International Journal of Medical Informatics 55 (1999) 87 – 101

www.elsevier.com/locate/ijmedinf

Patient care information systems and health care work: a

sociotechnical approach
Marc Berg *
Institute of Health Policy and Management, Erasmus Uni6ersity Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands

Received 9 November 1998; received in revised form 20 January 1999; accepted 25 January 1999

Abstract

Those who face the difficulties of developing useful patient care information systems (PCISs) often stress the
importance of ‘organizational issues’. Building upon recent sociological insights in the construction and use of
information technologies for (health care) work, this paper underscores the importance of these insights for the
development and evaluation of these systems. A sociotechnical approach to PCISs in health care is outlined, and two
implications of this empirically grounded approach for the practices of developing and evaluating IT applications in
health care practices are discussed. First, getting such technologies to work in concrete health care practices appears
to be a politically textured process of organizational change, in which users have to be put at center-stage. This
requires an iterative approach, in which the distinctions between ‘analysis’, ‘design’, ‘implementation’ and ‘evaluation’
blur. Second, a sociotechnical approach sheds new light on the potential roles of IT applications in health care
practices. It is critical of approaches that denounce the ‘messy’ and ‘ad hoc’ nature of health care work, and that
attempt to structure this work through the formal, standardized and ‘rational’ nature of IT systems. Optimal
utilization of IT applications, it is argued, is dependent on the meticulous interrelation of the system’s functioning
with the skilled and pragmatically oriented work of health care professionals. © 1999 Elsevier Science Ireland Ltd. All
rights reserved.

Keywords: Patient care information systems; Health care work; Sociotechnical approach; Electronic patient record; Medical
decision-making; System development; System implementation; Evaluation

1. Introduction putting a man on the moon had been’’ ([1] p.

‘Developing a comprehensive medical in-
formation system’, Morris Collen concluded
in his historical survey of medical informatics
in 1995, appears ‘‘a more complex task than
* Tel.: +31-10-4088555; fax: + 31-10-4522511.
E-mail address: m.berg@bmg.eur.nl (M. Berg)
464). Although we hear much more about
successes, certain benefits and the need to
implement patient care information systems
(PCIS)1 in health care, the fact is that most
applications to date have failed. Large num-
bers of systems never make it off the drawing
table—and if they do, they do not appear to
be transportable out of the specific context in

1386-5056/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.
PII: S 1 3 8 6 - 5 0 5 6 ( 9 9 ) 0 0 0 1 1 - 8
88 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
which they were developed [1–7]. Although
precise data are lacking, a general handbook
on management information systems esti-
mated that from the large systems that end
up being used as much as 75% should be
considered to be operating failures. They
might be in operation, but they are too cum-
bersome, too expensive or too functionally
deficient to be even remotely called a ‘success’
[8]. And in a recent overview of clinical data
systems, Jeremy Wyatt mentions a staggering
figure of 98% of software built for US gov-
ernment use that was ‘unusable as delivered’
[9].
Those who are facing the enormous
difficulties of developing useful PCISs often
stress the importance of ‘organizational is-
sues’ [10,11]. Building upon recent sociologi-
cal insights in the construction and use of
information technologies for (health care)
work, this paper underscores the importance
of these insights for the development and
evaluation of these systems.2 A ‘sociotechni-
cal approach’ to PCISs in health care will be
outlined. After introducing the central start-
ing points of this empirically grounded ap-
proach (Section 2), two implications of this
approach for the practices of developing and
evaluating IT applications in health care or-
ganizations are outlined (Section 3). First,
getting such technologies to work in estab-
lished practices appears to be a politically
textured process of organizational change, in
which users have to be put center stage. This
requires an iterative approach, in which the
1
PCIS is a broader term than e.g. ‘electronic patient record’,
but no sharp terminological distinctions are intended here. All
these systems denote IT applications whose core users are
doctors, nurses and other health care professionals.
2
Development remains an important topic even in the cur-
rent era of off-the-shelf PCIS products. Such products are
never turnkey systems: they require extensive tailoring and
redesign to match local circumstances and needs.
distinctions between ‘analysis’, ‘design’, ‘im-
plementation’ and ‘evaluation’ blur. Second,
a sociotechnical approach sheds new light on
the potential roles of IT applications in
health care practices. It is critical of ap-
proaches that denounce the ‘messy’ and ‘ad
hoc’ nature of health care work, and that
attempt to structure this work through the
formal, standardized and ‘rational’ nature of
IT systems. Optimal utilization of IT applica-
tions, it is argued, is dependent on the metic-
ulous interrelation of the system’s functioning
with the skilled and pragmatically oriented
work of health care professionals. These im-
plications will and cannot take the form of
‘formulas’ for successful system development.
Rather, the sociotechnical approach is inher-
ently critical of overly methodical ‘guidelines’
promising success. The implications this pa-
per will sketch are intended to reorient the
way we tend to investigate and manage these
processes. It attempts to offer new perspec-
tives rather than a list of ‘critical success
factors’; it engages in constructive critique
rather than in delivering yet another set of
guidelines for design and implementation.
This does not mean that the sociotechnical
approach is ‘philosophical’ or merely aca-
demically, research oriented. To the contrary:
we claim that for a successful management
and realization of concrete development pro-
cesses, such an approach will ultimately
prove to be more useful.
2. The sociotechnical approach: starting
points
Some 20 years ago, the term ‘sociotechni-
cal system design’ was used to indicate design
approaches that stressed the importance of
job satisfaction, workers’ needs, and skill en-
hancement [12–14]. These approaches put
people and their working relationships center
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
stage and formed a long-needed antidote to
the technology-centered and top–down ap-
proaches that dominated system develop-
ment. In current times, the term has drifted
from this direct focus on workers’ emancipa-
tion. Embracing a user-oriented perspective,
sociotechnical approaches emphasize that
thorough insight into the work practices in
which IT applications will be used should be
the starting point for design and implementa-
tion [8,15,16]. This is especially important in
organizations that pivot around the work of
professionals [17]. Since any potential benefit
that IT might bring to health care has to be
realized at the level of the concrete interac-
tions with these tools, it is here that any
development or evaluation process should
start. Building upon previous research and
literature reviews [18–20], the specificity of
the sociotechnical approach to IT applica-
tions in health care is characterized in the
following three points.
2.1. Health care practices are seen as
heterogeneous networks
In the sociotechnical approach, work prac-
tices are conceptualized as networks of peo-
ple, tools, organizational routines, documents
and so forth [21,22]. An emergency ward,
outpatient clinic or inpatient department is
seen as an interrelated assembly of humans
and things whose functioning is primarily
geared to the delivery of patient care. The
work of doctors and nurses articulates with
the functioning of monitors, of order forms
and laboratory routines to keep an Intensive
Care patient stabilized, to treat an acute
traffic accident victim or to provide long term
care to a chronic diabetes patient. In per-
forming these primary goals, secondary aims
have to be served as well. Teaching interns,
doing clinical research and meeting budget
objectives are all to be achieved by many of
89
the same people and tools, and often as part
of the care process itself [23,24].
The tools, documents and machines are
‘constitutive’ elements of these work prac-
tices. If one would pull away even a simple
object such as the order form from an aver-
age Intensive Care Unit, that work practice
could not continue functioning in its current
complex and smooth manner [25]. The roles
and tasks of doctors and nurses are tightly
interwoven with each other and with the op-
erations of these other elements. Just like the
tasks and position of doctors have changed,
historically, with the coming of the stetho-
scope and the emergence of laboratory tests,
the current roles and tasks of health care
professionals are intertwined with the func-
tioning of record systems and the architec-
tures of their work environments [26–29].
Simple forms and room designs structure the
way their work is organized, the way their
responsibilities are distributed, and the nature
of the doctor–nurse relationship.
The elements that constitute these net-
works should then not be seen as discrete,
well-circumscribed entities with pre-fixed
characteristics. Rather, these entities acquire
specific characteristics, roles and tasks ‘only
as part of a network’ [30]. A ‘physician’ is
only a ‘physician’ in the modern western
sense because of the network of which s/he is
a part, and which makes his/her work and
responsibilities a reality. Without nurses,
record systems or the stethoscope, the Medi-
cal Doctor as we know it would and could
not exist. Because of this tight interrelation
between elements in a network, the introduc-
tion of a new element (as when a PCIS is
implemented), or the disappearance of an
element (as when a hospital stops training
junior residents) often reverberates through-
out the health care practice. The introduction
of the patient centered record in the US at
the beginning of the twentieth century, for
90 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
example, necessitated changes in hospital ar-
chitecture and the emergence of the new pro-
fession of medical record managers. It was
also tied to the changing position of the
physician from an enterpreneuring gentleman
to a methodical member of a team: it
changed record keeping from a private affair
to a matter of cooperation with (and poten-
tial criticism of!) colleagues [31,32]. Similarly,
shifting from paper-based recording practices
to electronic record keeping often makes
work practices (and especially the record-cre-
ating activities themselves) more visible, in-
spectable and manageable. This can have
large consequences for the interprofessional
relations within the hospital [33,34].
Yet although the effects and functions of
IT applications in work practices can be
highly consequential, the precise shape they
take is not pre-determined. They evolve
within the specific, socio-political contexts of
these practices. As Kling and Scacchi phrased
it, the introduction of technologies in work
practices is ‘embedded in a larger system of
activity, (has) consequences which depend on
peoples’ actual behavior, and (takes) place in
a social world in which the history of related
changes may influence the new change’ [21].
The sociotechnical approach, then, is hesi-
tant to speak of ‘organizational issues’ or
‘human factors’ in its analysis of IT applica-
tions in work practices. Although it is crucial
to stress the importance of the ‘social’ in a
predominantly technology-oriented environ-
ment [17,35], there is no distinct set of ‘issues’
that has to be ‘dealt with’ in developing IT
applications. Nor are there pre-fixed ‘human
factors’ that inevitably come to play. In the
practices that are at stake, technologies and
humans are closely interwoven, with more or
less frictions, aligned towards the perfor-
mance of common tasks. Which ‘issues’ and
‘factors’ are at stake depends on the network
in question. When the network changes, the
‘issues’ and ‘factors’ are transformed as well.
An IT application in a health care practice
should be seen as forming a seamless web
rather than a Technology in an Organization
[22]. In addressing, managing and studying
this network, one should not attempt to pry
it apart in a ‘social’ and a ‘technical’ system.
‘Technology’ and ‘organization’ do not oc-
cupy separate domains or operate according
to separate logics; nor does their relationship
develop in some unilinear way (the former
‘causing change’ in the latter or vice versa).
The interrelated elements constitute an as-
sembly that should be dealt with as a whole
rather than as a Technical subpart for the IT
engineers and a Social subpart for the social
scientists [2,21,36,37].
2.2. The nature of health care work
The core activity of health care work prac-
tices is ‘managing patients’ trajectories’: do-
ing investigations, monitoring, intervening
and re-intervening in order to at least tempo-
rarily cure or palliate patients’ problems
[2,24]. In all but a few instances, managing
patients’ trajectories is a collective, coopera-
tive enterprise. Even the individually operat-
ing general practitioner communicates with
his/her colleagues. A fundamental character-
istic of this work is its pragmatic, fluid char-
acter. Like other complex work activities, it is
characterized by the constant emergence of
contingencies that require ad hoc and prag-
matic responses. Although much work fol-
lows routinized paths, the complexity of
health care organizations and the never fully
predictable nature of patients’ reactions to
interventions result in an ongoing stream of
sudden events. These have to be dealt with on
the spot, by whomever happens to be present,
and with whatever resources happen to be at
hand [2,24]. In addition, and partly because
of this phenomenon, health care work is typ-
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
ified by ongoing negotiations about the na-
ture of the tasks and the relationships be-
tween those who execute the tasks [38,39].
The sociologist Hughes [40], for example, has
documented how experienced nurses often
help inexperienced residents by suggesting the
way towards the diagnosis, or by hinting
towards the necessary treatment. These are
subtle interactions: the doctors are formally
responsible for these actions, and their social
status disallows them to be too ignorant too
openly. Likewise, physicians in Dutch hospi-
tals often informally negotiate with nurses
that the latter order routine drugs, or give
intravenous injections, while they are ‘offi-
cially’ not allowed to do that.
Health care work is further characterized
by its distributed decision making, by ‘multi-
ple viewpoints’ and by its ‘inconsistent and
evolving knowledge bases’ [38,41,42]. During
the patient’s trajectory, many authors and
events exert their influence on the course of
events. There is rarely one individual who
truly oversees the whole chain of events and
decisions. More often, the shape of the trajec-
tory is the contingent result of small decisions
and steps taken by individuals from diverse
backgrounds and with varying viewpoints
about what is the case and what should be
done.
All these features of health care work point
towards the importance of the last character-
istic to be discussed here: the ‘articulation’
work that keeps such complex practices go-
ing. The sociologists Gerson and Star, speak-
ing about complex work activities in general,
describe this work as such:
all the tasks involved in assembling,
scheduling, monitoring, and coordinating
all of the steps necessary to complete a
production task. This means carrying
through a course of action despite local
contingencies, unanticipated glitches, in-
commensurable opinions and beliefs, or in-
91
adequate knowledge of local
circumstances. …[Since] no formal descrip-
tion of a system (or plan for its work)
can…be complete…every realworld system
thus requires articulation to deal with the
unanticipated contingencies that arise. Ar-
ticulation resolves these inconsistencies by
packaging a compromise that ‘gets the job
done’, that is, that closes the system locally
and temporarily so that work can go on.
[38] (cf. [24,43])
Although it can be seen as the ‘glue’ that
holds complex work practices together, artic-
ulation work tends to be paradoxically invisi-
ble to outsiders [44,45]. It does not result in
clear-cut ‘products’, it is not highly valued,
and it is generally not even recognized in
work descriptions or by managers. It will
come as no surprise that in health care, pri-
marily nurses and assistants perform this
work.
The sociotechnical view of ‘work’, then, is
at odds with traditional views of work preva-
lent within IT development in at least two
ways. First of all, it emphasizes the need to
address ‘cooperative work processes’ rather
than discrete tasks for individuals. Most dis-
cussions and concrete attempts in health care
informatics focus on the individual doctor or
nurse, and model his/her ‘decision making
process’ as if that could be depicted as a
sequence of logically distinguishable steps
[18,46,47]. Most concrete designs ‘seem to
encompass a rather restricted view of collab-
oration’: a design that assumes individuals
using the computer alone, for individual
tasks, which have to be completed before
another user might continue [48]. They might
focus on the importance of order entry, but
they will rarely focus on the way order entry,
for example, is intermingled with the ‘larger
process of assessing/reassessing patient
status…and developing the care strategy’
[49]. And they may pay much attention to the
92 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
development of a nicely looking clinical
workstation, but just how this workstation’s
functions will be used in meetings or by
residents that are constantly interrupted
rarely receives sufficient attention [49,50].
Second, the sociotechnical view of ‘work’
fundamentally undermines the idea that the
‘essence’ of work practices can be caught in
pre-fixed workflows, clinical pathways, for-
mal task descriptions or other formal models
[38,51]. Such descriptions are useful (see also
further), but it should not be forgotten that
they are only highly incomplete, summarized
and rigid depictions of the modeled work
practices. There is no a priori or algorithmic
connection between any particular formal de-
scription of an action and its specific occur-
rence [52]. Any concrete work activity only
unfolds ‘in the doing’, in constant interaction
with the contingent circumstances that make
up the situation in which it is located [51]. A
gynecologist confronted with a newly preg-
nant patient, for example, can leave the ‘stan-
dard’ path of actions because of a myriad of
reasons. The patient might be so insecure
that she foregoes some routine tests in order
to comfort her; she might skip a standard
blood test because she knows that a colleague
whom is also treating this patient will per-
form this test anyway; she might not ask
certain ‘standard’ questions because of a par-
ticularly painful history of a stillborn child.
The list is endless.
This issue is important to stress because
rationalist, technology-centered discourses
are still all-pervasive within our field. Such
discourses emphasize the messiness of current
work practices, the need to weed out variabil-
ity in practice, and the opportunities of
PCISs, protocols and other such tools to
finally bring ‘structure’ and ‘rationality’ to
the work of doctors and nurses. This issue is
complex and multifaceted, and cannot be sat-
isfactorily dealt with here [2,53–57]. Yet
given the dangers of importing too much
pre-fixed structure in health care work, and
given the resources that are currently spent
on creating more fine-grained, more directive
protocols for more aspects of this work, it is
imperative to continue querying what ‘good’
all these efforts will bring. In the light of
increasing evidence that too much structure
obstructs worktasks and puts additional bur-
dens on health care personnel [55,58,59],
careplans are being produced and promoted
as if ‘more is better’. In the light of abundant
sociological evidence as to information-rich-
ness of free text and the practical efficacy of
brief, handwritten notes [16,28,60], such prac-
tices are being scolded as ‘non-scientific’, ‘un-
professional’ and ‘outdated’ [61,62]. This is
not to say that there are no problems with
current practices, or that formal tools hold
no promise-far from it (see Section 3.2). Yet
as I will argue in Section 3, we will only be
able to fully realize these benefits when we
adapt a more empirically informed view of
the work of doctors and nurses, and a more
modest view of what these newest ‘solutions’
can bring.
2.3. Empirical orientation, with emphasis on
qualitati6e methods
It follows from the above characteristics
that the sociotechnical approach emphasizes
the importance of deep empirical insight into
the work practices in which an IT application
will be used. It is imperative to acquire in-
sight into the ongoing workflow and ‘negoti-
ated orders’ [63] of these practices before we
can even begin to consider developing (or
buying) a system. Similarly, we need to know
what the specific network that constitutes a
health care practice looks like before we can
think of realizable implementation strategies
or meaningful evaluation criteria. The socio-
technical approach is thus skeptical of ratio-
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
nalist models in which the existence of
common goals, predetermined tasks and a
limited number of formal procedures is
assumed and therefore found wherever one
looks [15]. Models may aim at universality,
at being generalizable over individual
health care practices, but this should be a
bottom – up exercise, generalizing from
empirical cases. It should not be a top–
down approach of creating universal ‘domain
information models’ in which all concrete
instances of actual health care practices
should fit. Nor should we loose ourselves
in trying to map the ‘basic structure of
medical knowledge’; to create a ‘foundation’
upon which then all individual instances of
medical action could be easily mapped. Such
exercises are fruitless at best—and all too
often result in stifling and rigid frameworks
[64].
The required empirical knowledge can
be made available in two ways, which
should both be used. First, end users should
be involved (see Section 3.1), and second,
qualitative research methods need to be em-
ployed. The use of interviews is generally not
adequate for the level of insight required:
ideally, participant observation forms the
starting point of any sociotechnical develop-
ment or evaluation process. It is difficult to
acquire a feeling for the intricate interrela-
tions between health care professionals and
(paper or electronic) documentation tech-
niques without having seen the work patterns
itself. Likewise, to get a grasp on the flows
and forms of information a health care
professional handles in a specific ‘patient
care scenario’ [49], we cannot limit our-
selves to interviews or surveys. Without
detailed, on-site insight, an adequate grasp
of what IT functionalities should be avail-
able in what form is practically impossible
[65].
93
Qualitative research methods are similarly
essential to any thorough evaluation of an IT
implementation. Simple quantitative mea-
sures (such as user satisfaction, usage indica-
tors, time studies) may be useful, but
they need to be grounded in qualitative data
so that their meaning can be understood.3
Why is satisfaction high for function A
but low for B? What is the reason for fluctu-
ating use times? Such questions can only
be thoroughly answered through detailed,
qualitative research [35,67]. The broad,
socio-cultural and political implications that
IT applications can have in health care
practices, moreover, is an additional argu-
ment for using qualitative evaluation meth-
ods. Since impacts can be wide-ranging and
unpredictable, pre-set measurement instru-
ments often miss the most relevant changes
that take place, or loose track of the way
variables affect one another [68,69]. Finally,
qualitative research methods are the most
suitable way to study changes in tasks, roles
and responsibilities. We may do a time-study
of documentation practices, but if the nature
of a nurse’s tasks changes, the time-study
looses most of its relevance. Typically, imple-
menting PCISs generates such changes, and a
central focus of the system’s evaluation
should be a thorough investigation of these
[2].
3
More complex quantitative measures — such as overall eco-
nomic return, or overall impact on health care outcomes—
have proven to be methodological nightmares, and have only
rarely yielded unequivocal results. Reasons are, amongst oth-
ers, the broad and diffuse impact of the implementation of IT
systems in organizations, and the state of general structural
change in which most health care organizations find them-
selves. In a network in which professionals’ tasks are changing,
new views on quality are emerging and cultural notions of
what proper documentation means are transforming, it is
virtually impossible to distill and causally link the influence of
one fixed variable — the IT application — on another. See Ref.
[66] for a fair juxtaposition of quantitative and qualitative
evaluation methods.
94 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
3. Implications of a sociotechnical approach
Adoption of a sociotechnical approach has
implications for a range of issues within the
development and evaluation of IT applica-
tions in work practices. It changes the way
we think about the very concept of medical
information [55,70,71], it transforms our view
of the way IT is embedded within organiza-
tions [72,73] and it problematizes any sugges-
tions of a ‘formula’ for successful system
development [3,11,37]. It is impossible to do
justice to all these implications in the scope of
this article. Here, two implications will be
developed in some detail: the merging of the
activities of development and evaluation in
an iterative, cyclical process, and the posi-
tioning of the formal IT application within
the fluid and pragmatic nature of health care
work.
3.1. An iterati6e process in which system
de6elopment and e6aluation acti6ities merge
Developing an IT application in health
care practices can never be a process of ‘sim-
ply installing and using a new technology’
[15]. The deep intertwinement between tech-
nological and human elements of the net-
works at stake implies that any design and
implementation attempt is necessarily related
to widespread transformations in these net-
works. Whether the system developers are
aware of this phenomenon or not, these pro-
cesses are inevitably political. They affect the
distribution of responsibilities and the hier-
archies between professionals and between
professionals and management [45,70,74].
Different groups might see the technology as
a way to achieve their potentially conflicting
goals [59,75], and the technology, conse-
quently, may embed values and assumptions
that are not shared by everyone [64,76,77].
The sociotechnical approach argues that
development projects should be seen and
managed as being the politically textured,
organizational change processes that they in-
evitably are. Closing one’s eyes for these real-
ities can only lead to failure through
‘resistance’ and even ‘sabotage’ by users who
are not taken seriously [3,17]. System devel-
opment projects, then, need to be user-cen-
tered. This implies more than a GUI, good
communication or adequate training pro-
grams. To obtain this support, to generate
commitment and to ensure user-driven design
and implementation, users need to be in-
volved early, thoroughly and systematically
[49,78,79]. This is easy to state as a slogan,
yet hard to achieve in practice: balancing the
do-ability of a project and the granting of
user-input is a very complex task. Preventing
conflicts between users to thwart the project,
preventing the project to become a mere
users’ wish list, and keeping an overall, co-
herent information strategy in place are just a
few of the challenges that await any such
project [8,80,81].
The work of PCIS analysis, design and
implementation, then, can itself be described
in the same terms as health care work has
been described above: it is a collective, coop-
erative enterprise, which is characterized by
the constant stream of contingencies. Ongo-
ing negotiations about the scope and aims of
the project, the role of the project manage-
ment and the responsibilities of the project
team members are the rule rather than the
exception. Here as well, decision making is
distributed, and work is guided by multiple
viewpoints and inconsistent and evolving
knowledge bases [38,82–84].
Because of the political nature of these
processes and the importance of the user, and
because of the fundamentally unpredictable
nature of these change processes, an iterative
approach to development is required. Start-
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
ing with central organizational needs,
systems have to be developed step by step,
so that the changes in technology and
work practices can evolve together [11].
Contrary to top–down design and implemen-
tation attempts, iterative approaches ac-
knowledge the impossibility to foresee
all consequences, and they can creatively
draw upon encountered problems or unantic-
ipated use in the further development of the
system. For traditional, top–down ap-
proaches, such inevitable effects can only ob-
struct the implementation and proper
functioning of the technology. Iterative ap-
proaches, on the other hand, allow for cre-
ative, organizational and technological
co-development.
Successful development processes, then, do
not emerge from attempts to install ‘hege-
monic’ systems whose rationality and power
subsumes all other elements. Rather, ‘power-
ful technical systems comprise…artful inte-
grations’ between working practices and new
and old devices [37]. Such integrations should
be allowed to gradually emerge rather than
brought about through enforced revolution
[39]. This is even true for such seemingly
technical and elementary issues such as stan-
dards (compare, for example, the success of
the pragmatically structured, flexible and bot-
tom – up developed Internet protocols with
the small impact of the formalistic, layered,
inflexible and top–down structure of the
ISO/OSI alternatives) [72,85,86]. ‘Growing’
such systems implies entering ‘into an ex-
tended set of working relations, of contests
and alliances’ [11,37]. In such a process, the
distinctions between system ‘analysis’, ‘de-
sign’, ‘implementation’ and ‘evaluation’ blur.
Strictly speaking, within a sociotechnical per-
spective it hardly makes sense to consider
them as separate activities [81,83]. Users are
involved from early on in the analysis, and
95
feedback from early implementations imme-
diately informs further analysis and design.
With design continuing during implementa-
tion, and ‘evaluation’ informing analysis and
design, ‘analysis’, ‘design’, ‘implementation’
and ‘evaluation’ become co-occurring activi-
ties. These processes should be organized and
managed as such, and the different expertise
required should be integrated in any PCIS
project team.
A modest and imperfect example of such
a process has been described in this journal
before [16]. In the adaption of a com-
mercial PCIS package for Intensive Care
departments to the specific needs of a IC
in a Dutch Research Hospital, IC nurses
and an anaesthesiologist from the depart-
ment in which the system would be placed
were trained so that they could take full
responsibility for the tailoring process. In
this way, the distinction between ‘users’
and ‘designers’ truly blurred, and ongoing
evaluations from real-time use were con-
stantly fed back into the ongoing tailoring
process. The pressure and presence of
real-time demands, and the ongoing co-
occurrence of design and evaluation continu-
ally ensured an artful integration of the sys-
tem with the surrounding working routines.
Part of the success of this implementation,
however, was due to the small scale of the
sociotechnical change: record keeping proce-
dures in the wards around the Intensive Care
remained unchanged. This obviously gener-
ated problems when patients were transferred
between wards. A larger sociotechnical
change might involve introducing similar
packages on these surrounding wards—yet
one problem that is encountered is that the
vendor of the IC PCIS package does not
provide ‘general ward’ packages that would
suit the needs of a surgical or internal
medicine ward. (See for other examples, e.g.
Refs. [11,49]).
96 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
3.2. The PCIS as a tool in health care work
What is it that IT applications can do in
health care work? If we ban all advertisement
rhetoric and tendencies towards hegemonic
systems, we can discern two modest yet po-
tentially powerful roles that PCISs can play
in health care work. Information technologies
enable professionals and organizations to ‘ac-
cumulate’ data-elements into meaningful
wholes and to ‘coordinate’ complex processes
of interaction and collaboration [87]. They
collect and aggregate data entered (into a
graph, list or possibly a reminder) and so
afford new levels of overview, and they link
activities of doctors and nurses across differ-
ent times and spaces [88]. IT’s potential to
take over these tasks is often greatly exagger-
ated. The active and efficient way in which
paper-based documentation technologies can
perform these roles is generally underesti-
mated. Also, even when IT applications are
omnipresent in a workpractice they only
touch upon a part of the information han-
dling and coordination that takes place
[18,87,89 – 91]. Yet PCISs obviously may per-
form accumulation and coordination func-
tions more powerfully than traditional paper
based documents could. IT applications can
bring their computational power to bear to
create summaries, graphs or reminders [92–
94], and fast electronic links can ensure coor-
dination between events taking place in
geographically separate locations.
In general terms, then, information tech-
nologies afford networks to span over larger
numbers of entities. Larger numbers of data
can be assembled and dealt with by the same
professional, and more events in more dis-
tinct spaces and times can be brought to-
gether. This is the generative power of formal
tools: in the interlocking of such tools with
human work activities, new competencies for
health care professionals can be produced,
and higher levels of complexity in work tasks
can be achieved [20,57,87,91].
Several important observations follow
from seeing the tool as being interrelated
with other elements in the performance of
overall tasks. First of all, conceptualizing IT
applications as tools that form ‘artful integra-
tions’ with health care professionals and
other instruments focuses the attention away
from overly ambitious attempts to ‘replace’
the paper record, or to ‘clean up’ medical
decision making. IT applications should be
developed so that their practical strengths
articulate optimally with the practical
strengths of health care professionals and, for
example, paper documents. This is a different
and more fruitful focus than attempts to
‘mirror’ doctor’s reasoning or to create a
‘paperless’ environment. Too often, designers
and managers mistakenly invest in ‘smart
machine’ solutions because they overestimate
the self-sufficient powers of IT, and overlook
the skills that are already present in the work
practice. In such circumstances, a more
‘dumb’ solution, with a retraining of already
rather skilled personnel, would often be
cheaper and easier [95]. Similarly, because
designers and managers underestimate the
richness and ‘ecological flexibility’ of paper
documents, useful resources and work rou-
tines are often destroyed and replaced with
much less flexible and more expensive ‘solu-
tions’ [48,95]. IT is better in repetitively
amassing and monitoring data than in mak-
ing patient-specific judgments about them. It
is excellent in fast transmission and search
capabilities, but it lacks paper’s ability to be
shuffled around, leafed through, and provide
overview [48,89].
Finally, this analysis returns us to the
seeming conflict between the fluid, coopera-
tive and necessarily ‘messy’ nature of work
practice and the formal, standardized and
comparatively rigid functioning of IT. As
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
stated in Section 2, one of the largest mis-
takes that PCIS developers can make is at-
tempting to ‘replace’ the messy and ad hoc
nature of that work with the straightforward-
ness and ‘rationality’ that the system seems to
promise. One of the most dangerous yet
widespread tendencies in the field of PCIS
development is taking the tool’s characteris-
tics as ideals to which the work practice
should be molded. Rather, the optimal uti-
lization of PCISs is dependent on the meticu-
lous ‘interrelation’ of the system’s functioning
with the skilled and pragmatically oriented
work of health care professionals. First of all,
formal tools such as PCISs only survive in
health care practices because of the skilful
work of health care professionals. Tools that
embed pre-fixed sequences of steps in a care
process, or that only allow certain modes of
data input would perish amidst the contin-
gencies and pragmatic needs that characterize
health care work—were it not for the balanc-
ing acts of health care professionals. These
professionals translate vague answers into
one of the preset ‘codes’ on the form, or
modify a proposed care path so that a patient
can spend a weekend with his family, or do
several diagnostic tests in one batch. They
often establish ‘workarounds’ to trick the
system so that it keeps on functioning with-
out interfering with the acute, practical situa-
tion at hand [96,97]. (One workaround, for
example, is a nurse resetting the computer’s
clock so that orders can be entered post hoc
[87]). Formal tools such as PCISs, then, para-
doxically seem to be kept ‘alive’ by the very
same ad hoc and pragmatic activities that
they are often set out to erase. The ‘irra-
tional’ practices that the ‘rationalizing’ tools
would smooth out actually are a ‘sine qua
non’ for the tools’ smooth functioning
[2,96,98].
The skilful activities of health care profes-
sionals keep PCISs functioning— and in
97
properly developed PCISs, the functioning of
these systems in their turn enhances the pro-
fessional’s responsibilities and competencies.
In well-designed systems, coding data and
working upon preset carepaths can generate
the possibility to do research, or alleviate the
burden of routine tasks [57,87]. In the IC
PCIS system described above, for example,
the system helped calculate the fluid balance
in a way which saved the nurses time and
gave them continuous access to the current
value of the fluid balance (whereas with the
paper system, the actual value would only be
calculated once every 24 h) [16]. Similarly,
several general practitioner PCIS systems
have well-developed, simple coding schemes
built in, so that the GPs can ‘code’ their
diagnoses and actions with a few simple
mouseclicks [94]. In a well-guided develop-
ment process, the application and the health
care professionals ‘enable each other to affect
each other’ in their mutual interactions.
When tools are carefully inserted in the prag-
matically structured activities of nurses and
doctors rather than seen as a ‘solution’ to this
‘messy’ nature, their skilful interaction with
the tool will afford it to exert, in its turn, its
generative power.
4. Conclusion
The sociotechnical approach cannot be
seen as a simple ‘solution’ to the many prob-
lems haunting PCIS development. It does not
automatically yield a list of superior system
requirements, nor does it answer the every-
day problems of a manager in charge of an
implementation project. The approach raises
several issues that have no easy answers: how
to find the optimal form for the iterative
development process in an environment full
of economic pressures for ‘fast results’, diver-
gent interests, and inflexible IT applications?
98 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
Where to find the optimal interrelation be-
tween the formal tool and the health care
professionals’ skills? Yet its view on health
care practices and health care work, and its
emphasis on empirical, qualitative insight
into these practices and processes will not
only improve our analytical understanding of
these phenomena. Approaching health care
practices as heterogeneous networks, and
recognizing the pragmatic and skilled charac-
ter of doctors’ and nurses’ work is a crucial
first step towards PCISs that will articulate
more powerfully and more artfully with their
surrounding networks. Seeing the PCIS as
the outcome of politically textured negotia-
tion processes, moreover, will help us face
this challenge up-front. It will also help us
recognize this technology for what it is: a tool
that may radically transform the shape and
structure of current health care practices.
Acknowledgements
I would like to thank Emilie Gomart, Els
Goorman and Berti Zwetsloot for the inspi-
ration and comments that led to this paper.
This research has been made possible by a
grant from the Netherlands Organization for
Scientific Research (NWO).
References
[1] M.E. Collen, A history of medical informatics in
the United States, 1950 to 1990, Am. Med. Infor-
matics Assoc. (1995).
[2] M. Berg, Rationalizing Medical Work. Decision
Support Techniques and Medical Practices, MIT
Press, Cambridge, 1997.
[3] C. Sauer, Why Information Systems Fail: A Case
Study Approach, Alfred Waller, Henley-on-
Thames, 1993.
[4] J.B. Chessare, K.E. Torok, Implementation of
COSTAR in an academic group practice of general
pediatrics, MD Comput. 10 (1993) 23–27.
[5] J.I. Levitt, Why physicians continue to reject the
computerized medical record, Minn. Med. 77
(1994) 17 – 21.
[6] P.B. Marr, et al., Bedside terminals and quality of
nursing documentation, Comput. Nurs. 11 (1993)
176 – 182.
[7] S.C. Schoenbaum, G.O. Barneu, Automated am-
bulatory medical records systems. An orphan tech-
nology, Int. J. Technol. Assess. Health Care 8
(1992) 598 – 609.
[8] K.C. Laudon, J.P. Laudon, Management Informa-
tion Systems. New Approaches to Organization
and Technology, 5 ed., Macmillan, New York,
1998.
[9] J.C. Wyatt, Clinical data systems, part 3: develop-
ment and evaluation, Lancet 344 (1994) 1682–
1688.
[10] N.M. Lorenzi, R.T. Riley, M.J. Ball, J.V. Douglas
(Eds.), Transforming Health Care Trough Infor-
mation: Case Studies, Springer, New York, 1995.
[11] C. Atkinson, V.J. Peel, Growing, not building, the
electronic patient record system, Methods Inf.
Med. 37 (1998) 206 – 310.
[12] R. Hirschheim, H.K. Klein, K. Lyytinen, Informa-
tion Systems Development and Data Modeling
Conceptual and Philosophical Foundations, Cam-
bridge University Press, Cambridge, 1995.
[13] E. Mumford, M. Weir, Computer Systems in
Work Design: the ETHICS Method, New York:
Wiley, 1979.
[14] A. Friedman, D.S. Comford, Computer Systems
Development. History, Organization and Imple-
mentation, Wiley, New York, 1989.
[15] R. Kling (Ed.), Computerization and Controversy:
Value Conflicts and Social Choices, 2nd ed., Aca-
demic Press, San Diego, 1996.
[16] M. Berg, C. Langenberg, I.V.D. Berg, J. Kwakker-
naat, Considerations for sociotechnical design: ex-
periences with an electronic patient record in a
clinical context, Int. J. Med. Inf. 52 (1998) 243–
251.
[17] N.M. Lorenzi, R.T. Riley, A.J.C. Blyth, G.
Southon, B.J. Dixon, Antecedents of the people
and organizational asprects of medical informatics:
review of the literature, J. Am. Med. Inform. Soc.
4 (1997) 79 – 93.
[18] M. Berg, Medical work and the computer-based
patient record: a sociological perspective, Methods
Inf. Med. 37 (1998) 294 – 301.
[19] M. Berg, The politics of technology: on bringing
social theory in technological design, Sci. Techn.
Hum. Val. 23 (1998) 456 – 490.
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
[20] M. Berg, Of forms, containers and the electronic
medical record: some tools for a sociology of the
formal, Sci. Techn. Hum. Val. 22 (1997) 403 – 433.
[21] R. Kling, W. Scacchi, The web of computing:
computer technology as social organization, Adv.
Comput. 21 (1982) 1–90.
[22] W.E. Bijker, J. Law (Eds.), Shaping Technology-
Building Society. Studies in Sociotechnical
Change, MIT Press, Cambridge, 1992.
[23] P. Atkinson, Medical Talk and Medical Work,
Sage, London, 1995.
[24] A. Strauss, S. Fagerhaugh, B. Suczek, C. Wieder,
Social Organization of Medical Work, University
of Chicago Press, Chicago, IL, 1985.
[25] M. Berg, Practices of reading and writing. The
constitutive role of the patient record in medical
work, Soc. Hea. Illn. 18 (1996) 499–524.
[26] S.J. Reiser, Medicine and the Reign of Technol-
ogy, Cambridge University Press, Cambridge,
1978.
[27] B.L. Craig, The role of records and of record-
keeping in the development of the modem hospital
in London, England and Ontario, Canada, c.
1890–1940, Bull. History Med. 65 (1991) 376 – 397.
[28] H. Garfinkel, Studies in Ethnomethodology, Pren-
tice–Hall, Englewood–Cliffs, NY, 1967.
[29] H. Tellioglu, I. Wagner, Work practices surround-
ing PACS: the politics of space in hospitals. (1997),
in press.
[30] B. Latour, Science in Action, Cambridge Univer-
sity Press, Cambridge, 1987.
[31] J.D. Howell, Technology in the Hospital. Trans-
forming Patient Care in the Early Twentieth Cen-
tury, Johns Hopkins University Press, Baltimore,
MD, 1995.
[32] R. Stevens, In Sickness and in Wealth: American
Hospitals in the Twentieth Century, Basic Books,
New York, 1989.
[33] J. Bowers, Making it work: a field study of a
‘CSCW Network’, Inf. Soc. 11 (1995) 189 – 207.
[34] E. Egger, I. Wagner, Negotiating temporal orders:
the case of collaborative time management in a
surgery clinic, Comp. Supp. Coop. Work 1 (1993)
255–275.
[35] B. Kaplan, Addressing organizational issues into
the evaluation of medical systems, J. Am. Med.
Inform. Soc. 4 (1997) 94–101.
[36] G.C. Bowker, S.L. Star, W. Turner, L. Gasser
(Eds.), Social Science, Technical Systems, and Co-
operative Work: Beyond the Great Divide,
Lawrence Erlbaum, Mahwah, NJ, 1997.
99
[37] L. Suchman, Working relations of technology pro-
duction and use, Comp. Supp. Coop. Work 2
(1994) 21 – 39.
[38] E. Gerson, S.L. Star, Analyzing due process in the
workplace, ACM Trans. Office Inf. Syst. 4 (1986)
257 – 270.
[39] G. Button, R.H.R. Harper, Taking the organisa-
tion into accounts, in: G. Button (Ed.), Technol-
ogy in Working Order. Studies of Work,
Interaction, and Technology, Routledge, London,
1993, pp. 98 – 107.
[40] D. Hughes, When nurse knows best: some aspects
of nurse/doctor interaction in a casualty depart-
ment, Soc. Hea. Illn. 10 (1988) 1 – 22.
[41] S. Timmermans, M. Berg, Standardization in ac-
tion: achieving universalism and localization in
medical protocols, Soc. Stud. Sci. 27 (1997) 273–
305.
[42] S.L. Star (Ed.), Ecologies of Knowledge: Work
and Politics in Science and Technology, State Uni-
versity of New York Press, New York, 1995.
[43] K. Schmidt, L. Bannon, Taking CSCW seriously:
supporting articulation work, Comp. Supp. Coop.
Work 1 (1992) 7 – 40.
[44] S.L. Star, The sociology of the invisible: the pri-
macy of work in the writings of Anselm Strauss,
in: D.R. Maines (Ed.), Social Organization and
Social Process: Essays in Honor of Anselm
Strauss, Aldine de Grnyter, Hawthome, 1991, pp.
265 – 283.
[45] I. Wagner, Hard times. The politics of women’s
work in computerized environments, Eur. J. Wom-
en’s Stud. 2 (1995) 295 – 324.
[46] J.M. Nyce, I.I.I. W. Graves, The construction of
knowledge in neurology: implications for hyperme-
dia system development, Al. Med. 2 (1990) 315–
322.
[47] T. Winograd, F. Flores, Understanding Comput-
ers and Cognition. A New Foundation for Design,
Ablex, Norwood, NJ, 1986.
[48] P. Luff, C. Heath, D. Greatbatch, Tasks-in-inter-
action: paper and screen-based documentation in
collaborative activity, in: J. Turner, R. Kraut
(Eds.), Proceedings of the Conference on Com-
puter Supported Cooperative Work, ACM Press,
New York, 1992, pp. 163 – 170.
[49] E.L. Drazen, J.B. Metzger, J.L. Ritter, M.K.
Schneider, Patient Care Information Systems: Suc-
cessful Design and Implementation, Springer, New
York, 1995.
[50] K. Henderson, Flexible sketches and inflexible
data bases: visual communication, conscription
100 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
devices, and boundary objects in design engineer-
ing, Sci. Techn. Hum. Val. 16 (1991) 448 – 473.
[51] L. Suchman, Plans and Situated Actions. The
Problem of Human–Machine Communication,
Cambridge University Press, Cambridge, 1987.
[52] M. Robinson, Double-level languages and co-oper-
ative working, AI Soc. 5 (1991) 34–60.
[53] S.L. Star (Ed.), The Cultures of Computing,
Blackwell, Oxford, 1995, p. 1995.
[54] M. Berg, Problems and promises of the protocol,
Soc. Sci. Med. 44 (1997) 1081–1088.
[55] M. Berg, E. Goorman, The contextual nature of
medical information, Int. J. Med. Inf. (1998) in
press.
[56] P.E. Agre, Formalization as a social project, Q.
Newslett. Lab. Comp. Hum. Cognition 14 (1992)
25–27.
[57] M. Robinson, ‘As real as it gets…’ taming models
and reconstructing procedures, in: G. Bowker, L.
Gasser, L. Star, B. Turner (Eds.), Social Science,
Technical Systems, and Cooperative Work,
Lawrence Erlbaum, Mahwah, NJ, 1997.
[58] E. Goorman, M. Berg, Modelling nursing activi-
ties: electorinc patient records and their discon-
tents, in press.
[59] G. Symon, K. Long, J. Ellis, The coordination of
work activities: cooperation and conflict in a hos-
pital context, Comp. Supp. Coop. Work 5 (1996)
1–31.
[60] R.H.R. Harper, K.P.A. O’Hara, A.J. Sellen,
D.J.R. Duthie, Toward the paperless hospital?, Br.
J. Anaesth. 78 (1997) 762–767.
[61] R.S. Dick, E.B. Steen (Eds.), The Computer-Based
Patient Record: An Essential Technology for
Health Care, National Academy Press, Washing-
ton, DC, 1991.
[62] C.J. McDonald, Physician’s needs for computer-
based patient records, in: M.J. Ball, M.F. Collen
(Eds.), Aspects of the Computer-Based Patient
Record, Springer, New York, 1992, pp. 3 – 11.
[63] A. Strauss, R. Bucher, D. Ehrlich, L. Schatzman,
M. Sabshin, The hospital and its negotiated order,
in: E. Freidson (Ed.), The Hospital in Modern
Society, Free Press, New York, 1963, pp. 147 – 169.
[64] O. Hanseth, E. Monteiro, Inscribing behavior in
information infrastructure standards, Acc. Man.
Inf. Techn. 7 (1997) 183–211.
[65] D.E. Forsythe, Using ethnography to build a
working system: rethinking basic design assump-
tions, in: 16th Symposium on Computer Applica-
tions in Medical Care, McGraw-Hill, New York,
1992.
[66] C.P. Friedman, J.C. Wyatt (Eds.), Evaluation
Methods in Medical Informatics, Springer, New
York, 1997.
[67] B. Kaplan, Qualitative research methods for evalu-
ating computer information systems, in: J.G. An-
derson, C.E. Aydin, S.J. Jay (Eds.), Evaluating
Health Care Information Systems, Sage, Thousand
Oaks, CA, 1993.
[68] R. Kling, Computerization and social transforma-
tions, Sci. Techn. Hum. Val. 16 (1991) 342 – 367.
[69] D.E. Forsythe, B.G. Buchanan, Broadening our
approach to evaluating medical information sys-
tems, in: P. Clayton (Ed.), Proceedings of the
Fifteenth Symposium on Computer Applications
in Medical Care, McGraw-Hill, New York, 1992.
[70] P.E. Agre, Institutional circuitry: thinking about
the forms and uses of information, Inf. Tech. Libr.
14 (1995) 225 – 230.
[71] S. Kay, I.N. Purves, Medical records and other
stories: a narratological framework, Methods Inf.
Med. 35 (1996) 72 – 87.
[72] S.L. Star, K. Ruhleder, Steps towards an ecology
of infrastructure: design and access for large infor-
mation spaces, Inf. Syst. Res. 7 (1996) 111.
[73] W.J. Orlikowski, The duality of technology: re-
thinking the concept of technology in organiza-
tions, Org. Sci. 3 (1992) 398 – 427.
[74] K. Schneider, I. Wagner, Constructing the ‘dossier
représentatif’: computer-based information-sharing
in French hospitals, Comp. Supp. Coop. Work 1
(1993) 229 – 253.
[75] Y. Engestrom, R. Engestrom, O. Saarelma, Com-
puterized medical records, production pressure and
compartmentalization in the work activity of
health centre phsysicians, in: Proceedings of the
Conference on Computer-Supported Cooperative
Work CSCW ’88, ACM Press, New York,
1988:65 – 84.
[76] D.E. Forsythe, New bottles, old wine: hidden cul-
tural assumptions in a computerized explanation
system for migraine sufferers, Med. Anthr. Quart.
10 (1996) 551 – 574.
[77] J. Bowers, The politics of formalism, in: M. Lea
(Ed.), Contexts of Computer Mediated Communi-
cation, Harvester, Hassocks, 1992, pp. 232 – 262.
[78] J. Greenbaum, M. Kyng (Eds.), Design at Work:
Cooperative Design for Computer Systems,
Lawrence Erlbaum, Hillsdale, NJ, 1991.
[79] F. Kensing. J. Simonsen, K. Bødker, MUST—a
method for participatory design, in: Participatory
Design Conference ’96, 1996.
 M. Berg / International Journal of Medical Informatics 55 (1999) 87–101
[80] J. Weedrnan, The structure of incentive: design
and client roles in application-oriented research,
Sci. Techn. Hum. Val. 23 (1998) 315–345.
[81] H. Scarbrough, J.M. Corbett, Technology and Or-
ganization. Power, Meaning and Design, Rout-
ledge, London, 1992.
[82] L. Suchman, R.H. Trigg, Artificial intelligence as
craftwork, in: S. Chaiklin, J. Lave (Eds.), Under-
standing Practice. Perspectives on Activity and
Context, Cambridge University Press, Cambridge,
1993, pp. 144–178.
[83] G. Walsham, Interpreting Information Systems in
Organizations, Wiley, Chichester, 1993.
[84] W.J. Orlikowski, Improvising organizational trans-
formation over time: a situated change perspective,
Inf. Syst. Res. 7 (1996) 63–92.
[85] O. Hanseth, K. Thoresen, L. Winner, The politics
of networking technology in health care, Comp.
Supp. Coop. Work 2 (1994) 109–130.
[86] O. Hanseth, E. Monteiro, M. Hatling, Developing
information infrastructure: the tension between
standardization and flexibility, Sci. Techn. Hum.
Val. 21 (1996) 407–426.
[87] M. Berg, Accumulating and coordinating: occa-
sions for information technologies in medical
work, Comp. Supp. Coop. Work (1999) in press.
[88] K. Schmidt, C. Simone, Coordination mecha-
nisms: towards a conceptual foundation of CSCW
systems design, Comp. Supp. Coop. Work 5 (1996)
155–200.
101
[89] E. Nygren, P. Henriksson, Reading the medical
record. I. Analysis of physicians’ ways of reading
the medical record, Comput. Methods Programs
Biomed. 39 (1992) 1 – 12.
[90] J. Lave, Cognitzon in Practice, Cambridge Univer-
sity Press, Cambridge, 1988.
[91] E. Hutchins, Cognitzon in the Wild, MIT Press,
Cambridge, 1995.
[92] J.C. Wyatt, Clinical data systems, part 2: compo-
nents and techniques, Lancet 344 (1994) 1609–
1614.
[93] P. Degoulet, et al., The component-based architec-
ture of the HELIOS medical software engineering
environment, Comp. Progr. Med. Biol. 45 (Suppl.)
(1994) S1 – S11.
[94] J. van der Lei, et al., The introduction of com-
puter-based patient records in The Netherlands,
Ann. Intern. Med. 119 (1993) 1036 – 1041.
[95] K. Henderson, The role of material objects in the
design process: a comparison of two design cul-
tures and how they contend with automation, Sci.
Techn. Hum. Val. 23 (1998) 139 – 174.
[96] L. Gasser, The integration of computing and rou-
tine work, ACM Trans. Office Inf. Syst. 4 (1986)
205 – 225.
[97] S.L. Star, Layered space, formal representations
and long-distance control: the politics of informa-
tion, Fundamenta Scientiae 10 (1989) 125 – 154.
[98] H.M. Collins, Artificial Experts. Social Knowledge
and Intelligent Machines, MIT Press, Cambridge,
1990.