This document summarizes a study of diagnostic processes in a neuro-oncology clinic in Israel. The study observed clinic operations for 6 months through participant observation and interviews. It finds that brain tumor diagnoses involve complex interactions between multiple experts, including neuro-oncologists, radiologists, pathologists, and others. Diagnoses are negotiated and refined over time through mechanisms like hierarchies of expertise, sequencing of tests, negotiations between experts, consideration of peripheral symptoms, and pragmatism in treatment decisions. Diagnoses are also fluid and lifelong processes as patients receive ongoing testing and care.

1. Experts and ‘knowledge that counts’: A study into the world of brain
cancer diagnosis
Sky Gross*
The Department of Sociology and Anthropology, The Hebrew University of Jerusalem, Israel
a r t i c l e i n f o
Article history:
Available online 12 October 2009
Keywords:
Israel
Expertise
Epistemology
Diagnosis
Hospital
Fieldwork
Authoritativeness
Brain cancer clinic
a b s t r a c t
This paper provides a close, in situ look into the life of a neuro-oncology (brain cancer) clinic of a large
hospital in Israel, based on a six-month participant observation. It points to the many challenges involved
in the solidification of brain tumour diagnoses by different experts, and presents these epistemological
and practical complexities as they uncover in daily routine. The paper’s task is two-fold: first, to
underline the technological and epistemological grounds of ‘expertise’ in the medicoscientific practice of
diagnosis, and their roles in the assertion of expert authoritativeness; and second, to provide analytical
tools to approach the complexity of diagnostic processes, the potential frictions it may create, and the
related mechanisms of resolution. These mechanisms include Hierarchisation: ranking the relative val-
idity and reliability of the different sources of information, eventually prioritising reports from more
authoritative expertises (e.g. imaging reports would be considered more reliable than patients’
accounts); Sequencing: relying upon the temporal dimension, and defining the discrepancy itself as
a diagnostic sign (e.g. the degradation or amelioration of the disease); Negotiation: adjusting diagnoses
via a preliminary exchange between experts and a consequent ‘‘fine tuning’’ of the reports (e.g. radiol-
ogists being aware of clinical evaluations before finalising their reports); Peripheralising: turning to other
expertises to ‘‘explain away’’ symptoms that do not fit with a well established initial diagnosis
(e.g. asserting that a symptom’s source was orthopaedic rather than neurological); and pragmatism: using
information only as far as it provided sufficient grounds for treatment decisions, leaving ambiguities
unresolved. These five mechanisms are presented here in the context of the daily work of the clinic.
Ó 2009 Elsevier Ltd. All rights reserved.
The majority of medical texts represent brain tumours as
involving fairly clear-cut entities and categories of entities,
amenable to ‘‘discovery’’ through different practices. A closer look
into the life of a neuro-oncology clinic proves, however, that the
solidification of neuro-oncological objects is less than straightfor-
ward, in both epistemological and practical terms. This study
sought to provide an in situ understanding of the ways in which
diagnoses are handled in medical experts’ daily routine via an
immersion into the everyday practice of a wide range of experts.
The aim of this paper is two-fold: first, to underline the techno-
logical and epistemological grounds of ‘expertise’ in the medi-
coscientific practice of diagnosis, and their roles in the assertion of
expert authoritativeness; and second, to propose analytical tools to
approach the complexity of diagnostic processes, the potential
frictions they may create, and their mechanisms of resolution.
As suggested by such scholars such as Philip Brown (1990, 2008)
and Annemarie Jutel (2009), works on the social and cultural
aspects of diagnosis require framing within the field of a ‘sociology
of diagnosis’. Following this literature, this paper will attempt to
provide potentially useful definitions of epistemological issues
underlying the practice of diagnosis.
Methodology
Over recent decades, a new approach to the study of medi-
coscientific work has arisen, seeking closer attention to the
everyday practices of ‘making science’ (e.g. Latour & Woolgar, 1986;
Lynch, 1988; Pickering, 1992; Rabinow, 1996). Drawing on this now
considerable body of research, this paper will present an ethno-
graphic study which allowed for a sustained in situ observation of
the daily micropolitics of diagnostic practice.
In the first half of 2006, I followed the work of members of
a neuro-oncology outpatient clinic situated in a large hospital in
Israel. The team consisted of four neuro-oncologists (henceforth
NROs), a head nurse, a social worker, and a neuropsychologist. The
NROs were observed during routine consultations; in rounds in the
inpatient ward; and in the course of three weekly professional
* Tel.: þ972 2 6437381.
E-mail address: crosss@mscc.huji.ac.il
Contents lists available at ScienceDirect
Social Science & Medicine
journal homepage: www.elsevier.com/locate/socscimed
0277-9536/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.socscimed.2009.09.017
Social Science & Medicine 69 (2009) 1819–1826

2. conferences: one internal conference, one with the radiology
experts, and one with the neurosurgical team. I spent some twenty
hours a week at the clinic, where I gathered extensive field notes
(simultaneously translated and composed in English) along with
both formal interviews and casual exchanges with staff members,
patients, and close-ones. Forming a small share of the data
collected, formal interviews – 13 overall, 4 with NROs, 2 with family
members, and 7 with 5 different patients – were later transcribed
verbatim. The field data was then analysed around themes that
seemed to recur, namely, the dynamics of the exchange of infor-
mation among different team members and patients. In line with
‘‘grounded theory’’ methodological approaches, the analytical
scheme was built and refined in constant conjunction with the
empirical data, and the data itself came to be read with this scheme
in mind.
Being, in a way, the physician’s home ground, the hospital
constitutes a crucial setting for understanding the world of medical
experts (e.g. Anspach, 1987; van der Geest & Finkler, 2004;
Mizrachi, Shuval, & Gross, 2005). The lengthy work of negotiation
and clarification, namely with the head of the clinic, led to the
gradual gaining of the staff’s confidence. Finally, formal permission
from the hospital’s Helsinki ethics committee allowed me to begin
my work. By the end of this process, I was accorded access to the
innermost stages of the clinic. Anonymity is ensured by the use of
pseudonyms throughout, again, for both patients and staff.
Considering the level of intimacy I developed with the patients,
I felt more comfortable using first names as pseudonyms, while
referring to staff members in a manner that would reflect their
professional roles, rather than on ad-hominem aspects of their work
(i.e. Dr. X.).
In view of the intricate nature of the field of neuro-oncology and
the complex terminology used in interactions, I spent months prior to
my insertion in the field studying related medical texts. In fact, it
seems one can hardly communicate the ‘‘real life’’ work of the
professionals without referring, however critically, to these schemes.
I had to repeatedly make clear to patients that I retained neither
a medical nor counselling role in the clinic, especially as I wore, as
requested by the staff, the traditional white robe. However, as the
proper ethical directives were followed, and as cooperation was
ensured, this did not seem to eventually form any significant obstacle.
At the clinic: the diagnostic process
According to textbook medical knowledge, a brain tumour arises
out of the proliferation and multiplication of a single cell, which, for
reasons little understood, goes through molecular mutation. At the
histological level, these will reproduce, create their own blood
supply, and eventually become numerous enough to apply pressure
on adjacent nervous tissues. The tumour presents itself through
a neurological realm of signs and symptoms, which may include
epileptic seizures, total or partial paralysis, speech malformations,
visual disturbances, loss of sensation in the limbs or, in some cases,
personality changes. Reporting diagnostic information such as
biopsy and MRI studies, along with clinical impressions, allows NRO
to determine the locality of the tumour, its size, the types of cells
involved, and its degree of malignancy.
At the clinic, this process begins with the first encounter with
the patient. There is a fairly ritualistic intake of new patients, where
the NRO assesses previous findings, gathers a clinical history, and
performs a physical examination. If judged necessary, immediate
intervention is considered, albeit only once options are discussed in
conjunction with other experts, such as radiologists and neuro-
surgeons, and, if applicable (viz. the tumour is metastatic), with the
primary oncologist. In either case, treatment options (including,
usually in this order, neurosurgery, chemotherapy, radiotherapy,
clinical trials, and palliative care) are deliberated among a relatively
large array of biomedical professionals.
Routine assessments of cases are typically performed in the
course of three weekly conferences. First there is the NROs’ staff
meeting where each physician is expected to present the files of the
patients he/she has seen throughout the week, with his/her
colleagues offering their opinions and evaluations. The second
meeting proceeds at the MRI unit, where the radiologists go
through recent imaging tests while the NROs provide clinical
information on their patients. The former then report their
impressions, first orally to the staff (‘‘things look bad, tut tut tut’’),
and later in writing (‘‘a slight compression of midbrain structures was
observed’’). The third meeting, the ‘tumour board’ –which includes
the NROs, neurosurgeons, and radiotherapists- seeks a collective
appraisal of the applicability of different surgical or radio-
therapeutic protocols (‘‘treatment algorithms’’). Its place within the
diagnostic complex is limited, and the approach is largely task-
oriented. The only common participants in these three sessions are
the NROs, a fact that reinforces their centrality in the diagnostical
process. Indeed, as the integration of reports into a compiling
diagnosis requires the adoption of a common way to ‘speak of’ the
disease, the participants (i.e. the reporting experts) are expected to
conflate their way of ‘speaking of’ the disease to the NROs, thus
placing the latter in a position of significant power (Mol, 2002).
The interaction between professionals is critical both before and
after initial diagnosis. Patients attend the clinic up to once a week
while under a radiotherapeutic regime, and once a month when
under chemotherapy – typically referred to imaging tests every few
weeks. In fact, for a brain cancer patient, diagnosis is literally
a lifelong process; For the NRO, each diagnostic function defines
a point on a temporal line which presumably depicts a logical
evolution of the disease, usually towards a worsening in all diag-
nostic values. The ‘‘case’’ thus remains in a certain state of fluidity,
where each re-assessment demands further exchanges of ‘‘reports’’
through the web of expertise.
Medicoscientific diagnosis
With the proliferation of diagnostic technologies, the practice of
medicine has become increasingly dependent on expert medi-
coscientific observation, adding to the intricacy of medical work
(Dew, 2001; Howell, 1995). This is accompanied by the need for
a constant reshaping of the definition and conceptualisation of
disease, most prominently in fields where diagnosis remains
complex and critical (Casper & Koenig, 1996; Clark & Mishler, 1992;
Reiser, 1978).
Facing sets of signs and symptoms, physicians are expected to
‘‘reveal’’ the one underlying element: ‘the Disease’. This quest aims
at a reconciliation of accumulated diagnostic data (e.g. clinical
examinations, blood test results) and the preconceived nosological
scheme of knowledge regarding diseases’ characteristics, expected
course, and likelihood to respond to certain treatments (Foucault,
1963; Lynch, 1988; Moreira, 2000). This integration of pathology
and nosology requires a ramified process of creating, changing,
communicating, and alternating definitions of ‘‘the Disease’’ as one
epistemological object (e.g. Berg & Mol, 1998; Mol, 2002; Moreira,
2004). In many biomedical settings, this process involves a multi-
plicity of experts: within each expertise, diagnostic elements are
observed, explained, assigned attributes and boundaries within its
specific ‘‘styles of reasoning’’ entrenched within formal and tacit
knowledge (e.g. Bos, 2004; Good, 1994). As Polanyi (1998) and
others (e.g. Collins, 2001; Hacking, 1982) have shown, expert con-
ceptualisation of pathologies is further mediated by technology
(e.g. an MRI machine), a form of attention (e.g. the analysis of the
spatial characteristics of a tumour) and a hermeneutic agent
S. Gross / Social Science & Medicine 69 (2009) 1819–18261820

3. (i.e. what does this finding mean?) (Berg & Harterink, 2004; Fou-
cault, 1963). This ‘gaze’ then provides further interpretation of
reality within these styles of reasoning (Haraway, 1997; Latour,
1993; Latour & Woolgar, 1986), and mediates the definition of the
disease as a thing-out-there, an object (e.g. Boumans, 2004; Koenig,
1988; Rabinow, 1996). Different forms of expertise are sustained
within their own technological, rhetorical, and epistemic systems
(Goodwin, 1994; Knorr-Cetina, 1999), each creating what shall be
referred here as ‘‘expert-mediated objects’’. These will then be
made comparable and manipulable within an established noso-
logical order, where they can be ‘‘spoken of’’ in a common language.
They will they be placed within a delineated, organised conceptual
chart of identified pathologies (Hacking, 1992; Lynch, 1985), and
made amenable to integration or alignment with other objects.
Both the ontological status and communicable qualities of these
mediated-objects are maintained by professional rhetoric in the form
of reports, i.e. the artefactual products of processes of observation and
translation (Anspach, 1987; Bos, 2004; Rose, 1996; Young, 1981).
Reports may be as numerous as the number of expertises involved,
and create as many objects as the number of gazes laid upon it
(Beaulieu, 2002; Berg & Mol, 1998; Mol, 2002). While mediated-
reports may ‘‘make sense’’ in the context of the single expertise,
epistemological inconsistencies may arise once juxtaposed, often
demanding an implicit evaluation of the forms of authoritativeness
associated with each of its sources. As this work will seek to
demonstrate, neuro-oncologists – being ultimately in charge of the
diagnosis – tended to strive towards the assertion of the nosopatho-
logical uniqueness and integrity of the ‘‘brain tumour’’ object. With
thisaim, theywereseentoapply–oftenimplicitlyandalwayssubtly–
several mechanisms, which will be discussed later in the text.
The web of expertise
In order to create the neuro-oncological object as a coherent
‘‘ontologically asserted’’ referent (Bos, 2004), the NRO has to draw
beyond his/her own field of expertise well into the other areas of
knowledge and technology. Within the scheme proposed here, this
involved exchanges between what I shall define as three relatively
distinct forms of diagnostic expertise, namely reporting, indepen-
dent, and compiling.
Reporting experts provide expert-mediated objects aimed at
assisting in the overall assessment of each particular case. These
included several professionals, including the histopathologist, radiol-
ogist, clinician, and neuropsychologist. In Betty’s case, forexample, her
fileshowsthattheradiologistreportedonthediseaseasanobjectwith
measurable spatial characteristics (‘‘a two centimeters neoplasm
located in the left parietal lobe’’), the histopathologist, on a qualitatively
distinctobject (‘‘a mixed-cell oligodendrogliomawith a necrotic centre’’),
the clinician, on a set of observable patterns (‘‘a contra-lateral hemi-
paresis with frequent grand-mal seizures’’), and the neuropsychologist
on a series of scores along different scales (‘‘a normal IQ test, low
tendenciestowardsanxiety; normal resultsinthecapacityto comprehend
abstract notions’’). The fact that these expert-mediated objects are
communicated in different forms seldom challenged either the
ontological status or the pathological definition of ‘‘the Disease’’.
However, reaching a sensible overall diagnosis did require, as shall be
shown, translation and juxtaposition of these heterogeneous reports.
Independent experts remain external to the neuro-oncological
sphere while potentially involved in the assessment of a patient’s
condition, depending on his/her known co-morbidities and general
health status. These experts (e.g. haematologists or general oncol-
ogists) served as autonomous physicians holding their own clinical
complex and definitions of disease. Still, while conducted inde-
pendently, their diagnostic and record-keeping activities
(including, typically, a separate file) were occasionally integrated
into the patient’s neuro-oncological case, although chiefly in its
margins. This placed them, in relation to the NROs, at the periphery
of the diagnostic web.
The compiling expert faces a multiplicity and disparity of
objects presented in reports gathered from independent and
reporting experts. Occasional disparity of mediated-objects,
however, could threaten the establishment of an integrated and
congruent image of the disease as a relatively well-defined entity.
This tension positioned the compiling expert (here, the NRO) at
a unique and most instructive junction which will be expanded
upon throughout the paper.
On the authoritativeness of reports
The report is both an artefact allowing communication, and
a channel through which independent and reporting experts
translate the observed disease into expert-mediated objects (Berg
& Harterink, 2004; Joyce, 2005). As suggested by Callon and Latour
(2006), it is through such translation processes that epistemologi-
cally distinct objects come to be linked together.
In practice, the translation (i.e. in the production of radiology,
histopathology, clinical, or other forms of reports) was often poor:
the reports seemed to serve more as means of asserting expert
authoritativeness than to communicate information among experts
or to patients. For example, the clinical report was handed over to
the patient at the end of the consultation, whether asked for or not.
As the reports were characteristically composed in a cryptic
language, the patient typically read and re-read it once having left
the room, often baffled by the swiftness of the consultation. ‘‘This’’,
one confused patient told me, ‘‘is a summary of my disease, so I know
whether I am better or not. I will show it to my children’’. She would
remain unsure of whether she was ‘‘doing well’’ or not, until having
the report deciphered for her.
The radiology reports were no more‘‘user-friendly’’. Forexample,
Noah’s MRI report read the following: ‘CVB 3, in MRS CHO/CR ¼ 4/3,
CR/MI ¼ 0.9’. This was obviously incomprehensible, even for Noah,
who usually tried to keep informed. Indeed, reports, say Good (1994)
and others (Beaulieu, 2002; Bourdieu, 1992) tend to use jargon and
rhetorics which make them appear more reliable, conceal areas of
uncertainty, and ultimately reify the experts’ position as holding
exclusive knowledge and thus epistemological authoritativeness.
This can be said of reports produced by each and every reporting
expertise at the clinic.
The sight of the tumour: radiology
The radiology report included several carefully operationalised
components, each conceivably contributing to the diagnostic effort:
the location of the tumour, the colouring of its centre and borders,
the number of foci, the regularity of its margins, etc. This, along
with the clinical and histopathology reports, enabled a grading of
the tumour, and a characterisation of its cellular bases.
This report played an important role in the continual exami-
nation and re-examination of the state of the disease. As far as the
team was concerned, it was arguably the most straightforward and
informative means of assessing the tumour’s progression. ‘‘I see
now.’’ or ‘‘It’s right there’’ were typical ways to refer to the radi-
ology-mediated object during radiology meetings. This has grounds
on several broader cultural stances, including the medi-
coscientifical tendency to regard sight as evidence, with ‘seeing’
being almost inseparable from ‘knowing’ (Draper, 2002; Dumit,
2004; Jenks, 1995; Joyce, 2006; Lynch & Woolgar, 1990) and vision
as an unmediated perceptual source of knowledge: the images
themselves purportedly allowing non-interventional access to the
truth (Cartwright, 1995; Daston & Galison, 1992; Foucault, 1963;
S. Gross / Social Science & Medicine 69 (2009) 1819–1826 1821

4. Knorr-Cetina & Amann, 1990). Once the visual is assumed to
provide the most authoritative access to the object of disease,
radiology would undoubtedly represent the epitome of diagnostic
science (Burri, 2008; Gunderman, 2005; Joyce, 2008).
‘‘Space’’ is another crucial element in the assertion of the
authoritativeness of radiological knowledge (Leigh Star, 1989).
‘‘Wow, it is indeed large, now I get [why the symptoms are as serious]’’,
or: ‘‘How can he even walk around with a tumour that big?!’’ were
types of phrases I would often hear in radiology meetings. The
diagnosis of a brain tumour, in the NROs’ own terms ‘a space
occupying process’, indeed involved an emphasis on spatialisation:
brain cancer is a local, non-systemic disease, unlikely to metastasise
to other parts of the body; the symptoms tend to be in correlation
with the surfaces of the brain affected by the disease; the site of the
symptoms often directly reflects the location of the tumour, its
spatial characteristics, and the rate of its spreading. Radiology, thus,
as a space-oriented measure, has powerful epistemological affinity
and mimetic relations with the disease as it is conceived: a spatial
diagnostic technique, for a spatially defined pathology.
Sorting things out: histopathology
Following a biopsy, the tumour specimen is brought to the
laboratory and assigned a degree of malignancy based on a set of
criteria. This assessment is extremely complex in terms of corre-
lating pathology with existing (and continuously growing) noso-
logical schemes. To mention only a few criteria of histopathological
categorisation: one form of gliomas, astrocytomas, may be either
well differentiated or anaplastic, either fibrillary, pilocytic, or
gemistocytic, with low or high degrees of cellularity, low or high
degrees of pleomorphism, etc. Patients, one must note, were
seldom aware of such intricate details regarding their tumours –
they would rather turn to plain binary terms: as either cancerous or
not, or about their cancer as either malignant or benign. These
sophisticated schemes of identification and classification are,
however, central to the creation and preservation of the medi-
coscientific ethos (Cicourel, 1964; Foucault, 1963; Reiser, 1991a;
Keating & Cambrosio, 2003). Histopathology’s ever-growing cate-
gories of mediated-objects, its high-tech procedures, and its posi-
tion at the cutting-edge of research contribute to its prestige as
providing an exclusive corpus of knowledge.
Another characteristic appears to be involved in this assertion of
authoritativeness. In sharp contrast to the ‘‘hands-on’’ clinical
evaluation, the removal of biological tissue and its analysis outside
of its human context allows ‘‘social distancing’’. The living person
becomes literally out-of-sight and, in a way, out-of-time, isolated
from subjective time and space (and thus from subjectivity itself) to
create alternative sterile, motionless, and well-defined areas where
scientific work can proceed. This, in turn, provides the laboratory
a distinction drawn on a halo of pure science, clear of the ‘fuzziness
of reality’ (Bourdieu, 1977; Foucault, 1963; Rosenberg, 1999;
Thompson, 1995). Removing the personal particularities of the
patient enables a more clear-cut alignment with existing nosolog-
ical categories, both within the realm of histopathology and within
the broader neuro-oncological compiling diagnosis.
Hands-on: the clinical report
The first component of the clinical exchange, the physical exami-
nation, comprises a direct, usually non-instrumented, assessment of
a number of parameters. According to textbook directions (which are
rarely meticulously followed in practice) these include the tone of
each muscle group, muscle power, and reflexes. Changes in sensation
will be identified by means of a probing of different parts of the body,
and coordination by observing body movements. Also central to the
examination is the evaluation of gait, during which the patient will be
asked to walk back and forth.
Thus, the clinician is expected to provide expert deduction and
gather relevant information from the patient, while drawing on
haptic skills and limited technological aids (Foucault, 1963;
Moreira, 2006; Porter, 1993). This ‘low-tech’ practice is believed to
be highly susceptible to errors and inaccuracies: the patient’s body
is viewed as disordered and ‘messy’ in conveying valuable data, and
the method itself ‘primitive’ (Joyce, 2005).
The second component of clinical practice, history-taking,
demands a systematic gathering of information (e.g. headaches,
seizures, visual disturbances, motor or sensory dysfunctions, pain,
nausea) which must be characterised by time of onset, duration,
location, frequency, severity, progression or remission, precipi-
tating circumstances, and associated complaints. A history of
administered treatments must also be sought, as well as a more
general background of co-morbidities, past medical interventions,
and cases of severe illness in the family.
The lack of neutrality, objectivity, and preciseness attributed to
this practice is critical in the evaluation of the validity – and thus
authoritativeness – of knowledge portrayed in such reports. Argu-
ably, patients have a ‘privileged access’ to their own inner world of
experience, and are able to reach a ‘truth’ that is beyond scientific
exploration (Gertler, 2001; Heil,1988). This impinges upon the ethos
of physician’s epistemological supremacy as holding total and
absolute knowledge of the phenomenon of disease. In order to retain
authoritativeness, the expert must bind his/her perspective to
observable and accessible aspects of the ‘reality’ he/she faces, rather
than to the patient’s life-world: his/her belief system, background,
personality, particularities, emotional issues, etc (Anspach, 1988;
Cicourel, 1983; Gunderman, 2005; Mishler, 1984). The patients
themselves cannot be assigned ‘‘expertise’’, a clear form of power,
even in the context of reporting their own subjective experience:
doing so would challenge the hierarchical physician/patient rela-
tions in the sacred biomedical space of the clinic. While the patients
may influence the actual process of diagnosis (e.g. overemphasise
some minor symptoms in order to assure full attention and treat-
ment), they are denied any status of expertise: the expert is the
‘‘reader’’ of the body and the deciphering agent of the narrative.
In written reports, patients’ accounts tended to be preceded by
a sort of ‘‘disclaimer’’, noting their relative unreliability. Typically,
the reports included clauses such as: ‘‘according to the patient...’’ or,
quoting from one clinical report: ‘‘[the patient] describes events of
abnormal smells, tastes, and panic feeling up to 10 times a day’’, or:
‘‘he complains of some difficulty getting his words out in conversa-
tions’’. A similar trend was found in oral exchanges: the ‘‘had’’ – as
in "the patient had headaches"- was presented during meetings as of
higher authoritative status than the ‘‘described’’, and –to a lesser
degree– than the ‘‘complained of’’. The first was assigned some
legitimacy as the reporting physician ipso facto provided a seal of
approval to the actuality of the symptom. The second portrayed the
patient as able to observe, reflect on, and effectively communicate
the nature of the symptom by offering a description, rather than
speak of a vague, emotionally-laden complaint. These variations,
however subtle, provided clues not only as to the authoritativeness
assigned to clinical knowledge, but also to the conceived credibility
of different forms of clinical information.
Figuring it out: neuropsychology
Unlike the psycho-oncologists, whose task was to provide patients
emotional support, the neuropsychologist was regarded as a diag-
nostic source. She was assigned patients that seemed to suffer from
cognitive or emotional deficiencies related to the organic effects of the
tumour, or related to side-effects of the treatment provided for the
S. Gross / Social Science & Medicine 69 (2009) 1819–18261822

5. tumour. Arguably with the aim of establishing medicoscientific status,
the neuropsychological evaluations tended to be presented as highly-
detailed statistical reports. However, such knowledge remained
illustrative rather than demonstrative and the neuropsychologist was
seldom consulted again, probably based on the low authoritativeness
of psychological methods in the medicoscientific complex (Hacking,
1992; Porter, 1995; Reiser, 1991b; Rose, 1985).
The meta-diagnosis: does it matter?
As a rule, the neuro-oncology clinic saw patients only once
a brain-tumour diagnosis was reasonably established. At this point,
however, the diagnosis remained vague: ‘‘there is a finding in your
brain, we are still unsure of its exact nature’’ was a classical opening.
Patients rarely demanded further detail, and, more often than not,
avoided using the term ‘‘cancer’’, but rather referred to hazy notions
such as ‘‘my condition’’, or ‘‘the lump’’. Oren, for example, was
surprised to hear me refer to the ‘‘oncology clinic’’ when we set
a location for an interview. He said he never noticed it was a clinic
that ‘‘also deals with cancer’’- this after five years of follow-up.
Recently diagnosed patients tended to use the more hesitant ‘‘I
was diagnosed with.’’ than the definite ‘‘I have.’’. Both patients and
physicians entered a sort of unspoken pact, where the precise diag-
nosis was regarded by both as tentative. In line with the ‘‘pact’’, the
physician would respond with such uninformative remarks as: ‘‘let’s
not jump our horses’’, ‘‘let’s talkafterwehavesomefurther testsdone’’,or
‘‘we’re never sure ofanything, but we have to dowhat needs to be done’’.
In the course of more than one hundred such consultations
observed, only a few patients inquired into expected consequences
of their diagnoses: the ‘‘how long do I have, doc?’’ question was never
raised, under any variation. Prognostic information was only
vaguely sought, with a question such as ‘‘can this be treated?’’,
rather than ‘‘will I live?’’. Physicians themselves never went beyond
a general: ‘‘it’s all just statistics’’, or ‘‘each case its own’’. While family
members occasionally sought online information on the expected
lifespan of their loved-ones (‘‘for practical reasons’’, they would
typically claim), as far as patient-physician interaction was con-
cerned, precise diagnosis did not serve any prognostic aim.
Mechanisms of integration
The raison d’eˆtre of neuro-oncological expertise lies in the
contention that brain cancer is a distinct, well-defined, and
conceptually unified disease entity. Asserting the coherence
between mediated-objects not only reaffirmed the soundness of
each individual form of inquiry and validated the meta-diagnosis
of one particular case, but also served to support the general notion
of brain cancer as a cellular based, spatially spread, and scientifically
detectible form of disease, one for which a specialised compiling
expert should be ascribed. In terms of the web of experts, the hier-
archical distinction between neuro-oncologists and general oncol-
ogists was continuously sought to be preserved not only by the NROs
forming a ‘‘clique of their own’’ as one onco-psychologist put it, but
also via the establishment of disease categories and diagnoses
pointing to the particularities of brain cancer over other oncological
diseases, and the affirmation of the diagnosis as requiring neuro-
oncological, rather than oncological, expertise. When faced with
a collection of disparate reports, the NROs thus sought to establish
the case as of a well-asserted neuro-oncological nature. In this sense,
the meta-diagnosis also served to delineate, and then preserve the
areas of jurisdiction and epistemological boundaries of the
compiling expertise (Abbott,1988; Gieryn,1983,1999). This secured
the NRO’s prerogative to elect and determine treatment options, to
provide reliable prognosis, and to centralise documentation and
records for future research: to be the compiling expert. This
endeavour involved five relatively distinct mechanisms: hierarch-
isation, sequencing, negotiation, peripheralising, and pragmatism.
Hierarchisation
Forms of knowledge, as well as the objects produced by this
knowledge, are created and sustained by having its different forms
become authoritative, and others subjugated, marginalised, or
excluded from the sphere of legitimacy (Foucault, 1980; Latour &
Woolgar,1986; Rose,1996). As shown throughout the text, the forms
of expertise associated with neuro-oncologic diagnosis differed not
only in terms of instruments and types of knowledge applied to the
mediated-object, but also in terms of their perceived authoritative-
ness and, consequently, hierarchical position. The NRO, confronted
with the task of incorporating the disparate reports, could (and often
did) choose to constrict the range of available knowledge by dis-
carding information from less authoritative sources, settling for
a diagnosis based on narrower – albeit more solid – foundations.
This was the case with Tal. At the radiology meeting, his latest MRI
took Professor Tamir by surprise: ‘‘this looks bad.’’, she told her
colleagues, ‘‘it’s amazing, he has no symptoms! This guy walks around
with a ticking bomb in his head, but is completely asymptomatic.’’. In
other words, the MRI suggested a progression, while the clinical
picture pointed to a more stable and benign condition. The pictures
were unquestionably there, thus the ‘‘anomaly’’ had to be located at
the clinical level, the less authoritative of the two forms of knowledge.
Tamir declared a worsening inTal’s symptoms: his chemotherapeutic
treatment was discontinued, as it was now judged ineffective.
Sequencing
Brain tumours are conceived as having a dynamic constitution.
First, it is granted that the tumour may progress from one grade to
another, may respond to therapy – at times long after the treatment
itself has been interrupted or concluded– or may simply remain
unchanged. Moreover, symptoms tend to accumulate and aggravate,
and their response to medication may decrease over time. Sequencing
uses this set of temporal factors to account for diagnostic changes
without challenging the unity of ‘‘the Disease’’, substituting ontolog-
ical fluidity for temporal fluidity: the object may not be stable as an
entity, but still retains a continual consistency as a neuro-oncology-
mediated object.
On one occasion, for example, David’s wife called Professor Berger
toreportthatherhusbandhad anewsymptom:hefelt‘‘tingling’’ inhis
left arm. She was surprised as the MRI, taken two weeks earlier, sug-
gested a shrinking of the tumour. Considering the authoritativeness
assigned to medicalimages,Berger did not readily discard the imaging
report as unreliable. Rather, she answered that the disease had prob-
ably worsened since. As MRI images were not compared unless taken
at least one to two months apart, this was an assertion that, while
sensible, was impossible to be put to the test, and thus also impossible
to challenge. Yet, as it allowed the initial diagnosis to remain valid, it
was accepted, at the very least until other evidence came along (the
next MRI suggested that the tumour was indeed growing).
Negotiation
It is during the weekly radiology meeting that ‘negotiated
processes’ (Moreira, 2006) could be observed. There, the four NROs
would sit behind a radiologist who gave ‘‘live’’ interpretations of the
images. Led by the NROs’ instructions, he focused on particular sites,
paid attention to specific patterns, all the while well aware of ‘what it
is we are looking for’. The clinical and the radiological were often –
again, never deliberately – adjusted ‘on the spot’, that is, well before
the written report was issued.
S. Gross / Social Science & Medicine 69 (2009) 1819–1826 1823

6. On some occurrences, the NROs were provided with non-negoti-
ated reports. These, compared to reports resulting from the MRI-NRO
meeting, tended to be vague and obscure. One such report was Ben-
ny’s. According to Professor Tamir, the clinical picture suggested
a slowing down of the tumour’s progression, while the MRI report
remained ‘‘incomprehensible’’. She advised Benny to phone her on
the subsequent Monday for a more definite answer. In the course of
Sunday’s meeting, Tamir briefed the radiologist on Benny’s clinical
improvement, emphasising the fact that he had recently been
through treatment. The highly negotiable nature of the imaging
report permitted the team to agree that the MRI showed some
decrease in the tumour’s size.
The use of this mechanism was also evident where clinical
reports were concerned. The NRO, one may recall, performed a dual
role in the web of expertise: once as a compiling expert and once as
a clinical reporting expert. The NRO, when clinically evaluating one
particular patient, could thus favour interpretations that were most
consistent with his/her overall preconceived impression. Hence, by
the time the file reached the NRO as the compiling expert, some
degree of compatibility could already have been incorporated into
the reports, facilitating the reaching of the meta-diagnosis.
Negotiation was less apposite where histopathology reports
were concerned. First, unlike MRI reports, which were delivered
orally, the histopathology report was customarily communicated
solely in writing. Secondly, histopathology was conceived, overall,
as involving a lesser degree of human interpretation, and thereby as
being more rigid to processes of negotiation. Thirdly, the NROs
tended to regard themselves as more competent in reading MRI
pictures than in the deciphering of cellular formations, thus having
little to add to the histopathologists’ interpretations. Finally, the
importance of the histopathology report was often only tangential
to decisions on treatment, and was conducted significantly less
frequently than radiology tests.
Peripheralising
Sarit, a 52 year old woman with a highly malignant cancer, was
toldbyProfessorBergerthat‘‘thelatestMRIshows[sheis] doingbetter’’.
Now almost completely blind, holding her face in her two hands, she
mumbled in response: ‘‘but my head aches so much that I can’t even
touch the tips of my hair without feeling like screaming!’’. Berger
interrupted her: ‘‘That, I can’t explain’’. As she attempted to persuade
him that her painwasreal,he referred her to a dermatologist: ‘‘itcould
be a different problem’’, he concluded. Here, as in other cases, the
independent expert was sought once a diagnostic element could not
be accounted for within the neuro-oncological sphere: the derma-
tologist could, however, explain away the ‘‘hair ache’’ without openly
dismissing the patient’s complaints on the one hand, or challenge the
already well-established diagnosis on the other.
This dermatology-mediated object could then be incorporated
into the view of the neuro-oncological disease, albeit in its
periphery. Such objects could then serve either as contextualisa-
tions of the meta-diagnosis or as evidence of its accuracy. As these
objects belonged to a sphere of expertise defined as ‘independent’-
i.e. extrinsic rather than intrinsic – the NRO held considerable scope
in their re-interpretation or, alternatively, their thrusting aside.
Overall, the place of independent-expert-mediated objects was to
a large extent assigned by the compiling expert, providing the NROs
a certain degree of slack in their meta-diagnostic work.
Pragmatism
In neuro-oncology, treatment options are comparatively scarce.
They include a few forms of chemotherapy, a limited array of
surgical interventions, and relatively little room for intense
radiotherapy. The primary brain tumour will almost invariably be
fatal to the patient: more than half will succumb to the disease
within the twelve months following diagnosis. Only one fifth will
be expected to survive for more than five years (Coldman, Phillips,
& Reid, 2004). In practice, once treatment options run out, the NRO
usually ceases to be involved. The patient will then be offered
supportive and palliative care, typically in a hospice or among his/
her relatives. This being the case, elaborate and minute diagnosis
may, in some cases, seem somewhat less crucial.
In effect, patients themselves were little interested in precise
diagnosis. Treatment prospects seemed more salient: ‘‘will I have to
go through chemo?’’ or ‘‘will I lose my hair?’’. Simon, for example,
spent months visiting the clinic without showing any interest as to
whether his tumour was graded III or IV. Although undoubtedly
aware of the grading system, he ‘‘never thought of asking’’. Indeed,
for many patients, beyond general notions of improvement or
degradation (‘‘has it grown?’’, ‘‘am I doing better?’’), formal diagnosis
remained futile.
This trend was also evident among another significant group of
patients. These were referred to the clinic after the treating
oncologist (here, the compiling expert) suspected the presence of
cerebral metastases. At this point, the neuro-oncologist (here,
serving as a independent expert) tended to provide an assessment
that was more quantitative than qualitative: the tumours were
either present or absent, single or multiple, active or stable. Patients
were then less concerned about the precise nature of the neuro-
oncological aspect of their disease. Also, as the presence of
metastases remained an almost definite death sentence, as far as
the NRO was concerned, a more pragmatic approach was adopted,
focusing on treatment options and palliative care rather than on
elaborate diagnostic undertakings.
In both cases, NROs often favoured the use of as little informa-
tion as necessary for the determination of the most advantageous
course of therapy. In practice, this stance of pragmatism typically
involved tacit avoidance of definite and elaborate diagnosis. Still,
while unspoken, an official diagnosis invariably appeared on the
reports, stating, for example, whether the tumour was defined as
a grade-IV or –III, whether a glioblastoma multiforme or an
anaplastic astrocytoma. This was in many cases, however, little
more than a bureaucratic formality, which the NROs regarded as of
weak empirical validity. In these cases, the NROs accepted the
fluidity of this definition and proceeded to consider treatment
options. Perhaps most interestingly, pragmatic considerations
could eventually have some bearing on the overall diagnostic
process: once a tentative object of disease was defined, other forms
of diagnosis could align. This process of alignment usually involved
reporting and independent experts: these would typically treat this
diagnosis as ascertained, being little aware of the looseness of its
epistemological grounds.
Concluding words
This study demonstrates the significance of the web of expertise
in the production of diagnosis, most particularly where the latter
tends to be ubiquitous. While widely acknowledged in past litera-
ture, little research was based on thick descriptions of expert
exchanges in a hospital setting. The ability to witness the processes
of diagnosis from various angles, including different actors and
different settings, allowed for a thorough and detailed presenta-
tion, and schematisation of the microdynamics of a central aspect
of medical work, i.e. the undertaking of diagnosis.
As defined above, the expert gaze incorporates sets of episte-
mologies and technologies, both of which mediate diagnosis and
reduce a multileveled phenomenon into one single manageable
object. The flow of artefacts, concepts, and interpretations among
S. Gross / Social Science & Medicine 69 (2009) 1819–18261824

7. experts were shown to have an important role in asserting both
reporting and independent expert authoritativeness, but also in
reifying the legitimacy of neuro-oncology as a compiling expertise.
Participant observation of the in-situ workings of the clinic sug-
gested that this reification involved, on the part of the compiling
experts, the use of five mechanisms of consolidation: hierarchisa-
tion, sequencing, negotiation, peripheralising, and pragmatism.
These mechanisms were shown to involve different cases at
different conjunctures, and with different sets of expert relations,
emphasising the complexity of diagnostic work.
Clearly, sociology has much to say on the issue of diagnosis and
the role of expertise in medical work (Brown, 1990, 2008; Jutel,
2009). I hope this paper will be able to provide some ground for the
development of a suggested sub-field of the sociology of diagnosis
in the social studies of medicine. A body of literature that would
integrate field studies and theoretical works can clearly profit from
the laying of a basic common conceptual grid, such as suggested
here.
References
Abbott, A. (1988). The systems of professions: An essay on the division of labor. Chi-
cago: The University of Chicago Press.
Anspach, R. R. (1987). Prognostic conflict in life-and-death decisions: the organi-
zation as an ecology of knowledge. Journal of Health and Social Behavior, 28(3),
215–231.
Anspach, R. R. (1988). Notes on the sociology of medical discourse: the language of
case presentation. Journal of Health and Social Behavior, 29(4), 357–375.
Beaulieu, A. (2002). Images are not the (only) truth: brain mapping, visual
knowledge, and iconoclasm. Science, Technology and Human Values, 27, 53–86.
Berg, M., & Harterink, P. (2004). Embodying the patient: records and bodies in early
20th century US medical practice. Body and Society, 10(2–3), 13–41.
Berg, M., & Mol, A. (1998). Differences in medicine: an introduction. In M. Berg, &
A. Mol (Eds.), Differences in medicine: Unravelling practices, techniques and
bodies. Durham, NC: Duke University Press.
Bos, J. (2004). Everyday life objectivity. Social Epistemology, 18(2–3), 123–138.
Boumans, M. (2004). The reliability of an instrument. Social Epistemology, 18(2–3),
215–246.
Bourdieu, P. (1977). Outline of a theory of practice. Cambridge: Cambridge University
Press.
Bourdieu, P. (1992). Language and symbolic power. Cambridge: Polity Press.
Brown, P. (1990). The name game: towards a sociology of diagnosis. The Journal of
Mind and Behavior, 11(3–4), 385–406.
Brown, P. (2008). Naming and framing: the social construction of diagnosis and
illness. In P. Brown (Ed.), Perspectives in medical sociology. Long Grove, Ill:
Waveland Press.
Burri, R. V. (2008). Doing distinctions: boundary work and symbolic capital in
radiology. Social Studies of Science, 38(1), 35–62.
Callon, M., & Latour, B. (2006). Le grand leviathan, s’apprivoise-t-il? Sociologie de la
traduction. Paris: Textes Fondateurs, Presses de l’Ecole des Mines.
Cartwright, L. (1995). Screening the body: Tracing medicine’s visual culture. Minne-
apolis: University of Minnesota Press.
Casper, M., & Koenig, B. (1996). Reconfiguring nature and culture: intersections of
medical anthropology and technosciences studies. Medial Anthropology Quar-
terly, 10(4), 223–236.
Cicourel, A. V. (1964). Method and measurement in sociology. New York: Free Press.
Cicourel, A. V. (1983). Language and the structure of belief in medical communi-
cation. In S. Fisher, & A. Dundas (Eds.), The social organization of doctor-patient
communication (pp. 221–240). Washington, DC: Center for Applied Linguistics.
Clark, J. A., & Mishler, E. (1992). Attending to patient’s stories: reframing the clinical
task. Sociology of Health and Illness, 14(3), 344–372.
Coldman, A., Phillips, N., & Reid, S. (2004). Cancer survival statistics – 2004. Canada:
The surveillance and outcomes unit, population and preventive oncology. BC
Cancer Agency.
Collins, H. M. (2001). Tacit knowledge, trust and the Q of Sapphire. Social Studies of
Science, 31(1), 71–85.
Daston, L., & Galison, P. (1992). The image of objectivity. Representations, 40, 81–128.
Dew, K. (2001). Modes of practice and models of science in medicine. Health, 5(1),
93–111.
Draper, J. (2002). ‘It was a real good show’: the ultrasound scan, fathers and the
power of visual knowledge. Sociology of Health and Illness, 24(6), 771–795.
Dumit, J. (2004). Picturing personhood: Brain scans and biomedical identity. Prince-
ton: Princeton University Press.
Foucault, M. (1963). Naissance de la clinique. Paris: Presses Universitaires de France.
Foucault, M. (1980). Power/knowledge: Selected interviews and other writings. New
York: Pantheon.
van der Geest, S., & Finkler, K. (2004). Hospital ethnography: introduction. Social
Science & Medicine, 59, 1995–2001.
Gertler, B. (2001). Introspecting phenomenal states. Philosophy and Phenomenal
Research, 63(2), 305–328.
Gieryn, T. F. (1983). Boundary-work and the demarcation of science from non-
science: strains and interest in professional ideologies of scientists. American
Sociological Review, 48(6), 781–795.
Gieryn, T. F. (1999). Cultural boundaries of science: Credibility on the line. Chicago:
University of Chicago Press.
Good, B. (1994). How medicine constructs its objects. In B. Good (Ed.), Medicine,
rationality and experience: An anthropological perspective. Cambridge, NY:
Cambridge University Press.
Goodwin, C. (1994). Professional vision. American Anthropologist, 96(3), 606–633.
Gunderman, R. B. (2005). The medical community’s changing vision of the patient:
the importance of radiology. Radiology, 234, 339–342.
Hacking, I. (1982). Language, truth and reason. In M. Hollis, & S. Lukes (Eds.),
Rationality and relativism (pp. 48–66). Oxford: Blackwell.
Hacking, I. (1992). ‘‘Style’’ for historians and philosophers. Studies in the History and
Philosophy of Science, 23(1), 1–20.
Haraway, D. (1997). Modest-witness@second-millenium. Femaleman(c)-meets-onco-
mouse: Feminism and technoscienceTM. New York: Routledge.
Heil, J. (1988). Privileged access. Mind (new series), 97(386), 238–251.
Howell, J. D. (1995). Technology in the hospital: Transforming patient care in the early
twentieth century. Baltimore, MD: John Hopkins University Press.
Jenks, C. (1995). The centrality of the eye in western culture: an introduction. In
C. Jenks (Ed.), Visual culture. London: Routledge.
Joyce, K. (2005). Appealing images: magnetic resonance imaging and the produc-
tion of authoritative knowledge. Social Studies of Science, 35(3), 437–462.
Joyce, K. (2006). From numbers to pictures: the development of magnetic
resonance imaging and the visual turn in medicine. Science as Culture, 15(1),
1–22.
Joyce, K. (2008). Magnetic appeal: MRI and the myth of transparency. Ithaca, NY:
Cornell University Press.
Jutel, A. (2009). Sociology of diagnosis: a preliminary review. Sociology of Health and
Illness, 31(2), 278–299.
Keating, P., & Cambrosio, A. (2003). Beyond ‘bad news’: the diagnosis, prognosis and
classification of lymphomas and lymphoma patients in the age of biomedicine
(1945–1995). Medical History, 47, 291–313.
Knorr-Cetina, K. (1999). Epistemic cultures: How the sciences make knowledge.
Cambridge, MA: Harvard University Press.
Knorr-Cetina, K., & Amann, K. (1990). Image dissection in natural scientific inquiry.
Science, Technology and Human Values, 15(3), 259–283.
Koenig, B. (1988). The technological imperative in medical practice: the social
creation of a ‘‘routine treatment’’. In M. Lock, & D. R. Gordon (Eds.), Biomedicine
examined. Dordrecht: Kluwer.
Latour, B. (1993). The pasteurization of France. Cambridge, MA: Harvard University
Press.
Latour, B., & Woolgar, S. (1986). Laboratory life: The construction of scientific facts.
Princeton, NJ: Princeton University Press.
Lynch, M. E. (1985). Discipline and the material form of images: an analysis of
scientific visibility. Social Studies of Science, 15(1), 37–66.
Lynch, M. E. (1988). Sacrifice and the transformation of the animal body into
a scientific object: laboratory culture and ritual practice in the neurosciences.
Social Studies of Science, 18(2), 265–289.
Lynch, M. E., & Woolgar, S. (Eds.). (1990). Representation in scientific practice. Cam-
bridge, MA: MIT Press.
Mishler, E. (1984). The discourse of medicine: Dialectics of medical interviews. Nor-
wood, NJ: Ablex.
Mizrachi, N., Shuval, J. T., & Gross, S. (2005). Boundary at work: alternative medicine
in biomedical settings. Sociology of Health and Illness, 27(1), 20–43.
Mol, A. (2002). The body multiple: Ontology in medical practice. Durham, NC: Duke
University Press.
Moreira, T. (2000). Translation, difference and ontological fluidity: cerebral angi-
ography and neurosurgical practice (1926–1945). Social Studies of Science, 30(3),
421–446.
Moreira, T. (2004). Coordination and embodiment in the operating room. Body and
Society, 10(1), 109–129.
Moreira, T. (2006). Heterogeneity and coordination of blood pressure in neuro-
surgery. Social Studies of Science, 36(1), 69–97.
Pickering, A. (Ed.). (1992). Science as practice and culture. Chicago: The University of
Chicago Press.
Polanyi, M. (1998). Personal knowledge. London: Routledge.
Porter, R. (1993). The rise of physical examination. In W. F. Bynum, & R. Porter (Eds.),
Medicine and the five senses (pp. 179–197). Cambridge, NY: Cambridge Univer-
sity Press.
Porter, T. (1995). Trust in numbers: The pursuit of objectivity in science and public life.
Princeton, NJ: Princeton University Press.
Rabinow, P. (1996). Making PCR: A story of biotechnology. Chicago: University of
Chicago Press.
Reiser, S. J. (1978). Medicine and the reign of technology. New York: Cambridge
University Press.
Reiser, S. J. (1991a). The clinical record in medicine, Part I: Learning from cases.
Annals of Internal Medicine, 114(10), 902–907.
Reiser, S. J. (1991b). The clinical record in medicine, Part II: Reforming content and
purpose. Annals of Internal Medicine, 114(10), 980–985.
Rose, N. (1985). The psychological complex: Psychology, politics and society in England
1869–1939. London: Routledge.
S. Gross / Social Science & Medicine 69 (2009) 1819–1826 1825

8. Rose, N. (1996). Power and subjectivity: critical history and psychology. In
C. Graumann, & K. J. Gergen (Eds.), Historical dimensions of psychological
discourse. Cambridge: Cambridge University Press.
Rosenberg, C. E. (1999). Meanings, policies, and medicine: on the bioethical
enterprise and history. Daedalus, 128(4), 27–46.
Star, S. L. (1989). Regions of the mind: Brain research and the quest for scientific
certainty. Stanford, CA: Stanford University Press.
Thompson, J. B. (1995). The media and modernity. Cambridge, MA: Polity Press.
Young, A. (1981). The creation of medical knowledge: some problems in interpre-
tation. Social Science & Medicine, 15, 379–386.
S. Gross / Social Science & Medicine 69 (2009) 1819–18261826