ARTICLE IN PRESS

0277-9536/$ - se

doi:10.1016/j.so

�Correspondfax: +4420 743

E-mail addr

Social Science & Medicine 62 (2006) 50–58

www.elsevier.com/locate/socscimed

Social determinants of diagnostic labels in depression

Susan McPherson�, David Armstrong

Tavistock & Portman NHS Trust, Belsize Centre, 94 Belsize Lane, London NW3 5NE, UK

Available online 11 July 2005

Abstract

The role of diagnostic labels in medicine is usually that of labelling an illness as a means of communication.

Control over labelling processes in medicine is ordinarily imposed via medical schools, textbooks, education or by

diagnostic manuals. Diagnostic labels often change following new discoveries in underlying pathology such as

‘consumption’ being relabelled as ‘TB’ or ‘cancer’. Sub-types of broad diagnostic labels also often emerge from such

discoveries e.g. ‘lung cancer’ or ‘throat cancer’. In mental health, underlying pathology is the subject of ongoing debate

spanning ideas including the brain as a faulty organ, faulty genetics and environmental problems. With controversy

over pathology comes controversy over labels and the idea that labels may be used not just for communication, but as

devices of social and professional control, arising out of a social process. This study explores the codification of the

diagnostic label ‘depression’ which emerged in the twentieth-century and has proliferated with numerous sub-types over

the last 40 years. The aim is to examine its social determinants and context. Medline is used as a data source for

professional label usage. A range of depression sub-type labels in professional use was identified. This exercise revealed

many official and ‘unofficial’ terms in professional use. Citation rate plots by year were then generated for these

depression sub-type labels. The rise and fall of different labels are examined in relation to social determinants and

context, including publication of diagnostic manuals DSM and ICD, power shifts in psychiatry, the discovery of

psychiatric drugs and the shift from inpatient to community care. Exploring the changing use of official and unofficial

labels over time in this way provides a novel historical perspective on the concept of depression in the late twentieth-

century.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Depression; History of medicine; Diagnostic labels; DSM; Community care

Introduction

Diagnosis in medicine involves a process of classifying

and attaching a label to an illness. Over the last two

centuries medical classification has increasingly been

based on the underlying pathological lesion or process

as, given this physical or organic referent, it becomes

easier to gain professional consensus about specific

e front matter r 2005 Elsevier Ltd. All rights reserve

cscimed.2005.05.021

ing author. Tel.: +4420 77945875;

5 8018.

ess: susan@mcpherson.org.uk (S. McPherson).

illnesses. What is pneumonia or appendicitis or cancer

can be agreed internationally with reference to the

presence or absence of certain clearly defined physical

characteristics. In psychiatry, however, there is no such

external biological referent to act as an anchor for

diagnosis. Essentially, psychiatry classifies on the basis

of patients’ patterns of symptoms which might vary

according to how they are elicited and interpreted.

The development of standardised psychiatric instru-

ments and interviews over the last few decades has

helped achieve some stability for researchers investigat-

ing mental illness but clinicians, historically, have had

d.

ARTICLE IN PRESSS. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–58 51

little constraint on their use of diagnostic labels. Indeed,

cross-national studies have shown the considerable

differences in diagnostic patterns in different countries

(Lee, 2002). There is also some evidence of differences

between individual psychiatrists within the same country

(Fabrega, Ahn, Boster, & Mezzich, 1990). Managing

this diagnostic heterogeneity has therefore been a major

task for maintaining the identity and credibility of

psychiatry: one of the main strategic approaches has

been to codify classification decisions in the form of

diagnostic manuals.

Diagnostic manuals came to the fore in medicine

during the twentieth-century when international com-

parisons of death rates became important. The Interna-

tional Classification of Disease (ICD) allowed some

standardisation of causes of death such that rates could

reliably be compared across countries and across time.

The success of the early ICD meant that subsequent

revisions gradually incorporated illnesses that did not

necessarily cause death, including mental illness. By the

sixth revision of 1948, guidance on diagnosing psychia-

tric illness was fully incorporated. In 1952 the American

Psychiatric Association (APA) produced their own

version of this guidance as the diagnostic and statistical

manual of mental disorders (DSM). Since then the two

systems have remained in parallel, each publishing

revisions every few years.

One of the aims of DSM and ICD was to provide a

language for communication (or control) both within

psychiatry as well as for other professionals dealing with

mental illness such as psychologists, primary care

workers, social workers and occupational therapists

(see Frances, First, Pincus, Widiger, & Davis, 1990).

However, labels given by these diagnostic manuals do

not in themselves dictate the range of terms employed by

such professionals and they have no legal status other

than for reporting national morbidity statistics to the

World Health Organisation (WHO) and for reimburse-

ment for health insurance in the USA and some other

countries with health insurance systems. ‘Unofficial’

labels are used by professionals both outside and within

psychiatry for purposes of intra- or inter-professional

communication. This was demonstrated in a study by

Pincus, Davis, and McQueen (1999) which aimed to find

out what terms and definitions were being used in

professional literature for ‘subthreshold depression’. A

literature review was carried out using the term

‘subthreshold depression’ as well as a variety of other

similar terms known to be relatively synonymous. The

study identified a wide variety of terms and definitions

employed in the published professional literature refer-

ring to depression that was ‘not clinically significant’.

These terms included some used in DSM and ICD as

well as other terms not found in either system. If a

variety of different labels exist for one such area of

depression, it is more than likely that a wide range of

sub-type labels exist for the broader totality of

depression.

The diagnosis of illnesses such as depression must

constitute a constant struggle between the standardising

forces—as represented by the manual authors—and the

idiosyncrasies of the labelling systems used by individual

clinicians. This tension between standardisation and

heterogeneity is the subject of this study that examines

the social determinants and context of how depression

became codified over the last 40 years. Given the lack of

reference to underlying pathology, these codes are likely

to be the product of a social process that includes the

publication and revisions of diagnostic manuals them-

selves, mental health policy such as the shift from

inpatient to community care, the discovery of successive

generations of psychiatric drugs and the power struggles

in psychiatry between biological psychiatry (i.e., those

who stressed the biological or organic causes of mental

illness) and psychoanalytic-oriented psychiatrists. These

will be the main social contexts explored in relation to

the codification of depression.

Methods

Medline, which electronically indexes articles from

1966 onwards, was used as the main source of data on

the use of psychiatric diagnostic terms for depression.

The database was first searched using the search term

‘depression.mp’ which identified articles where ‘depres-

sion’ featured in the title, keywords or abstract. The data

set was downloaded and cleaned in reference manager

(including the removal of duplicates) and imported into

SPSS. SPSS was then used to generate a random sample

of 50 articles for every fifth year in the data set. This

resulted in 350 articles, the titles and abstracts of which

were then scanned to identify any depression sub-type

labels used.

In order to explore further the range of sub-type labels

identified, the labels were divided into clusters which

represented the underlying type of illness classification.

Clusters were identified by an inductive process of

grouping sub-type labels in terms of an implicit under-

standing of their meaning (given that definitions were

not considered systematically in this process as these

may vary and change over time and are outside the

scope of this study). These clusters were agreed by both

authors. Bipolar disorder and manic depression were

excluded from this study in order to focus on the

concepts and labels for low mood rather than mania.

Each sub-type label identified from the search of a

sample of articles was then used to perform a Medline

search across the whole database to identify articles

where the term features in the title, keywords or

abstract. Each data set was then downloaded into

reference manager.

ARTICLE IN PRESSS. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–5852

Once each data set was cleaned, it was exported to

SPSS to allow easier manipulation of data. SPSS was

used to count the number of citations for each search

term by year. This figure was converted into a citation

rate by dividing the number of citations for the search

term by the number of citations for depression in that

year. This was multiplied by 10,000 to get whole

numbers. The citation rate for each term was graphed

over time with the number of ‘depression’ citations

plotted on a second axis.

Results and discussion

Changing patterns of diagnostic labelling reflect a

number of contextual events. The main events explored

in this paper are the publication dates of diagnostic

manuals, namely DSM II in 1968, DSM III in 1980 and

DSM IV in 1994. But the successive editions of DSM

were themselves related to shifts in psychiatry itself. In

particular, there was a major power shift in the APA

that occurred around the time of preparation and

publication of DSM III during which the psychoanalytic

psychiatrists lost political ground to biological psychia-

trists, e.g. Grob (1991), Bayer and Spitzer (1985) and

Havens (1985). Also, developments in the use of

psychiatric drugs, illustrated by key shifts from mono-

amine oxidase inhibitors (MOAIs) falling out of grace

following reports of their severe hepatotoxic and

hypertensive effects through the domination of the

tricyclics, to the introduction of the very successful

selective serotonin reuptake inhibitors (SSRIs) (Ban,

2001), are likely to have had an effect on depression

diagnosis. This period also witnessed a significant shift

from inpatient to community care as well as a number of

key governmental and legal landmarks (Geller, 2000),

which are summarised in Fig. 1.

Depression sub-type labels identified and used in this

study are listed in Table 1. Labels were also clustered

into categories which refer to an implicit understanding

of their meaning.

Table 2 shows information for each sub-type label.

The first column gives the year that the label first

Fig. 1. Social context timeline. 1Mental Retardation Facilities and Co

88–164) passes; it contains funds for constructing community mental

appeals rules that an individual could not be committed to the hospita

facility available to care for her. 3A U.S. district court judge in the Dist

since the District’s Hospitalization of the Mentally Ill Act becam

controversial throughout the rest of the century. 4Congress overrides P

Sharing on Health Services Act. The bill includes expanded funding for

must provide. 5The National Institute of Mental Health awards con

Support Program. 6Surgeon General Julius B. Richmond, M.D., rele

Chronically Mentally Ill. 7The State Comprehensive Mental Health Se

and implement comprehensive mental health plans for community-ba

appeared in Medline (i.e. the citation rate was greater

than zero for the first time). The peak rate column gives

the highest citation rate for each label and the year in

which this citation rate occurred. Finally, the table

shows the citation rate in 2000. As can be seen from

Table 2, peak citation rates range from 9.05 (‘exogenous

depression’) to 1097.5 (‘major depression’). However, in

spite of this very large range, the majority of labels

cluster at the bottom of this range. The three most

dominant labels are ‘major depression’, ‘endogenous

depression’ and ‘dysthymia’. Fig. 2 plots citation rates

for each of the three dominant sub-type labels on the

left-hand (primary) axis. The right-hand (secondary)

axis shows the total number of ‘depression’ citations.

This line can be seen to be rising steadily over the period

under consideration. The most prominent feature of

Fig. 2 is that ‘major depression’ can be seen to increase

very steeply from zero around 1980 and to dominate the

plot thereafter.

The term major depression was first introduced in

1980 with DSM III. Since this term was not found in any

indexed articles before DSM III yet became very

common afterwards, it is reasonable to conclude that

this significant increase in the use of the term was at least

in part a product of the publication of DSM III. The

parallel ICD system did not use the term major

depression, suggesting a particular influence for DSM.

The APA had a much greater budget for producing,

disseminating and training for the DSM than the WHO

had for the ICD. This may have therefore made the

DSM more accessible, certainly to American audiences

if not further afield with global sales of video training

materials. Furthermore, while the ICD eventually made

headway in producing operationalised versions of its

manuals, the DSM made faster progress by incorporat-

ing these various versions into training for relevant

professionals. Perhaps at an even more practical level,

published medical literature tends to be dominated in

terms of volume and impact by US journals. North

American journals, while open to authors from around

the world, may also have an inherent publication bias in

favour of US authors using DSM language, in that the

language is familiar and the writing style will conform

mmunity Mental Health Centers Construction Act of 1963 (P.L.

health centers (CMHCs). 2In Lake vs. Cameron a U.S. court of

l until hospital officials determined there was no less restrictive

rict of Columbia orders an outpatient commitment, the first ever

e effective in 1964. Outpatient commitment would remain

resident Ford’s veto of the Nurse Training and Health Revenue

CMHCs and increases the number of essential services CMHCs

tracts to 16 states under a new project called the Community

ases a 457-page report entitled Toward a National Plan for the

rvices Plan Act of 1986 (P.L. 99–660) requires states to develop

sed services for people with severe mental illness.

ARTICLE IN PRESS

Manuals Power struggles

Drugs Community Care (USA) Psychiatric

inpatients(USA)

1963 MOAI CMHCs established1

1966scandals Lake vs Cameron2

452,000

1967

1968 DSM II 399,000

1969

1970 338,000

1971

1972 Outpatient commitment order3 275,000

1973

1974 216,000

1975

APA

pre

dom

inan

tly

psyc

hoan

alyt

ic

CMHC expanded funding4

1976 171,000

1977 NIMH Community Support Program5

1978 154,000

1979 ICD 9

1980

DSM III APA powerstruggles

Toward a National Plan for the Chronically Mentally Ill 6

1981

Lith

ium

fir

st u

sed

to tr

eat

man

ic d

epre

ssio

n

1982

1983 Tri

cycl

ics

dom

inat

e m

arke

t

1984

1985

1986

State Comprehensive Mental Health

Services Plan Act 7 111,000

1987 DSM IIIR

1988

1989

1990 91,000

1991

1992 ICD 10 83,000

1993

APA

P pr

edom

inan

tlybi

olog

ical

1994 DSM IV 72,000

1995

1996 62,000

1997

1998

SSR

Is in

trod

uced

1999

2000 DSM IVTR

↑ ↑

Psychiatric Psychiatry

138,000

↓ ↓

↓ ↓

↓ ↓

S. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–58 53

ARTICLE IN PRESS

Table 1

Sub-type labels and clusters

Cluster Label

Causal Endogenous depression

(relating to the aetiology

of depression)

Exogenous depression

Reactive depression

Temporal Postnatal depression

(relating to the timing

or duration of illness)

Recurrent depression

Chronic depression

Acute depression

Dysthymia

Severity of symptoms Major depression

(relating to the severity

of depression)

Minor depression

Mild depression

Severe depression

Nature of symptoms Atypical depression

(relating to the nature

of symptoms)

Melancholic depression

Psychotic depression

Neurotic depression

Melancholia

Treatment response Treatment resistant depression

(relating to the patient’s

response to treatment)

Refractory depression

S. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–5854

more often to the US method of scientific reporting.

Once the DSM discourse had established a hegemony

within the USA, it would have been able to spread as an

international discourse, with an ever-increasing trend to

use language that can be globally understood.

The dramatic success of ‘major depression’ is not

matched by any other sub-type label introduced by other

versions of DSM or ICD. Given that the DSM is written

by and controlled by the APA, the success of ‘major

depression’ as a label can also be seen as a success for the

profession of psychiatry in controlling language use

around this diagnosis. The APA was given authority by

the US government to devise a classification system for

mental disorders that would be used for official national

statistics and financial processes in the health system.

The explicit authority over mental health labelling given

to psychiatrists by this process is one aspect of the social

context in which the development of the manual took

place.

The coining of the term ‘major depression’ may also

be considered in the context of the earlier asylum

closures and the ongoing development of community

care, as illustrated in Fig. 1. These developments posed

both an opportunity and a threat for the profession of

psychiatry. In one sense, community care posed a threat

to psychiatry in that the responsibility of care for the

mentally ill could become dispersed among other

professions such as primary care physicians, nurses

and new mental health professions such as community-

based psychiatric nurses. In another sense, community

care provided an opportunity for the mental health field

as a whole to expand and to create more work for

psychiatrists as mental illness was construed as a

continuum, from the severely psychotic to the general

worries and stresses of everyday life. With this dual

threat and opportunity posed by community care,

psychiatry may have needed to set boundaries around

areas of mental health that psychiatry would own or

control and to keep them apart from other professions

involved in mental health. In terms of depression, this

may have meant creating categories of psychiatric and

non-psychiatric depression—hence ‘major depression’

which might be considered severe enough to require

psychiatric input and, by implication, ‘minor depression’

to be dealt with by other professionals. Supporting this

interpretation, ‘minor depression’ is present in DSM III,

but is ill-defined, appearing within the ‘Not Otherwise

Specified’ category; it was considered by DSM authors

to be within the remit of primary care rather than

psychiatry (Pincus et al., 1999).

The terms major and minor depression fall into the

‘severity of symptoms’ cluster set out in Table 1. These

clusters, which were derived from labels identified in

Medline from 1966 to 2000, might usefully be compared

to previous clusters or dimensions of mental illness in

order to establish whether ‘severity’ is indeed a new

concept in mental illness. The eight criteria that were

used to classify mental disorder by the nineteenth-

century French alienists provide a useful comparison.

According to Berrios (1999), these were the cause of the

disorder (aetiological), the substratum of the disorder

(e.g. anatomy), the clinical outcome (whether curable or

incurable), actuarial (according to what was observed in

the statistics of the main asylums in France), phenom-

enological (according to whether or not the disorder

included delusions), ‘natural’ (i.e. whether it corre-

sponded to ‘real types’ as given in nature), psychological

(i.e. what mental faculty was assumed to be impaired)

and disease course. There are some clear similarities with

the diagnostic clusters identified in this paper. The

aetiological criterion is identical to the causal cluster; the

clinical outcome criterion similar to the treatment

response cluster; the phenomenological criterion similar

to the nature of symptoms cluster and the disease course

criterion similar to the temporal cluster. The only cluster

identified in the present study that does not have a

counterpart in these eight criteria is the ‘Severity of

symptoms’ category, which could therefore be seen as

something new to twentieth-century psychiatric think-

ing. This may suggest that the domain of psychiatry has

ARTICLE IN PRESS

Table 2

Sub-type labels: appearance in Medline, peak citation rate and year 2000 citation rate

Label Appearance year Peak year Citation rate 2000

Year Citation rate

Endogenous depression 1966 1982 300.26 15.79

Reactive depression 1966 1981 58.85 3.51

Chronic depression 1966 1996 50.78 49.12

Dysthymia 1966 1994 198.25 135.09

Mild depression 1966 1987 57.23 35.09

Psychotic depression 1966 1999 66.06 36.84

Neurotic depression 1966 1979 84.78 8.77

Melancholia 1966 1967 149.05 31.58

Postnatal depression 1968 1997 70.41 63.16

Recurrent depression 1968 1994 51.92 29.82

Severe depression 1968 1999 160.44 119.3

Atypical depression 1968 1984 45.2 14.04

Acute depression 1970 1999 28.31 12.28

Treatment resistant depression 1974 1996 44.16 28.07

Minor depression 1975 1997 55.91 35.09

Exogenous depression 1978 1981 9.05 0

Refractory depression 1978 1996 35.33 21.05

Major depression 1979 1997 1097.54 985.96

Melancholic depression 1981 1997 37.27 14.04

0

200

400

600

800

1000

1966

1967

1968

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Cit

atio

n R

ate

0

1000

2000

3000

4000

5000

6000

All

Dep

ress

ion

Cit

atio

ns

Dysthymia Endogenous Depression Major Depression All Depression Citations (2nd Axis)

All

othe

r la

bels

belo

w t

his

line

Fig. 2. Citation rate of dominant depression sub-type labels 1966–2000.

S. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–58 55

broadened since the nineteenth-century, when psychia-

trists’ work was largely restricted to asylums where

psychoses were usually present. Twentieth-century

psychiatry may be seen to require the new dimension

of ‘severity’ in order to reflect its expansion out of

asylums and into the community.

ARTICLE IN PRESSS. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–5856

The ability of the APA to control diagnostic practice

through its DSM manuals, however, was not charac-

terised by total success. On the one hand, it failed to

suppress some terms and, on the other, was not always

successful in its prescription of what terms were

acceptable. ‘Endogenous depression’, for example, was

removed from the classification system by the authors of

DSM III, yet remained in use for years later and is the

second most dominant label up until the mid-1990s (see

Fig. 2). Furthermore, ‘dysthymia’, a term which DSM

III redefined from reflecting a problem of personality to

a depressive disorder (Ramana & Paykel, 1992), did not

have anything like the success of ‘major depression’ and

did not get adopted until the late 1980s.

The term ‘endogenous depression’ refers to a depres-

sion with biological origin. Although a fashionable term

before 1980 (see Fig. 2), DSM III abandoned all causal

terminology regarding depression, claiming to be

aetiologically neutral. DSM III was essentially written

by biological psychiatrists, for whom a classification

term reflecting a biological basis might seem a natural

choice, but the very existence of the term conjured into

opposition the idea that there could also be an

alternative type of depression that did not have a

biological origin (‘exogenous’ or ‘reactive’ in the

previous classifications). It is possible that biological

psychiatry would benefit from a labelling system which

excluded the possibility of any non-biological causes.

This effect can be noted in the objections to the removal

of this dimension and to the claims by DSM III authors

of aetiological neutrality, e.g. Sadler, Hulgus, and Agich

(1994) and van Praag (1990).

The control that biological psychiatrists exerted over

DSM III was a major element of the power struggle

within the APA (indicated in Fig. 1) around the time of

its publication. Previous versions of DSM had been

written by psychoanalytically orientated psychiatrists

and contained psychoanalytic terminology. DSM III

abandoned the term ‘neurotic’ under a great deal of

controversy and much resistance from the psycho-

analytic psychiatrists (Bayer & Spitzer, 1985). ‘Neurosis’

was perceived as referring to a psychological aetiology

and DSM III authors claimed to be intent on removing

aetiological labels from the classification system. Fol-

lowing heated debates and much media attention, a

compromise was reached between Spitzer (who led work

on DSM III) and the APA, whereby the term ‘neurosis’

and other psychoanalytically related terms would appear

in brackets next to their new equivalents (Bayer &

Spitzer (1985)).

The removal of endogenous depression from the

classification system, however, did not receive anything

near as much attention, perhaps because it was a term

allied to biological psychiatry—which had no wish to

defend it. Hence, while attention was focused on the

‘neurosis’ debate, the removal of endogenous depression

may prove to have been a more subtle assertion of

biological psychiatry’s power, removing the implication

that there could be any depression that did not have a

biological origin.

Why was it important for biological psychiatrists to

champion the complete medicalisation of depression?

Physicians outside of psychiatry might have been more

inclined towards this goal than psychiatry; moreover,

the psychopharmacological advances being made at the

time were sufficient to mark out a biological aspect of

mental illness without needing to rule out the possibility

of both biological and environmental causes. In fact, the

medicalisation of depression may have been a by-

product of the power struggle with psychoanalytic

psychiatry. It may also have been important to

psychiatrists in the context of a general need for the

greater professionalisation of psychiatry. Professional

status for any occupational group was enhanced

through control of an esoteric knowledge base, in

particular that of science. The discoveries of psychoac-

tive substances around this time as well as an improved

understanding of brain morphology and metabolism

created an opportunity for psychiatry to proclaim its

biomedical scientific credentials and ally itself to the rest

of medicine. This strategy had been partly compromised

by the deaths from hepatotoxicity arising from the

widespread use of the monoamine oxidase inhibitors

(MAOIs), so by the time of DSM III when the new

(and safer) tricyclics had been relatively established

(see Fig. 1), biological psychiatry may have needed to

recover ground from the MAOI scandals and over-

emphasise the view that depression could be treated with

drugs on a par with any other medical illness.

In spite of the success of biological psychiatry, the

increasing medicalisation of depression and publication

of DSM III, it is nevertheless notable that ‘endogenous

depression’ did not disappear from professional dis-

course. ‘Endogenous depression’ remained the second

most dominant term overall, even after 1980, despite a

gradual decline over subsequent years. It continued to

be used to some degree up to 2000 (see Fig. 2). This may

be testament to its great popularity before the publica-

tion of DSM III and the reluctance to abandon a term,

which perhaps dominated the professional trainings and

careers of the majority of those professionals who were

at the height of their careers in the 1980s and 1990s. It

may alternatively represent some resistance among the

rank and file of the profession to the removal of an

aetiological dimension and signal the existence of

members within the profession who resisted polarisation

into a medical or psychoanalytic camp and retained the

possibility of different causes of depression.

Before DSM III, the term ‘dysthymia’ referred to a

personality disorder (Ramana & Paykel, 1992). In 1980,

dysthymia was given an official place in DSM III’s

categorical axis I, being defined as a chronic mild form

ARTICLE IN PRESSS. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–58 57

of depression. However, the previous concept of

dysthymia as a personality disorder would likely cause

it to be more closely linked with chronicity rather than

mildness. Why did the authors of DSM III feel it

important to alter the definition of dysthymia? It may be

that as with ‘endogenous depression’, the concept of

personality disorder implied some alternative origin for

depression other than a biological origin. It may also

have implied that the illness was not treatable with drugs

if it was ingrained in a person’s personality. However,

the citation rate of ‘dysthymia’ indicates that the term

was not successful in the way that ‘major depression’

was; its usage did not rise significantly until the late

1980s and early 1990s (see Fig. 1). However, from 1994

to 2000, dysthymia is the most dominant label

after ‘major depression’. This perhaps reflects the

greater definition given to it in DSM IV and also a time

lag in the mental health professional community

adjusting to its reframing as a mood disorder rather

than a personality disorder. It may also reflect the

problematic issues surrounding the concept of chronicity

in DSM III.

DSM III was seen as a breakthrough for psychiatric

classification in terms of its multiaxial system covering

the biopsychosocial model of mental illness and might

therefore be expected to capture the dimensions of

depression reflected in the clusters identified in the

present study. The five axes of DSM do cover some of

the clusters, but DSM noticeably lacks a ‘temporal’ axis

and a ‘treatment response’ axis. Notable also are the

labels identified in this study that do not appear in either

DSM or ICD—‘chronic depression’, ‘treatment resistant

depression’ and ‘refractory depression’, which all have

implications for the duration or course of illness.

‘Recurrent depression’ does not appear in DSM, but

does appear in ICD 10. Keller (Keller et al., 1995),

leading the DSM IV mood disorder group, described the

concern of some that DSM III had no provision for

chronic major depression in the distinction between

depression and dysthymia. Keller went on to defend the

DSM IV system which omitted ‘chronic major depres-

sion’, but did provide a classification frame for the

course of depression. This embraced the six potential

courses of depression which supposedly fitted the

majority of patients in the DSM IV mood disorder field

trials (Keller et al., 1995) and was reflected in DSM IV

largely within the categorical distinction between ‘dys-

thymia’ and ‘major depression’. Moreover, Williams,

who would later lead the DSM IV working group on the

multiaxial system, mentioned in 1985 the possibility of

considering a treatment-response axis (Williams, 1985)

to resolve the confusion of course of illness with severity,

problematic in DSM III. However, in a later report of

the DSM IV work group in 1990, this potential new axis

was firmly ruled out (Williams, Goldman, Gruenberg,

Mezzich, & Skodol, 1990).

The decisions by DSM III and IV authors to resist

either a temporal- or treatment-response axis, in spite of

problems associated with not having them, may be

considered within the social context of psychiatry. In

particular, a chronic dimension to depression may have

been considered a threat to the psychiatry profession in

the context of the rise of community care in that a

chronic illness lends more to the idea of long-term

management in the community rather than acute

treatment by psychiatrists. Chronicity also implies the

lack of a cure and this may also threaten psychiatry’s

shift to a medical model in which a disease can be cured

through advancement of science. However, as with the

ongoing use of ‘endogenous depression’, the slow

development of popularity of the term ‘dysthymia’ and

ongoing use of terms such as ‘chronic depression’ not

sanctioned in DSM III may indicate that the DSM

system failed to provide a meaningful representation of

course of illness acceptable to all mental health

professionals and that within the field are those who

employ concepts of the chronicity and the cyclical nature

of depression, in spite of the avoidance of these issues in

DSM III and IV.

Conclusion

Although the Oxford English Dictionary (2nd Edi-

tion, 1989) cites the first professional use of ‘depression’

in 1905 in ‘Psychological Review’, Snaith (1987)

describes a slightly earlier usage by Bevan Lewis in

1899: ‘‘In simple pathological depressiony the patient

exhibits a growing indifference to his former pur-

suitsy’’. Whatever its exact origins, however, it is clear

that the term ‘depression’ used as a medical label to

describe mood and mental health, especially in the

context of the ‘neuroses’, is a Western twentieth-century

phenomenon (Armstrong, 1980). In other words, the use

of the diagnostic label of depression is historically and

culturally located in very recent times. Furthermore, the

proliferation of sub-types of depression also reflects

contemporary social events and processes.

To a great extent, the integrity of psychiatry as a

medical specialty depends on the consistency of diag-

nostic labelling. Without consistency there would be

difficult communication within the professional com-

munity, as meetings, correspondence, textbooks and

journals would use conflicting language. Psychiatry

therefore needs a certain level of diagnostic stability

for its very survival. Yet, as pointed out earlier, it is

more difficult to achieve this degree of stability in an

area devoid of biological correlates with which to

anchor diagnostic definitions. The solution that has

therefore emerged in the second half of the twentieth-

century has been the Diagnostic Manual that would seek

ARTICLE IN PRESSS. McPherson, D. Armstrong / Social Science & Medicine 62 (2006) 50–5858

to impose consistency, stability and control on the

centripetal forces of clinical diagnosis.

In large part the Diagnostic Manuals have been

successful, albeit their influence has not always been

immediate. The most significant new diagnostic sub-type

introduced (by DSM III) has been ‘major depression’,

which rapidly dominated psychiatric thinking, for

reasons described. Other labels, such as dysthymia, have

taken longer to become established. But there have also

been notable ‘failures’, as in the continuing use of

‘endogenous depression’ despite its removal from the

formal classification. Ironically, as argued, some of these

survivals may be related to internal conflicts within

psychiatry on the best way to maintain a unified

hegemony.

Clearly, struggles internal to psychiatry are important

in developing and maintaining diagnostic classifications.

At its most crude, this is illustrated by the group that

manages to achieve authorship of any Diagnostic

Manual; the triumph of biological over psychoanalytic

psychiatry (at least in the US) therefore had a major

impact on the diagnoses that patients were to subse-

quently receive. But psychiatry has also existed in a

rapidly changing healthcare and technological milieu

and in many ways, diagnosis has had to respond as

psychiatry adapted to a new environment, whether it

was the era of community care or pharmacological

developments that potentially transformed the thera-

peutic potential of its practitioners.

In summary, this study has argued that the publica-

tion of Diagnostic Manuals in psychiatry may have been

driven by and reflected the need for psychiatry to

establish and maintain a specific area of expertise in the

mental illness marketplace in order to survive as a

profession. In part this involved strengthening its

alliance with the rest of medicine through the medica-

lisation of mental illnesses such as depression and in part

developing medical models of mental illness in order to

create a dominant power base for itself as a profession:

for both of these goals, control over the diagnostic

labelling process was vital. This study provides an

example of how trends in labelling and language in

depression and mental health should not be taken at face

value, but considered as an indicator of important social

processes going on within the mental health professions.

References

Armstrong, D. (1980). Madness and coping. Sociology of

Health and Illness, 2, 293–315.

Ban, T. A. (2001). Pharmacotherapy of depression: A historical

analysis. Journal of Neural Transmission, 108, 707–716.

Bayer, R., & Spitzer, R. L. (1985). Neurosis, psychodynamics,

and DSM-III. A history of the controversy. Archives of

General Psychiatry, 42, 187–196.

Berrios, G. E. (1999). Classifications in psychiatry: A con-

ceptual history. Australian and New Zealand Journal of

Psychiatry, 33, 145–160.

Fabrega, H., Jr., Ahn, C. W., Boster, J., & Mezzich, J. E.

(1990). DSM III as a systemic culture pattern: Studying

intracultural variation among psychiatrists. Journal of

Psychiatric Research, 24, 139–154.

Frances, A., First, M., Pincus, H. A., Widiger, T., & Davis, W.

(1990). An introduction to DSM-IV. Hospital and Commu-

nity Psychiatry, 41, 493–494.

Geller, J. L. (2000). The last half-century of psychiatric services

as reflected in psychiatric services. Psychiatric Services, 51,

41–67.

Grob, G. N. (1991). Origins of DSM-I: A study in appearance

and reality. American Journal of Psychiatry, 148, 421–431.

Havens, L. (1985). Historical perspectives on diagnosis in

psychiatry. Comprehensive Psychiatry, 26, 326–336.

Keller, M. B., Klein, D. N., Hirschfeld, R. M., Kocsis, J. H.,

McCullough, J. P., Miller, I., et al. (1995). Results of the

DSM-IV mood disorders field trial. American Journal of

Psychiatry, 152, 843–849.

Lee, S. (2002). Socio-cultural and global health perspectives for

the development of future psychiatric diagnostic systems.

Psychopathology, 35, 152–157.

Pincus, H. A., Davis, W. W., & McQueen, L. E. (1999).

‘Subthreshold’ mental disorders. A review and synthesis of

studies on minor depression and other ‘brand names’.

British Journal of Psychiatry, 174, 288–296.

Ramana, R., & Paykel, E. S. (1992). Classification of affective

disorders. British Journal of Hospital Medicine, 47, 831–835.

Sadler, J. Z., Hulgus, Y. F., & Agich, G. J. (1994). On values in

recent American psychiatric classification. Journal of

Medicine and Philosophy, 19, 261–277.

Snaith, R. P. (1987). The concepts of mild depression. British

Journal of Psychiatry, 150, 387–393.

van Praag, H. M. (1990). The DSM-IV (depression) classifica-

tion: to be or not to be? Journal of Nervous and Mental

Disease, 178, 147–149.

Williams, J. B. (1985). The multiaxial system of DSM-III:

Where did it come from and where should it go? I. Its

origins and critiques. Archives of General Psychiatry, 42,

175–180.

Williams, J. B., Goldman, H. H., Gruenberg, A., Mezzich, J. E.,

& Skodol, A. E. (1990). DSM-IV in progress. The multiaxial

system.[comment]. Hospital and Community Psychiatry, 41,

1181–1182.