International Journal of Epidemiology, 2014, i53–i66
doi: 10.1093/ije/dyu106
Advance Access Publication Date: 3 June 2014
Original article
Original article
Psychiatric epidemiology, or the story of a
divided discipline
Steeves Demazeux
Université Bordeaux Montaigne, Laboratoire SPH – Philosophy Bordeaux, France.
E-mail: [email protected]
Accepted 16 April 2014
Abstract
This article traces the historical decisions, concepts and key professional collaborations
that laid the foundations for the formation of American psychiatric epidemiology during
the 20th century, up to the discipline’s institutional consolidation, circa 1980, when the
third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) was
published. Thomas Kuhn’s ‘disciplinary matrix’ is mobilized as a framework that allows
the institutional and intellectual construction of a discipline to be analysed as separate
but intertwined components, without assuming that the two evolve in tandem. The identification of the strengths as well as the frailties and internal divisions of the discipline as
it developed reveals a paradoxical situation: a time lag between psychiatric epidemiology’s institutionalization and public recognition, on the one hand; and the weak coherence of its intellectual components, on the other hand. We briefly trace the origins of split
among the discipline’s aetiological models of mental disorders and suggest that the lack
of coherence among them has prevented psychiatric epidemiology from achieving the
status of a normal scientific discipline, in the Kuhnian sense. Without a more explicit
attention to the intellectual rationale of the discipline, psychiatric epidemiology will continue to maintain a strong institutional dimension and weak intellectual matrix.
Key Messages
• Compared to the history of general epidemiology and to that of most epidemiological sub-disciplines, psychiatric epi-
demiology has developed more slowly, both institutionally and intellectually;
• From the late 1950s on, psychiatric epidemiology in the USA was gaining institutional recognition, but it remained
weak (or “pre-paradigmatic” in the Kuhnian sense) intellectually;
• By 1980, American psychiatric epidemiology was undergoing a profound conceptual crisis. There was a split in its
disciplinary matrix between two radically opposite approaches : a medical-centred approach (as embodied by the
DSM-III) and a socio-ecological approach (favored by some of the most influent actors in the history of psychiatric
epidemiology).
C The Author 2014; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association
V
i53
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
Introduction
Since the 19th century, philosophers of science have utilized
the term ‘discipline’ to characterize a specialized branch of
knowledge with its specific objects, theories, methods, academic journals and degree courses in universities. A ‘discipline’ implies the existence of both a structured community
of scholars (its institutional or organizational dimension)
and a specific set of methods, concepts, practices, problems
and solutions (its intellectual dimension).1–3 Compared
with the history of general epidemiology and that of most
epidemiological sub-disciplines,4 psychiatric epidemiology
has developed more slowly, both institutionally and intellectually. By the 1980s, the institutional relevance of psychiatric epidemiology had been fully accepted in the USA. This
paper questions, however, whether the intellectual coherence of psychiatry epidemiology was similarly assured by
that period. Without such coherence, psychiatric epidemiology cannot be considered to have achieved the status of a
strong, let alone fully developed, scientific discipline.
To follow this line of enquiry, this paper uses Thomas
Kuhn’s concept of a ‘disciplinary matrix’ (defined below) as
a relevant framework in that it captures both the institutional and the intellectual constructions of a discipline,
without assuming that they must evolve in tandem. After
introducing Kuhn’s concept, the paper traces the historical
decisions, concepts and professional collaborations that laid
the foundation for the formation of North American psychiatric epidemiology in the 20th century. It signals the relationship of psychiatric epidemiology to both psychiatry and
general epidemiology, as well as the unique research partnership between
psychiatrists,
psychologists
and
sociologists, many of whom would become the pioneers of
the discipline. It then evaluates these developments by comparing the institutional and the intellectual accomplishments, taking 1980 [the publication year of the third edition
of the American Psychiatric Association’s Diagnostic and
Statistical Manual of Mental Disorders (DSM-III)] as both
the culmination of and the turning point for the discipline’s
development. The last section discusses reasons for a lag in
the two dimensions of the psychiatric epidemiology’s disciplinary matrix, namely the split within the intellectual component (‘matrix’) of the discipline, whose coherence is
nevertheless necessary to the achievement of what Kuhn
considers a normal science. In conclusion, the relevance of
this analysis to further research on the epistemological history of biomedical disciplines is outlined.
Thomas Kuhn’s elaboration of the scientific
paradigm: the disciplinary matrix
Thomas Kuhn’s concept of ‘paradigm’, introduced in
The Structure of Scientific Revolutions,1 has been very
i54
influential in philosophy of science. But it has also often
been criticized for its ambiguity and multiple meanings.
Kuhn himself was quite cognizant of these critiques by the
time the second edition of The Structure of Scientific
Revolutions appeared. In his 1969 Post-script to that edition and in further essays,5,6 he attempted to clarify the
paradigm concept in at least three principal ways.
First, Kuhn envisioned the paradigm in its most global
sense as that which a community of scientists share and to
which they are committed. To capture this, he suggested
replacing the term ‘paradigm’ with that of ‘disciplinary
matrix’, which referred at once to that which is possessed
in common by the practitioners of a particular profession
(‘disciplinary’) and the ordering of a set of intellectual
elements (‘matrix’) within the discipline.
Second, by breaking down the disciplinary matrix in
this way, Kuhn allowed for the contrast between its institutional aspect, or ‘community structure’, and its intellectual
or cognitive aspect. In the latter, he included: (i) symbolic
generalizations; (ii) metaphysical models; (iii) shared values; and (iv) exemplars. One strength of Kuhn’s approach
lies in its requirement that a ‘normal science’ can progress
only if all the aspects of the disciplinary matrix are adequately developed. Or, as Alexander Bird puts it, if ‘there
is the strong commitment by the relevant scientific community to their shared theoretical beliefs, values, instruments
and techniques and even metaphysics’.7 Kuhn further insists that one can speak of a paradigm only when a discipline is fully constituted, that is when the institutional
aspects (or organizational norms) and the theoreticalmethodological aspects (or cognitive norms) of the discipline overlap in a dynamic way.
But whereas institutional and cognitive dimensions are
intertwined, an actual overlap does not always occur: a
gap may open up between the achievement of a normalized
institutional framework (the community of scientists, with
their academic departments, journals, professional societies and so forth) and a still ambiguous intellectual project, itself split into numerous definitions and models. For
example, as the philosopher Jean Gayon has shown, this is
the case for genetics: after the rediscovery of Mendel’s laws
in 1900, genetics very quickly became a specialized discipline, and it underwent intense institutionalization in the
1940s and the 1950s. Yet today it would be difficult from
the Kuhnian perspective to still grant genetics the status of
a scientific ‘discipline’ in its own right, given its fissured
theoretical identity. The reason for this loss of theoretical
identity, at a time when the tools and methods of genetics
now exist throughout the biological and medical sciences,
can be traced to the concept of the gene itself. For, as
Gayon demonstrates, the ‘gene’ evolved from a theoretically robust notion at the beginning of the 20th century into
i55
a polymorphous and difficult-to-grasp notion by the century’s end.3
Third, Kuhn emphasized a subset of the disciplinary
matrix, the exemplar, as being more important than the
matrix’s other elements for attaining the status of a scientific discipline.5 He defined exemplars as standard examples of problem-solving within a specific discipline,
more closely linked to practice than are other aspects of
the disciplinary matrix, such as symbolic generalizations or
models. As Kenneth F Schaffner has pointed out, the
Kuhnian notion of ‘exemplar’ has ‘some interesting implications for the analysis of particularity in medicine’.2 As an
exemplar commonly denotes in biomedicine a particularly
successful example of an experiment (or the solving of a
problem, whether or not experimentally), it inspires other
members of the discipline as they take on new puzzles, and
it plays an essential role in training and transmitting a discipline’s way of working, which always operates to a certain extent by analogy with previous problem-solving.
Hence exemplars are crucial to the success of the overall
disciplinary matrix.2 (It is useful for purposes of clarity to
note Peter Godfrey-Smith’s qualification of the disciplinary
matrix as a paradigm in the broad sense, and the exemplar
as a paradigm in the narrow sense.)8
Kuhn’s original conception of a ‘disciplinary matrix’ is
particularly useful for questioning the intellectual coherence of a complex discipline with regards to its institutional functioning.9–12 We now turn to one such case: the
evolution of psychiatric epidemiology; its intersection with
general epidemiology, psychiatry and the social sciences;
and its paradoxical development.
The multifaceted early history of psychiatric
epidemiology: mental diseases in the fold of
the ‘new epidemiology’
The conditions for the emergence of an epidemiology of
mental diseases were established early on, but changes
within general epidemiology created a conceptual space for
psychiatric epidemiology to move beyond the application
of statistics to treated cases of mental illness. The early
developments that fostered the formation of psychiatric
epidemiology can be divided into three stages.
Proto-epidemiology of mental diseases
The origins of psychiatric epidemiology are anchored in the
administrative statistics tradition of 19th century asylums
and the very early statistical studies that attempted to correlate mental diseases with various environmental factors.13
The British medical statistician, William Farr (1807–1883),
contributed to the understanding of the relationship between disease and the environment; Sir Arthur Mitchell
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
(1826–1909) in the UK and Edward Jarvis (1803-1884) in
the USA applied statistics to examine the disease-environment relationship specifically for mental disease. Unlike
keepers of asylum statistics, which for decades would constitute the major and often only source of data on cases of
mental illness, Mitchell and Jarvis gathered their statistical
data both within and outside the asylum. Thus, they laid
very early the groundwork for psychiatric epidemiology
and provided an early solution to the thorny methodological problem of whether and how to include both treated
and untreated cases of mental illness.13–17
The broadening of general epidemiology
By the first half of the 20th century, the study of mental
diseases was potentially furthered by a broadening of the
scope of general epidemiology to encompass phenomena beyond infectious diseases. In 1928, the British physician Sir William Hamer employed the term ‘new
epidemiology’—possibly the first use of this expression—
to address the considerable progress that had been made in
the previous century in analysing and preventing epidemics.18 However, epidemiology was soon expanding to
include new areas of enquiry that would redefine its most
basic research object. Notably, the English epidemiologist
and statistician Major Greenwood demonstrated the value
of applying the epidemiological method to non-infectious
diseases such as cancer in a key 1935 textbook. He called
for nothing less than a theoretical shift of epidemiology’s
centre of gravity, arguing that for the epidemiologist, it
was no longer the contagiousness of a disease that mattered, but rather its status as a crowd disease, by which
he meant all pathological conditions with a significant
distribution in a given population.19
Over 20 years after Hamer proclaimed the advent of a
‘new epidemiology’, John Gordon, a professor in the
Department of Epidemiology at Harvard, used the term in
a very different way. In a 1950 article, he pronounced epidemiology as poised to become ‘a discipline with implications more far-reaching than the study of epidemics’.20
Seemingly disparate public health problems, ranging from
cancer and cardiovascular diseases to the prevention of
road accidents, would soon warrant consideration as genuine topics for epidemiological study. Gordon’s ‘new epidemiology’ reflected how epidemiologists worldwide were
coming to accept the transformation of the discipline into
what Greenwood had earlier called ‘the study of the mass
aspects of disease’.20 This change of focus met resistance
from some camps.21 However, the community of North
American researchers proved particularly open to the new
epidemiology, partly because of institutional concerns with
the problem of redefining the missions of the nation’s
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
public health officers in the aftermath of World War II. At
the same time, Gordon noted, people were witnessing ‘a
single epidemiological universe’ in a ‘shrinking world’, as
travelling became commonplace, trade between cities and
countries intensified and the need for water and food
reached global proportions.20 Once tropical diseases could
be transported more easily, former epidemiological subunits, based on geographical criteria or the particularities
of place, would lose their relevance. Moreover, the improvement of living conditions and the ageing of the population required a reordering of priorities, as infectious
diseases became less prevalent and the burden of chronic
diseases emerged as a major public health issue.20
Psychiatry and the new epidemiology
Within the discipline of psychiatry, support for the idea that
new types of statistical studies and new objects of enquiry
might warrant the term ‘epidemiological’ developed more
slowly. The issue was first raised explicitly in a 1937 paper,
entitled ‘Is there an epidemiology of mental disease?’ delivered to members of the American Public Health
Association by the Massachusetts physician, Henry Elkind.
He claimed that, to his knowledge, no published texts in
which mental diseases were clearly treated as epidemiological objects could be found in the psychiatric literature
before the early 1920s. However, he did note that in a lecture published in 1923, Professor Milton J Rosenau,
Director of the Department of Preventive Medicine and
Epidemiology at Harvard University, had suggested the possibility of applying epidemiology to mental diseases. Elkind
identified several sources of resistance that had impeded the
development of a full-fledged epidemiology of mental diseases. At the top of his list was the abstract and intangible
nature of psychiatric symptomatology. Elkind mentioned
other obstacles, such as: the conflict between theoreticallybased schools; the fact that the available statistical data, for
the most part drawn from asylum statistics, were partial
and often inexact; the ‘single aetiologically-minded’ character of most doctors; and, more generally, the infectious
thicket in which epidemiology as a whole seemed trapped.
Yet these obstacles did not prevent him from predicting that
‘during the next decade or two, mental disease [will] become an important part of epidemiological science.’22
An original partnership between psychiatry
and the social sciences
Beyond the proto-epidemiology period of asylum statistics,
the growth of an epidemiological science of mental disease
predicted by Elkind can be traced to the crucial contribution of social scientists, particularly in the USA.23,24 Three
i56
major types of contribution influenced the formation of
psychiatric epidemiology as a discipline.
A first strand developed within sociology. In fact, the
particular tenor of the first US studies of mental disease
outside the asylum, characterized by a concern for community, place and social status, came from sociologists—the
University of Chicago’s Department of Sociology, founded
in 1892, playing a central role in this endeavour. Within
the field of psychiatry, the Chicago School exerted a direct
influence on community studies, particularly through the
seminal research of Chicago-trained sociologists Robert
Faris and Warren Dunham. Their study, published in
1939, examined the medical records of over 30 000 hospitalized psychiatric patients and concluded that the rates of
schizophrenia in Chicago were higher in more deteriorated
urban areas.25 (For the influence of the Dunham and Faris
study on French sociology of mental disease, see Henckes26
in this issue of IJE.) During the interwar years, the
National Institutes of Health sponsored community studies
in various areas of non-infectious chronic disease, particularly cardiovascular disease. However, after World War II,
Faris and Dunham’s ecological perspective provided the
impetus for a new kind of socio-medical enquiry on mental
health, spearheaded by sociologists like August
Hollingshead, Leo Srole, Lee Robins, Bruce Dohrenwend
and the social psychologist, Barbara Snell Dohrenwend.
These scholars pioneered the development of psychiatric
epidemiology as a discipline, each in their own way.27
On the whole, their research illustrated the application
of sociology in medicine, in contrast to the contemporaneous field of the sociology of medicine.28 It intersected
with the more general concerns in the 1950s and 1960s of
physicians and epidemiologists working both inside and
outside mental health, including John C Cassel and
Mervyn Susser, whose research in South Africa contributed
to the development of epidemiological thinking focused on
social determinants,29,30 and the British epidemiologist
Jerry Morris, who wrote the first post-World War II
textbook for the burgeoning discipline of epidemiology.31
Cassell, Susser, Morris and their followers shared with the
pioneering mental health sociologists a particular concern
for the relationship of health and disease to socioeconomic
position and environmental factors, a concern later identified with social epidemiology.32
Social science thinking within psychiatry, notably in the
area of social psychiatry, constituted a second strand in the
construction of American psychiatric epidemiology. Social
psychiatry exerted a significant influence on the promotion
of epidemiological studies in Europe and in North
America. However, social psychiatry as it was understood
after World War II differed radically from its prewar definitions. Until the 1940s, no clearly defined school of
i57
thought characterized social psychiatry, and the orientations defended under its label varied widely from one country to the next. In Germany, for instance, social psychiatry
took on the quite broad project of studying, as historian
Michael Neve has noted, ‘the mental health of the whole
population’ in order to promote ‘an active programme as
to how to prevent mental pathology [from] increasing
within [the] population’.33 German social psychiatrists,
like some of their European counterparts, mobilized the
idea of heredity, as it had been expounded in theories of
degeneration, as a target of prevention. A similar notion of
prevention developed in the USA under the impetus of the
mental hygiene movement founded by the former asylum
patient, Clifford Beers, and the Swiss-born psychiatrist,
Adolf Meyer.27 Whereas this movement was anchored in
the combination of self-help practices, system reform and
community-based interventions with which historians of
psychiatry are more familiar, its principles of prevention
were nevertheless relayed and reinterpreted by eugenicists
in the 1920s.34
After World War II, social psychiatry acquired a sounder identity, partly due to the development of the mental
health paradigm in the USA and the influence of the World
Health Organization (WHO).35 At that stage, it became
more specifically defined as ‘a theoretical field that postulates the relevance of social variables to the causes of mental illnesses and their treatment and prevention’.36 From a
theoretical point of view, social psychiatry in the USA
resembled more closely the ‘social medicine’ defended
by the Englishman John Ryle, and aimed at effecting the
transition towards an epidemiology of non-infectious
diseases.37
A third strand, running alongside the contributions of
sociology and social psychiatry, involved another professional group, namely psychologists. Working with psychiatrists, psychologists exerted a discreet but decisive
influence on the nascent discipline of psychiatric epidemiology. As several historical accounts have noted, psychologists, working as assistants or technicians, affected the
organization of epidemiological studies in psychiatry38 and
were crucial to the design of the first standardized tools for
epidemiological use.39 They included eminent figures like
John R Wittenborn, Maurice Lorr, John Overall, Paul E
Meehl and Joseph Zubin, although the importance of these
psychologists, most of whom held respectable academic
positions, to the development of psychiatric epidemiology
tends to be under-acknowledged. For example, John R
Wittenborn, Maurice Lorr and John Overall made major
contributions to the development of standardized instruments as well as to the use of factor analysis in psychiatric
research. Their work notably influenced the case for using
dimensional approaches in classification as opposed to the
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
dominant neo-Kraepelinian categorical approach favoured
by the designers of DSM-III.40
Yet in the 1960s, at a time when US psychiatry was in
intellectual turmoil because of its internal divisions, many
of these psychologists were able to convince psychiatrists
not only to introduce more rigour into their diagnostic
approaches and quantitative analyses, but also to give
population studies precedence over clinical ones.41 To give
but one example, Joseph Zubin, a professor at Columbia
University, founder of the Biometric Research Unit at New
York State Psychiatric Institute and former President of the
American Psychopathological Association (APPA), contributed significantly to conceptualizing what a psychiatric
epidemiology might actually resemble. In a 1960 talk delivered to psychiatrists in Chicago, Zubin identified six
dominant models of psychopathology: the socio-cultural,
the developmental, the learning-centred, the internal environment-centred, the genetic and the cerebral.42 He highlighted a seventh model as superior, coining for it the label
‘epidemiological super-model’. For him, this was the only
theoretically powerful model. Epidemiology thus comprised a ‘super-model’ rather than simply a method or a
simple model because it integrated all of the other models,
from the purely environmental to the purely biological. It
could thus encompass the positive qualities of the range of
theoretical models, while avoiding their respective shortcomings. Finally, the super-model proved to be the only
truly inclusive one, Zubin argued, because it alone could
provide the researcher with ‘the whole picture’ of the
particular causes of mental diseases.
Owing largely to the efforts of influential scholars like
Zubin himself, the conditions were ripe for a disciplinary
consensus among epidemiologists, sociologists and clinicians. All of them stressed the necessity of establishing an
original partnership between the biological sciences and
the social sciences as a foundation for psychiatric
epidemiology.42–44
The 1960s: psychiatric epidemiology’s
institutional success
Neither an emerging consensus between theoreticians nor
the constitution of an original partnership between social
scientists and psychiatrists sufficed to establish psychiatric
epidemiology as a new discipline. For this to happen, Kuhn
reminds us, all the aspects of a disciplinary matrix must
be shared by a structured community of scholars. More
specifically, the medical historian and physician George
Rosenhas emphasized the degree to which professional
and scientific societies were essential to providing an
identity for epidemiologists in general. In his review of
John R Paul’s book on the history of the American
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
Epidemiological Society,45 Rosen noted that: ‘Specialized
knowledge and techniques may exist and be used by physicians, and yet there may be no sense of a separate identity.
Only when there are bonds between such practitioners
which take shape within an association based upon like
interests and common problems does an identifiable specialized group appear’.46
Rosen’s general conclusion is equally applicable to psychiatric epidemiology. In fact, an identifiable, specialized
group was gradually emerging from epidemiologists, physicians, social scientists and statisticians working on mental
health issues, and its members came to be publicly designated as ‘psychiatric epidemiologists’. In 1959, the
American Psychopathological Association devoted its
Forty-ninth Annual Meeting to the epidemiology of mental
disorders. Three years later, a reviewer of the just-published proceedings of this important event, writing in the
Journal of the American Medical Association, humorously
remarked that ‘there is now an epidemic of investigations
of the epidemiology of psychiatric disorders’.47 Indeed
from the late 1950s on, the private health policy foundation, Milbank Memorial Fund, and the National Institute
of Mental Health, funded studies specifically designated
as psychiatric epidemiology. Among the first journals to
devote space to psychiatric epidemiology was Social
Psychiatry, founded in 1966, although it did not change its
title to Social Psychiatry and Psychiatric Epidemiology
until 1998.36 In 1968, Mervyn Susser, by then working
in the USA, published Community Psychiatry.
Epidemiological and Social Themes, which reviewed relevant studies to date, including his own, and discussed
issues raised by the application of epidemiological methodology to the measurement of mental illness in the community.48 Psychiatric epidemiology was gaining institutional
recognition through its presence in North American universities. Columbia University founded the first psychiatric
epidemiology training programme in 1967 and obtained
funding for it from the National Institute of Mental Health
(NIMH) Program to Enhance Diversity in Institutional
Training (T32 Program) in 1972. From the late 1970s on,
NIMH funded Psychiatric Epidemiology Training (PET)
programs at several other universities as well. Meanwhile,
the first psychiatric epidemiology textbook, by British authors Brian Cooper and Howard Morgan, appeared in
1973.49
Psychiatrists had also turned to psychiatric epidemiology as it confronted the institutional watershed
represented by the second revision of the American
Psychiatric Association’s Diagnostic and Statistical
Manual of Mental Disorders (DSM-II). For the first time,
the DSM revision task force integrated a large number of
recognized epidemiologists and epidemiologically-oriented
i58
researchers, including Ernest Gruenberg (the first psychiatrist in the USA to hold a doctorate in epidemiology),50
the biostatistician and epidemiologist Morton Kramer, the
pioneer mental health researcher Benjamin Pasamanick
and the researcher and psychiatric administrator Lawrence
Kolb.
It is beyond the scope of this paper to explore the various complex socio-cultural forces that led to the institutional recognition of psychiatric epidemiology. From the
1960s on, psychiatric epidemiology clearly enjoyed all the
attributes of a young discipline, including the formation of
a critical mass of internationally renowned specialists.
Barbara and Bruce Dohrenwend’s categorization of the
discipline into three generations of community studies51,52
fits nicely with the gradual institutionalization of psychiatric epidemiology in the USA and elsewhere—even if their
schema represents only one aspect of the history of psychiatric epidemiology, given that the field was not limited to
community studies. The first generation, which for the
Dohrenwends ended with World War II, consisted primarily of asylum statistics. At that time, psychiatric epidemiology had yet to be identified as an autonomous scientific
discipline and, somewhat strikingly, the term ‘epidemiology’ was virtually absent from the 16 studies which the
Dohrenwends included in that first generation. It was only
with the second generation, comprising about 60 studies
undertaken between 1945 and 1980, that the term ‘epidemiology’ began to appear in the definition and the selfdesignation of many of the community studies in
psychiatry.
The third generation described by the Dohrenwends
encompasses all studies undertaken after the 1980 publication of the third edition of the DSM (DSM-III), that is with
the version of DSM which for the first time aimed to standardize all diagnostic criteria. With this new generation of
studies, psychiatric epidemiology moved into a new phase,
characterized by a profusion of epidemiological studies
based on categorical measures of mental illness, rather
than on the vaguer, continuous indicators of mental health
used in earlier studies. This generation also paid attention
to the inter-rater reliability of diagnoses, which in part
compensated for the absence of biological markers and for
other problems concerning the validity of what was
defined as mental disorders. Finally, some methods used in
third-generation psychiatric epidemiology resembled those
of general epidemiology.
Yet despite methodological breakthroughs and an
increased institutional stability, psychiatric epidemiology
was undergoing a profound conceptual crisis. As we shall
see, this involved some of the most fundamental convictions embraced by the discipline’s disciplinary matrix. The
result was a widening rift between the discipline’s
i59
institutional consolidation and its intellectual foundations,
a problem to which we now turn.
1980: the consolidation of a discipline?
Institutional stability and conceptual
controversy
By providing psychiatrists with operational criteria for
each mental disorder it covered, the DSM-III53 brought
about a major transformation in the manner in which epidemiological studies of mental disorders were carried out.
But as North American researchers had begun to develop
operational criteria for diagnostic measurements to be
used in epidemiological and other types of study before
DSM-III appeared in 1980, that year can be seen as a point
of culmination as well as a turning point. In the previous
decade, sociologist Lee N Robins and the then-NIMH
psychiatrist and researcher, Darrel Regier, had developed
the Diagnostic Interview Schedule (DIS), a structured instrument capable of generating diagnoses according to the
then-future DSM-III criteria, for the Epidemiological
Catchment Area (ECA) Study. That study, which they
spearheaded, was initiated in 1978 and launched 2 years
later in five areas of the USA.24
Although it would spur an abundance of epidemiological research, DSM-III nevertheless appeared on the crest
of controversies and debates built up over the previous
years around the direction psychiatric epidemiology should
take. The controversy first arose during the 1960s and
early 1970s, in the series of round tables organized by the
Milbank Memorial Fund to promote psychiatric epidemiology as a basis for mental health policy.35 It crystallized
more clearly—and before a nationwide public of psychiatric scientists—on the occasion of DSM-III’s publication.
The essence of the controversy is illustrated by an exchange
published in several issues of the major North American
psychiatric research journal, Archives of General
Psychiatry, between the Columbia University sociologists,
Leo Srole and Anita K Fisher on the one hand, and Yale
University researchers Myrna Weismann and Gerald
Klerman on the other hand.54–57 The catalyst was Srole
and Fischer’s now classic paper, ‘The Midtown Manhattan
Longitudinal Study’ vs ‘The Mental Paradise Lost’
Doctrine’, published in the Archives in February 1980.
Srole and Fischer used the data from their impressive
socio-epidemiological survey, launched in the early
1950s,58 to attack a commonly held view among psychiatrists and psychoanalysts of the day, namely that mental
health tends to deteriorate with the progress of civilization.
As their study examined an area of Midtown Manhattan
characteristic of large, modern cities, Srole and Fischer
were well positioned to argue against the ‘mental paradise
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
lost’ view of the modern era. The two sociologists attributed that view to a sexist bias on the part of psychiatrists
seeking to show that women had more mental disease than
men, especially in the modern era. They countered that
claim with results from the Midtown Manhattan study to
suggest the opposite tendency: women’s subjective wellbeing actually tended to improve as they became more
emancipated.
The same issue of the Archives included a letter to the
editor signed by Myrna Weissman, then professor of psychiatric epidemiology at Yale University, and her husband,
the psychiatrist Gerald Klerman. Their letter did not criticize the conclusions of Srole and Fischer’s article but rather
questioned the underlying assumptions of the Midtown
study and the methodology which had produced the findings on which the two sociologists based their argument.
For instance, Weissman and Klerman opposed the idea
that there are no discrete mental diseases, but only a continuum between the normal and the pathological. This
‘continuum perspective’ undergirded the psychological
scales and mental health measures which the operational
criteria from DSM-III tried, at least partly, to displace, albeit with mixed success. But above all, Weissman and
Klerman attacked the assumptions behind Srole and
Fischer’s sociological hypotheses. They criticized the sociologists for perpetuating a style of epidemiological inquiry
that they considered to be outdated and out of sync with
the new epidemiology of ‘risk factors’ and the focus on specific, well-delineated diseases. In their view, the publication of the DSM-III with its operational criteria would
rapidly foster a renewed psychiatric epidemiology.56
This controversy foreshadowed the role that the DSMIII would come to play from a historical perspective.
Writing in the following decade, Klerman characterized
the advent of DSM-III as ushering in a radical transformation that was nothing less than a change of paradigm
within psychiatry, in the technical sense Kuhn gave to this
expression. That is, from an epidemiology centred on
the idea of mental health, the DSM-III triggered the shift
towards an epidemiology centred on specific psychiatric
disorders.9 Klerman thus addressed some of the same concerns as the present paper, concerning the status of psychiatric epidemiology as a scientific discipline. However from
a Kuhnian perspective as well as on the basis of the history
of the discipline presented thus far, his argument faltered
on several points.
First, Klerman addressed the problem of psychiatric epidemiology only from the point of view of psychiatry, as if
the two disciplines could be superimposed, or as if psychiatric epidemiology were merely a branch of psychiatry.
The paradigm change Klerman referred to did not involve
epidemiology in general. Rather, he was implying that
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
with DSM-III, psychiatric epidemiology would be able to
move closer to other types of medical research being practised at the time—in cardiology, oncology, and elsewhere—and which were turning to the search for risk
factors.56 Second, although the DSM-III would profoundly
change the nature of epidemiological studies in psychiatry,
Klerman was not sensitive to the potential for resistance
within the community of psychiatric epidemiology
researchers itself. In fact, a profound resistance arose immediately, as we will show below. Third, and more fundamentally, even if many theoretical assumptions and
methodological implications of DSM-III were to become
influential within North American psychiatry during ensuing decades,59 it can hardly be concluded that DSM-III was
leading (or has led) to a real ‘scientific revolution’ in the
field, which Klerman’s use of the paradigm concept
implied. In fact, the DSM continues to experience at best
an ambivalent success from a scientific viewpoint. The last
revision of the DSM, the DSM-5, published in 2013, has
been criticized by many clinicians and researchers for its
lack of validity.60,61 In any case, from a Kuhnian perspective, Klerman was misusing the paradigm concept by not
referring to the discipline of psychiatric epidemiology as a
whole. Rather, he tended to identify a current within
psychiatry as a paradigm. Klerman himself was aware of
this terminological problem. In a later article,62 he did recognize that the existence of five rival schools—biological,
psychoanalytical, interpersonal, cognitive-behavioural and
social—within psychiatry challenged the very concept of a
paradigm in the Kuhnian sense, unless one were willing to
admit the possibility that psychiatry as a whole was still in
a ‘pre-paradigmatic’ age. Thus his observation inadvertently illustrated the point we are arguing in this paper; that
is, that in the historical development of a discipline, a gap
may exist between its institutional and its intellectual components. Competing models, for example, result in an internal crisis within what is not yet a fully coherent
discipline. But what are the implications of Klerman’s observation concerning multiple schools of psychiatry?
The disciplinary matrix of psychiatric
epidemiology before 1980
What did the disciplinary matrix of psychiatric epidemiology actually resemble in 1980? Contrary to Klerman9
and Susser and Stein,63 who sum up the history of
epidemiology as the succession of several paradigms, we
propose that by 1980, psychiatric epidemiology in the USA
had not even reached the status of a ‘normal science’. That
is, the theoretical framework of psychiatric epidemiological studies was neither coherent nor unitary, hence not
paradigmatic. And paradoxically, from the 1980s on,
i60
the discipline continually strengthened its institutional dimension, but remained weak (or ‘pre-paradigmatic’)
intellectually.
To elaborate this point, let us return to the framework
presented at the beginning of this paper. As we saw,
Thomas Kuhn breaks down the disciplinary matrix into
its institutional (disciplinary) and cognitive (intellectual)
components. From the institutional or organizational perspective, psychiatric epidemiology had more or less consolidated by 1980. But what of the intellectual dimensions
of the disciplinary matrix, or the conjunction of its four
cognitive components—symbolic generalizations, shared
values, exemplars and metaphysical models? How cohesive
was this aspect of psychiatric epidemiology by 1980?
With regard to the first type, symbolic generalizations
and all forms of expressions ‘deployed without questions
or dissent by group members’,1 psychiatric epidemiology
adopted the vocabulary and statistical toolkit of general
epidemiology very early on, with notions such as comorbidity, continuum, density, incidence, validity, reliability
and different study designs, such as experimental, quasiexperimental, longitudinal etc.64,65 The second type of
cognitive element, or values of psychiatric epidemiology,
are also those of epidemiologists in general, and they contributed strongly to defining the scientific ambition of psychiatric epidemiology as a new discipline with regard to
the uses to which psychiatric statistics had been put in the
past. These values include scientific rigour in data analysis
as well as the humanitarian vocation of epidemiology as
promoted, for example by the WHO, with the primary objectives of human welfare and the respect of individuals.
This could not contrast more with North American and
European eugenic projects between the two world wars, in
which statistical research on mental health was used to
promote the existence of one kind of individual over another, and often through coercive means.34,35
With regard to the third type, exemplars or paradigmatic examples available in textbooks and systematically
cited by the members of the discipline, psychiatric epidemiology took shape around several studies which became
canonical references, despite the tremendous differences in
their scope and methodologies. Examples include Joseph
Goldberger’s work on pellagra in the mid 1910s,66 Faris
and Dunham’s study of urban areas and schizophrenia in
1939,25 Hollingshead and Redlich’s study of mental illness
and social class in 1958,67 the Stirling County study by
Leighton and his colleagues in 1959,68 the Midtown
Manhattan Project initiated in 1962,58 the International
Pilot Study of Schizophrenia initiated by the WHO in
196669 and the Epidemiological Catchment Area Study
launched in 1978.70 Yet the results yielded by most of these
studies are difficult to reproduce, and their methodological
i61
biases are commonly critiqued. Few epidemiologists would
claim that the studies conducted in psychiatric epidemiology before the 1980s can be summarized as a series of
successive discoveries. This absence of a cumulative history
reveals an inherent fragility of the matrix of psychiatric
epidemiology.
The final type of element in a disciplinary matrix is what
Kuhn calls its ‘metaphysical part’, by which he means the
adhesion of researchers to certain privileged theoretical
models. In the history of psychiatric epidemiology, three important concepts have functioned as ‘preferred or permissible analogies and metaphors’3 that enable researchers as a
group to adhere to common models: infectious contagion,
hereditary transmission and moral epidemics. Interestingly,
even if they did not contribute to decisive scientific progress,
these three models polarized the research enterprise of epidemiological puzzle-solving during the entire past century.
Infectious contagion
The idea that insanity could be transmitted by germs, or
that it stemmed from an infection, has recurred repeatedly
in the history of psychiatry. The theory of focal infection,
which goes back to antiquity, postulated that a local and
persistent, even slight, infection could be at the origin of
the most serious systemic disorders. In the early 20th
century, the heyday of bacteriology and the hygienist
movement, this theory met with considerable success
throughout the medical field. Many doctors and dentists
maintained that by treating dental infections, cleaning
sinuses and removing tonsils, diseases ranging from
rheumatism to gastric problems, anaemia, severe depression and psychosis could be fought. For instance Henry
Cotton, medical director of the Trenton Psychiatric
Hospital in New Jersey, became famous in the 1920s when
he undertook to cure patients by removing all sorts of
organs: teeth, tonsils, gall bladder, stomach and spleen. He
defended the idea of an insidious transmission of mental
diseases, notably within the family, through contact between individuals and perhaps even, in the case of pregnant
women, through in utero contamination.71 Similarly, examples of contagion in the case of mental illnesses were
presented in the 1950s and 1960s to support the scientific
rationale for applying epidemiological methods to mental
illnesses, as Anne Lovell points out in this issue of IJE.35
Although not dominant, the infectious contagion model
survives today in some psychiatric thinking (for example,
the idea that an ‘invisible plague’ explains the rise of mental illness from the 18th century to the present)72 and in
the mass media.
Even if this model has proved harmful in the past, the
possibility does exist that some mental disorders may be
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
due to the transmission of an unknown infectious agent.
(Think for example of the history of general paresis, which
psychiatrists discovered to be caused by syphilis.) Yet, in
the scientific literature, it is sometimes difficult to understand whether authors are referring to the infectious contagion model in the literal sense of the term (contagion as a
transmission through actual contact) or in a metaphorical
sense. A preferable alternative to mere ‘behavioural’ contagions (i.e. without hypothesizing the spread of an infectious agent) is ‘moral epidemics’, which will be discussed
below.
Hereditary transmission
With the rediscovery of Mendel’s laws in the early 20th
century, hereditary research in psychiatry proliferated. For
many psychiatrists, the idea of a hereditary transmission of
insanity was soon considered as a contagion risk far more
insidious than any infection caused by a germ. In the USA,
figures like: the lawyers Prescott Hall and Madison Grant;
the biologist Charles Davenport; the director of the
Eugenics Record Office, Harry Laughlin; as well as two
physicians often considered important precursors of psychiatric epidemiology, Horatio Pollock and Thomas
Salmon; contributed significantly to spreading the idea
that psychiatrists were responsible for protecting the ‘genetic vitality’ of the North American population, notably
against immigration, which was considered a vehicle for
spreading mental diseases.73,74 As an alternative model of
contagion,75 heredity proved more worrisome than the infectious model, in so far as a recessive gene could be transmitted in an inconspicuous way and might take several
generations before being expressed. If insanity could be
transmitted by recessive genes, and not merely by dominant genes, psychiatric examination of new immigrants
entering the country no longer sufficed to prevent mental
illness. Rather, investigations would have to include information on a family’s ancestors, or at least a sufficiently
long period of psychiatric surveillance of new arrivals in
North America. Far from focusing on the individual, eugenicists thus understood very early on that new statistical
methods and family studies would need to be mobilized so
that society could defend itself against the proliferation of
mental diseases. From the 1950s on, the development of
family and twin studies exerted an enormous influence
within psychiatric epidemiology.
Moral epidemics
Emile Durkheim’s sociology exerted a profound intellectual influence on the development of psychiatric epidemiology,76–78 although the discipline has not always
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
acknowledged this. His work on suicide, published in Paris
in 1897, quickly earned him international renown in the
social sciences. The first English translation, however, was
not published until 1951;79 and even though his ideas were
frequently opposed to those developed by the Chicago
School of sociologists, certain commentators have noted
significant similarities, especially with the earlier Chicago
School members.78,80
One of the most original features of Durkheim’s work
was his conceptualization of a phenomenon that he characterized as a moral epidemic. Holding that the sociologist’s
job was to bring attention to the social norms and pressures likely to explain the variation in rates of suicide according to geographical distribution, social origin or other
factors, Durkheim defended an original conception of epidemics. He explicitly contrasted epidemic to contagion as
two very different forms of normative pressure on the individual. Behavioural contagion is inter-individual and based
on effects of imitation. It involves a limited number of individuals. A moral epidemic, according to Durkheim, is
essentially of a sociological nature, characteristic of the
normative pressures on a group. He summed this contrast
as follows: ‘A [moral] epidemic is a social fact, produced
by social causes; contagion consists only in more or less
repeated repercussions of individual phenomena’.79 This
notion enabled him to defend a sociological approach to
disease and mental health in which the latter were characterized as social facts, given that their distribution in a
population followed patterns correlated with social characteristics such as gender, marital status and urban vs rural,
as well as social norms. Such a distribution provided
immunity against merely psychological interpretations.
Durkheim’s conception bears a family resemblance with
an important theoretical assumption that stimulated all
medical and sociological reflection in the early 20th century: the holistic perspective. In the 1920s USA, the first
partisans of constitutional medicine were inspired by a holistic approach.81 Later, pioneering epidemiologists like
Major Greenwood and John Gordon agreed that the epidemiological perspective should take the group as its unit
of analysis. In so doing, they assumed that the whole could
not be reduced to the sum of its parts. Epidemiology was
therefore defined as a population-based style of thinking
that was explicitly opposed to the clinical viewpoint of the
doctors.20,43
Until the 1980s, the disciplinary matrix of psychiatric
epidemiology attempted to integrate these three metaphysical guiding models. The spread of a noxious agent, the
presence of a gene and the distribution of social factors
were considered the three main candidates for explaining
high rates of mental disorders in the general population.
i62
Within psychiatric epidemiology, these ideas nurtured an
ideological common ground based on a multi-causal and
holistic theoretical framework, which in turn contributed
to the vitality of the institutional partnership between social scientists and physicians within psychiatric epidemiology. But it was not to last.
A split in the matrix of the discipline
Partly in reaction to the sociological criticism of psychiatry
in the 1970s,82 many psychiatrists began to adamantly reject the idea that true mental diseases could spread like
moral epidemics. A notable example is Samuel Guze, who
deplored the many ‘easy generalizations about the effects
of social factors on psychiatric disorders’ which he thought
posed a threat to the medical model in psychiatry.83
Others claimed that the disciplinary role of sociology inside and outside mental health was becoming ever more
confused,84 and sociologists were becoming ‘deskilled’ in
epidemiology.82 The result was an ‘exhaustion’ of the original interdisciplinary perspective adopted during the past
decades.82
In this light DSM-III, which as we noted earlier reinforced the medical model in a field influenced by sociological models, represents a turning point in psychiatric
epidemiology’s development. Psychiatric epidemiology
was henceforth torn in two radically opposite directions: a
medical-centred approach, which attempted to treat mental disorders as if they were diseases like any other; and a
socio-epidemiological32 or, better still, an ecological approach.10,12,63 The split between these two broad
approaches at the beginning of the 1980s still has implications for today’s psychiatric epidemiology.
The second direction in fact illustrates exactly where
the crack had formed in psychiatric epidemiology’s disciplinary matrix. The key culprit was the by then archaic
notion of ‘ecology’. In the 1920s, Chicago School sociologists had drawn inspiration from the botanist Frederic
Clements’s seminal book on plant ecology, Plant
Succession: An Analysis of the Development of Vegetation
(1916).85 The sociologist Roderick D McKenzie was the
first to justify the value of broadening the application of
the ecology concept to human communities.86 This idea
was also central to Faris and Dunham’s perspective when
they undertook their study of the ‘ecological distribution
of mental diseases’ in Chicago.25 The notion of human
ecology rapidly spread to social medicine, notably in the
writings of John Ryle. The chairman of the Royal Society
of Medicine, Sheldon Dudley, even argued that the ecological method was equivalent to the epidemiological
method.87 John Gordon reiterated this equivalence in
1950, when he defined epidemiology as a ‘medical
i63
ecology.’20 He touted the advantage of the ecological approach as allowing disease to be addressed as a mass phenomenon, taking into account the ‘dynamic balance’
existing between the triad of pathogenic agent, the host
and the general environment.88 After Gordon, the ecological perspective in psychiatric epidemiology continued to
be treated as an overarching perspective aimed at promoting: (i) the search for multiple causes; (ii) the emphasis on
the complex entanglement of causes; and (iii) a shift in
causality from attention to the pathogenic agent to emphasis on the host and the environment, with no a priori
hierarchy between the two sets of determinants.
Early on, these three goals characterized the specificity
of the ecological perspective, its intellectual content and its
political agenda. However, the dominant position attained
by the medical model in psychiatry at the beginning of the
1980s led many psychiatric epidemiologists to focus on individual-level factors in mental disorders. Social determinants were relegated to the hypothetical role of ‘modifying
or facilitating causes’83 rather than considered possible
true causes of mental disorders.
The division among scholars between these two radically opposite approaches—the medical-centred approach
the DSM-III embodied, and the ecological approach which
several of the most influent actors in the history of psychiatric epidemiology preferred—split the disciplinary matrix
of psychiatric epidemiology at the cognitive level. Hence
the paradox of psychiatric epidemiology: it had flourished
institutionally, but as an intellectual endeavour it would
lag behind. For without cognitive coherence, the disciplinary matrix was not crystallized. Psychiatric epidemiology
was stuck in a pre-paradigmatic phase. We can further
hypothesize that it remained pre-paradigmatic during the
following three decades, since the DSM medical approach,
although dominant, has not contributed to notable scientific progress in psychiatric epidemiology so far.
Conclusion
With the publication of the DSM-III, a milestone in the history of psychiatry, the discipline of psychiatric epidemiology stood poised to move closer to general epidemiology
by providing epidemiological studies with criteria for measuring specific mental disorders. Circa 1980, the DSM-III
had set the stage for relegating the socio-epidemiological
survey to medical sociology, thereby more or less excluding
it from the disciplinary field of psychiatric epidemiology itself. The dominance of the medical model further relegated
inquiry into distant causes of mental diseases to a domain
beyond the ‘epidemiological’. These moves did not eliminate the tension between the two approaches. Thus, rather
than constituting a veritable paradigm shift within the
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
discipline of epistemology, DSM-III’s arrival revealed a split
in the disciplinary matrix of psychiatric epidemiology that
had been under way for decades, and that would remain
discreet but persistent during the following decades.
The question of whether or not psychiatric epidemiology appears much more homogeneous today than it was
when DSM-III was launched over 30 years ago cannot be
examined within the scope of this paper. Whereas we
would caution against assuming that the split has been
replaced by a single model, given the discipline’s somewhat
insecure theoretical foundations, only future studies can
determine the extent to which the intellectual division
remains within the discipline.
As a scientific enterprise—and not merely as a
profession—psychiatric epidemiology is affected by the limitations of general epidemiology and by the theoretical conundrums of modern psychiatry. We suggest that the division
between a socio-epidemiological approach and a medically
centred approach described here characterizes general epidemiology as well, judging by the critique some epidemiologists make of the epidemiology of risk factors.89,90
However, this division is stronger in psychiatric epidemiology. Disagreements among researchers over how to use logical inference and integrate several levels of causation into
one coherent picture of illness run through many epidemiological
subspecialties
other
than
psychiatric
23,91–93
epidemiology.
Within psychiatric epidemiology, such
disagreements are intensified owing to the fact that the
biological causes of mental disorders are still massively unknown. This is a non-negligible consideration, for the
absence of a consensus about causation not only
undermines the coherence of psychiatric epidemiology, it
also undermines the coherence of the psychiatric field.
On the other hand, the increasing influence of nonreductive approaches in the psychiatric literature during
the past decades may indicate that the field in general is
experiencing a conceptual evolution towards a more pluralistic and complex discipline. Or, on the contrary, it may
indicate that psychiatry in general—and not only psychiatric epidemiology—still lacks intellectual coherence and
thus has not yet reached the status of a normal science, in
Kuhn’s sense.
These tendencies need to be verified. The criteria of the
Kuhnian framework concerning the status of a scientific
discipline are quite demanding. One might even speculate
whether the absence of coherence and convergence within
a disciplinary matrix is specific to psychiatric epidemiology
or is true for other epidemiological subspecialties (cardiovascular, genetic etc.) or even for general epidemiology—
questions that are beyond the scope of this paper. What is
certain is that nowadays researchers in psychiatric epidemiology are not merely confronted with ‘anomalies’ that
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
could be resolved by ‘normal means’. They still express
hesitancy and uncertainty about the most fundamental
conditions necessary for the full consolidation of their discipline. In conclusion, given the general ideological tendency in modern medicine to localize health problems
within the individual rather than in society,94 debates on
the relevance of the ecological perspective and its renewal
within psychiatric epidemiology, could prove crucial in the
next few years for reassessing causal explanation in that
discipline, as well as in psychiatry more generally.
Funding
This work was conducted while a Postdoctoral Fellow at the
Cermes3 research centre (Paris, France), with funding from the
PHS2M (‘Philosophie, Histoire et Sociologie de la Médecine
Mentale’) programme supported by the Agence Nationale pour la
Recherche (ANR-O8-BLAN-0055-01).
Acknowledgments
I thank Jerome Wakefield for his comments on an early draft on this
article and the referees of the International Journal of Epidemiology
for their helpful reviews. I express my deep gratitude to Anne
Lovell, who directed my postdoctoral fellowship. She was tremendously helpful during the preparation of this article. She read the
manuscript critically and provided invaluable assistance. Her many
insights, comments and suggestions have improved the different
versions of this text.
Conflict of interest: None declared.
References
1. Kuhn TS. The Structure of Scientific Revolutions. Chicago, IL:
University of Chicago Press, 1962.
2. Schaffner KF. Exemplar reasoning about biological models and
diseases: a relation between the philosophy of medicine and philosophy of science. J Med Philos 1986;11:63–80.
3. Gayon J. La génétique est-elle encore une discipline? [Genetics:
still a discipline?] Médecine/sciences 2004;20:248–53.
4. Morabia A. A History of Epidemiologic Methods and Concepts.
Basel: Birkhäuser, 2004.
5. Kuhn TS. The Structure of Scientific Revolutions. 3rd edn.
Chicago, IL: University of Chicago Press, 1996.
6. Kuhn TS. Second thoughts on paradigms. In: Kuhn TS (ed). The
Essential Tension. Chicago, IL: University of Chicago Press,
1977.
7. Bird A. Thomas Kuhn. In: Zalta EN (ed). The Stanford
Encyclopedia of Philosophy. Stanford, CA: Stanford University
Press, 2011.£
8. Godfrey-Smith P. Theory and Reality: An Introduction to the
Philosophy of Science. Chicago, IL: University of Chicago Press,
2009.
9. Klerman GL. Paradigm shifts in USA psychiatric epidemiology
since World War II. Soc PsychiatryPsychiatr Epidemiol
1990;25:27–32.
10. March D, Susser E. The eco- in eco-epidemiology. Int J
Epidemiol 2006;35:1379.
i64
11. Susser M, Susser E. Choosing a future for epidemiology: I. Eras
and paradigms. Am J Public Health 1996;86:668.
12. Susser M, Susser E. Choosing a future for epidemiology: II. From
black box to Chinese boxes and eco-epidemiology. Am J Public
Health 1996;86:674.
13. Grob GN. The origins of American psychiatric epidemiology.
Am J Public Health 1985;75:229–36.
14. Grob GN. Commentary: A psychiatric epidemiological pioneer—Arthur Mitchell’s follow-up of study of hospitalized
patients. Int J Epidemiol 2010;39:1425–29.
15. Susser E, Baumgartner JN, Stein Z. Commentary: Sir Arthur
Mitchell—pioneer of psychiatric epidemiology and of community
care. Int J Epidemiol 2010;39:1417–25.
16. Susser M, Adelstein A. An introduction to the work of William
Farr. Am J Epidemiol 1975;101:469–76.
17. Davey Smith G. The antecedents of epidemiological methodology in Arthur Mitchell’s surveillance and care of the insane.
Int Jf Epidemiol 2010;39:25.
18. Hamer W. Epidemiology Old and New. London: Kegan Paul,
Trench, Trubner & Co., 1928.
19. Greenwood M. Epidemics and Crowd-diseases: an introduction
to the study of epidemiology. London: Williams and Norgate
Ltd, 1935.
20. Gordon JE. Epidemiology—old and new. J Mich State Med Soc
1950;49:194–99.
21. Reid DD. Epidemiological Methods in the Study of Mental
Disorders. Geneva: World Health Organization, 1960.
22. Elkind HB. Is there an epidemiology of mental disease? Am J
Public Health 1938;28:245.
23. Susser M. Epidemiology in the United States after World
War II: the evolution of technique. Epidemiol Rev 1985;7:
147–77.
24. Campbell ND. The Spirit of St. Louis: addiction research and
psychiatric epidemiology. IJE 2014;43(Suppl 1):i19–28.
25. Faris R, Dunham HW. Mental Disorders in Urban Areas: an
Ecological Study of Schizophrenia and Other Psychoses.
Chicago, IL: Chicago University Press, 1939.
26. Henckes N. Mistrust in numbers: The difficult development of psychiatric epidemiology in France, 1940–80. IJE
2014;43(Suppl 1):i43–52.
27. March D, Oppenheimer G. Social disorder and diagnostic order:
The US Mental Health Movement, the Midtown Manhattan
Study and the development of psychiatric epidemiology in the
20th century. IJE 2014;43(Suppl 1):i29–42.
28. Straus R. The nature and status of medical sociology. Am Sociol
Rev 1957;22:200–04.
29. Cassel J. A comprehensive health program among South African
Zulus. In: Paul BD, Miller WB (eds). Health, Culture and
Community: Case Studies of Public Reactions to Health
Programs. New York: Russel Sage, 1955.
30. Susser M, Watson W. Sociology in Medicine. London: Oxford
University Press, 1962.
31. Morris JN. Uses of Epidemiology. Edinburgh: Livingstone, 1957.
32. Berkman L, Kawachi I (eds). Social Epidemiology. Oxford, UK:
Oxford University Press, 2000.
33. Neve M. A commentary on the history of social psychiatry and
psychotherapy in twentieth-century Germany, Holland and
Great Britain. Med Hist 2004;48:407.
i65
34. Kevles DJ. In the Name of Eugenics: Genetics and the Uses of
Human Heredity. Cambridge, MA: Harvard University Press, 1985.
35. Lovell AM. Constructing psychiatric epidemiology as an international discipline: one rope, many threads. IJE 2014;43(Suppl 1):
i6–18.
36. Sabshin M. Changing American psychiatry: a personal perspective.
Arlington, VA: American Psychiatric Publishing, 2008.
37. Buck C, Llopis A, Najera E, Terris M. The Challenge of
Epidemiology: Issues and Selected Readings. Washington, DS:
Pan American Health Organization, 1988.
38. Dohrenwend BP. Psychology, psychologists, and psychiatric epidemiology. Acta Psychiatr Scand 1994;90:13–20.
39. Holland WW, Olsen J, Florey C (eds). The Development of
Modern Epidemiology: Personal Reports from Those Who Were
There. Oxford, UK: Oxford University Press, 2007.
40. Blashfield RK. The Classification of Psychopathology: NeoKraepelinian and Quantitative Approaches. New York: Plenum,
1984.
41. Meehl PE. Psychodiagnosis; Selected Papers. Minneapolis, MN:
University of Minnesota Press, 1973.
42. Zubin J. The revolution in psychopathology and its implications
for epidemiology. Paper presented at the Symposium on the
Public Health Approach to Mental Research. Annual Meeting of
the American Psychopathological Association, 3 September
1960. Chicago, IL, 1960.
43. Gordon JE, O’Rourke E, Richardson FLW Jr, Lindemann E. The
biological and social sciences in an epidemiology of mental
disorder. Am J Med Sci 1952;223:316.
44. Graham S. Social factors in relation to chronic illness. In:
Freeman HE, Levine S, Reeder LG (eds). Handbook of Medical
Sociology. Englewood Cliffs, NJ: Prentice Hall, 1963.
45. Paul JR. An Account of the American Epidemiological Society: A
Retrospect of Some Fifty Years. New York: Academic Press, 1973.
46. Rosen G. An American epidemiologist views his discipline. Yale
J Biol Med 1973;46:1–2.
47. Ross W. Comparative epidemiology of the mental disorders.
JAMA 1962;179:669.
48. Susser M. Community Psychiatry: Epidemiologic and Social
themes. New York: Random House, 1968.
49. Cooper B, Morgan HG. Epidemiological Psychiatry. Springfield
(IL): Thomas, 1973.
50. Susser M. The Columbia University School of Public Health
75th anniversary issue. Introduction. Am J Epidemiol 1998;147:
197–200.
51. Dohrenwend BP, Dohrenwend BS. Perspectives on the past and
future of psychiatric epidemiology. The 1981 Rema Lapouse
Lecture. Am J Public Health 1982;72:1271.
52. Dohrenwend BP. A Psychosocial perspective on the past and
future of psychiatric epidemiology. Am J Epidemiol
1998;147:222–31.
53. APA. Diagnostic and Statistical Manual of Mental Disorders (DSMIII). Washington DC: American Psychiatric Association, 1980.
54. Srole L, Fischer AK. The Midtown Manhattan Longitudinal
Study vs the Mental Paradise Lost doctrine: A controversy
joined. Arch Gen Psychiatry 1980;37:209–21.
55. Srole L, Fischer AK. Debate on psychiatric epidemiology. Arch
Gen Psychiatry 1980;37:1424–26.
56. Weissman MM, Klerman GL. Psychiatric nosology and Midtown
Manhattan Study. Arch Gen Psychiatry 1980;37:229–30.
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
57. Weissman MM, Klerman GL. Debate on psychiatric epidemiology – reply. Arch Gen Psychiatry 1980;37:1423–24.
58. Srole L, Langner TS, Michael ST et al. Mental Health in the
Metropolis: the Midtown Manhattan study. New York: New
York University Press, 1962.
59. Demazeux S. Qu’est-ce que le DSM?: Genèse et transformations
de la bible américaine de la psychiatrie. Paris: Ithaque, 2013.
60. Insel T. Transforming diagnosis. NIMH director’s blog. 2013.
http://www.nimh.nih.gov/about/director/2013/transformingdiagnosis.shtml (13 May 2014, date last accessed).
61. Demazeux S, Singy P. The DSM-5 in Perspective. Philosophical
Reflections on the Psychiatric Babel. Dordrecht, The
Netherlands: Springer, 2014 (forthcoming).
62. Klerman GL. The significance of DSM-III in American psychiatry. In: Spitzer RL, Williams JBW, Skodol AE (eds).
International Perspectives on DSM-III. Washington DC:
American Psychiatric Press, 1983.
63. Susser M, Stein Z. Eras in Epidemiology: the Evolution of Ideas.
New York: Oxford University Press, 2009.
64. Burger H, Neeleman J. A glossary on psychiatric epidemiology.
J Epidemiol Community Health 2007;61:185–89.
65. Tohen M, Bromet E, Murphy JM, Tsuang MT. Psychiatric epidemiology. Harv Rev Psychiatry 2000;8:111–25.
66. Elmore JG, Feinstein AR. Joseph Goldberger: An unsung hero of
American
clinical
epidemiology.
Ann
Intern
Med
1994;121:372–75.
67. Hollingshead AB, Redlich FC. Social Class and Mental Illness: A
Community Study. New York: Wiley, 1958.
68. Leighton AH. My Name Is Legion: Foundations for a Theory of
Man in Relation to Culture. New York: Basic Books, 1959.
69. Sartorius N, Shapiro R, Kimura M, Barrett K. WHO International
Pilot Study of Schizophrenia. Psychol Med 1972;2:422–25.
70. Robins L, Regier D. Psychiatric Disorders in America: the Epidemiologic Catchment Area Study. New York: The Free Press, 1991.
71. Scull A. Madhouse: A Tragic Tale of Megalomania and Modern
Medicine. New Haven, CT: Yale University Press, 2007.
72. Torrey EF, Miller J. The Invisible Plague. New Brunswick, NJ:
Rutgers University Press, 2001.
73. Salmon TW. Immigration and the mixture of races in relation to
the mental health of the nation. In: White WA, Jellife SE (eds).
Modern Treatment of Nervous and Mental Diseases.
Philadelphia, PA: Lea and Febiger, 1913.
74. Pollock HM. Eugenics as a factor in the prevention of mental disease. Ment Hyg 1921;5:807–12.
75. Paul DB. The Politics of Heredity: Essays on Eugenics,
Biomedicine, and the Nature-Nurture Debate. Albany, NY:
State University of New York Press, 1998.
76. Eaton WW. Social facts and the sociological imagination: the
contributions of sociology to psychiatric epidemiology. Acta
Psychiatr Scand 1994;90:25–38.
77. Lawson A. A sociological and socio-anthropological perspective.
In: Williams P, Wilkinson G, Rawnsley K (eds). The Scope of
Epidemiological Psychiatry: Essays in Honour of Michael
Shepherd. New York: Routledge, 1989.
78. Susser E, Schwartz S, Morabia A, Bromet EJ (eds). Psychiatric
Epidemiology: Searching for the Causes of Mental Disorders.
New York: Oxford University Press, 2006.
79. Durkheim E. Suicide, a Study in Sociology. New York: Simon
and Schuster, 1951.
International Journal of Epidemiology, 2014, Vol. 43, Supplement 1
80. Snell P. From Durkheim to the Chicago school: Against the ‘variables sociology’ paradigm. J Class Sociol 2010;10:51–67.
81. Tracy SW. An evolving science of man: the transformation and
demise of American constitutional medicine, 1920-1950. In:
Lawrence C, Weisz G (eds). Greater Than the Parts: Holism in
Biomedicine, 1920-1950. New York: Oxford University Press,
1998.
82. Pilgrim D, Rogers A. The troubled relationship between psychiatry and sociology. Int J Soc Psychiatry 2005;51:228–41.
83. Guze SB. Nature of psychiatric illness: why psychiatry is a
branch of medicine. Compr Psychiatry 1978;19:295.
84. Wheaton B. The role of sociology in the study of mental health
and the role of mental health in the study of sociology. J Health
Soc Behav 2001;42:221–34.
85. Clements FE. Plant Succession: an Analysis of the Development
of Vegetation. Washington, DC: Carnegie Institution of
Washington, 1916.
86. McKenzie RD. The ecological approach to the study of the
human community. Am J Sociol 1924;30:287–301.
i66
87. Gordon JE. Epidemiology in modern perspective. Proc R Soc
Med 1954;47:564.
88. Gordon JE. The twentieth century – yesterday, today, and tomorrow (1920-). In: Top FH (ed). The History of American
Epidemiology. Saint Louis, MO: Mosby, 1952.
89. Taubes G, Mann CC. Epidemiology faces its limits. Science
1995;269:164–69.
90. Susser M. Does risk factor epidemiology put epidemiology at
risk? Peering into the future. J Epidemiol Community Health
1998;52:608–11.
91. MacMahon B, Pugh TF, Ipsen J. Epidemiologic Methods.
Boston, MA: Little, Brown, 1960.
92. Susser M. Causal Thinking in the Health Sciences: Concepts and
Strategies of Epidemiology. New York: Oxford University Press,
1973.
93. Lewontin RC. The analysis of variance and the analysis of
causes. Int Jf Epidemiol 2006;35:520.
94. Lewontin RC. Biology as Ideology: the Doctrine of DNA. New
York: Harper Perennial, 1992.