrenaline and
relevant clues as well as information from other
5-hydroxytryptamine by rat brain slices incubated in sources should be used as a basis for the rating in line
buffer or human plasma. European Journal of Clinical withcustomaryclinical practice.
Pharmacology, 5,229—35. The scale may be used for any time interval
HAMILTON, M. (1960) A rating scale for depression. between ratings, be it weekly or otherwise but this
Journal of Neurology, Neurosurgery and Psychiatry, 23, must be recorded.
56-62.
—¿ (1967) Development of a rating scale for primary Item List
depressive illness. British Journal of Social and Clinical
1. Apparent sadness
Psychology, 6,278-96.
2. Reportedsadness
—¿ (1976) Comparative value of rating scales. British
3. Inner tension
Journal of Clinical Pharmacology, 3,58—60.
4. Reduced sleep
MAss, J. W. (1975) Biogenic amines and depression.
Biochemical and pharmacological separation of two 5. Reduced appetite
types of depression. Archives of General Psychiatry, 32, 6. Concentration difficulties
1357—61. 7. Lassitude
MAITRE, L., STAEHELIN, M. & BEIN, H.J. (1971) Blockade 8. Inability to feel
of noradrenaline uptake by 34276 B a, A new anti 9. Pessimistic thoughts
depressant drug. Biochemical Pharmacology, 20, 2169—86. 10. Suicidal thoughts
MONTGOMERY, S., ASBERG, M., TRASKMAN, L. & MONT
GOMERY, D. (1978a) Cross-cultural studies on the use
1. Apparent Sadness
of the CPRS in English and Swedish depressed
patients. Acta Psychiatrica Scandinavica, Supplement Representing despondency, gloom and despair,
272, 33—9. (more than just ordinary transient low spirits)
388 A NEW DEPRESSION SCALE DESIGNED TO BE SENSITIVE TO CHANGE

reflected in speech, facial expression, and posture. 4. Reducedsleep

Rate by depth and inability to brighten up. Representing the experience of reduced duration or
0 No sadness. depth of sleep compared to the subject's own normal
pattern when well.
2 Looks dispirited but does brighten up 0 Sleeps as usual.
without difficulty.
3 2 Slight difficulty dropping off to sleep or
4 Appears sad and unhappy most of the time. slightly reduced, light or fitful sleep.
D
3
4 Sleep reduced or broken by at least two
6 Looks miserable all the time. Extremely
hours.
despondent.
5
2. Reported sadness 6 Less than two or three hours sleep.
Representing reports of depressed mood, regardless
5. Reducedappetite
of whether it is reflected in appearance or not.
Includes low spirits, despondency or the feeling of Representing the feeling of a loss of appetite
being beyond help and without hope. compared with when well. Rate by loss of desire for
Rate according to intensity, duration and the food or the need to force oneself to eat.
extent to which the mood is reported to be influenced 0 Normal or increased appetite.
by events.
0 Occasional sadness in keeping with the 2 Slightly reduced appetite.
circumstances. 3
4 No appetite. Food is tasteless.
2 Sad or low but brightens up without difficulty. 5
3 6 Needs persuasion to eat at all.
4 Pervasive feelings of sadness or gloominess.
The mood is still influenced by external 6. ConcentrationdU/Iculties
circumstances. Representing difficulties in collecting one's thoughts
5 mounting to incapacitating lack of concentration.
6 Continuous or unvarying sadness, misery or Rate according to intensity, frequency, and degree
despondency. of incapacity produced.
0 No difficulties in concentrating.
3. Inner tension
Representing feelings of ill-defined discomfort, 2 Occasional difficulties in collecting one's
edginess, inner turmoil, mental tension mounting to thoughts.
either panic, dread or anguish. 3
Rate according to intensity, frequency, duration 4 Difficulties in concentrating and sustaining
and the extent of reassurance called for. thought which reduces ability to read or hold
0 Placid. Only fleeting inner tension. a conversation.
5
2 Occasional feelings of edginess and ill 6 Unable to read or converse without great
defined discomfort. difficulty.
3
4 Continuous feelings of inner tension or 7. Lassitude
intermittent panic which the patient can only Representing a difficulty getting started or slowness
master with some difficulty. initiating and performing everyday activities.
5 0 Hardly any difficulty in getting started. No
6 Unrelenting dread or anguish. Overwhelming sluggishness.
panic.
2 Difficulties in starting activities.
 STUART A. MONTGOMERY AND MARIE ASBERG 389
3
4 Difficulties in starting simple routine activities 2 Fluctuating ideas of failure, self-reproach or
whicharecarried outwitheffort. selfdepreciation.
5 3
6 Complete lassitude. Unable to do anything 4 Persistent self-accusations, or definite but
without help. still rational ideas of guilt or sin. Increasingly
pessimistic about the future.
8. Inability tofeel 5
Representing the subjective experience of reduced 6 Delusions of ruin, remorse or unredeemable
interest in the surroundings, or activities that norm sin.Self-accusations which are absurd and
ally give pleasure. The ability to react with adequate unshakable.
emotion to circumstances or people is reduced.
0 Normal interest in the surroundings and in 10. Suicidalthoughts
other people. Representing the feeling that life is not worth
living, that a natural death would be welcome,
2 Reduced ability to enjoy usual interests. suicidal thoughts, and preparations for suicide.
3 Suicidal attempts should not in themselves
4 Lossof interest
in thesurroundings.Lossof influence the rating.
feelings
forfriendsand acquaintances. 0 Enjoys life or takes it as it comes.
5
6 The experience of being emotionally para 2 Weary of life. Only fleeting suicidal thoughts.
lysed, inability to feel anger, grief or pleasure 3
and a complete or even painful failure to feel 4 Probably better off dead. Suicidal thoughts
forclose relativesand friends. are common, and suicide is considered as a
possible solution, but without specific plans
o)@ Pessimistic thoughts or intention.
Representingthoughtsof guilt,inferiority,
self 5
reproach, sinfulness, remorse and ruin. 6 Explicit plans for suicide when there is an
0 No pessimistic thoughts. opportunity. Activepreparationsforsuicide.

* Stuart A. Montgomery, B.Sc., M.D.. M.R.C.Psych., Senior Lecturer, Academic Department of Psychiatry,
Guy's Hospital Medical School, London, S.E. 1,

Marie Asberg, M.D.,Karolinska Institute, Stockholm, Sweden

* Correspondence.

(Received 24 April; revised 30 August 1978)