Querying Social Practices in Hospital Context

John Bruntse Larsen

, Virginia Dignum

, Jørgen Villadsen

and Frank Dignum

DTU Compute, Technical University of Denmark, 2800, Kongens Lyngby, Denmark

Faculty of Technology, Policy and Management, Delft University of Technology, Delft, The Netherlands

Department of Information and Computing Sciences, Utrecht University, Utrecht, The Netherlands

Keywords:

Social Practices, Hospital Staff Planning, Multi-Agent Goals, Decision Support.

Abstract:

Understanding the social contexts in which actions and interactions take place is of utmost importance for

planning one’s goals and activities. People use social practices as means to make sense of their environment,

assessing how that context relates to past, common experiences, culture and capabilities. Social practices can

therefore simplify deliberation and planning in complex contexts. In the context of patient-centered planning,

hospitals seek means to ensure that patients and their families are at the center of decisions and planning of

the healthcare processes. This requires on one hand that patients are aware of the practices being in place at

the hospital and on the other hand that hospitals have the means to evaluate and adapt current practices to the

needs of the patients. In this paper we apply a framework for formalizing social practices of an organization

to an emergency department that carries out patient-centered planning. We indicate how such a formalization

can be used to answer operational queries about the expected outcome of operational actions.

1 INTRODUCTION

The importance of developing shared meaning and

understandings is one of the main aspects of main-

taining an organization. Hospital organizations are

no different and tend to develop their own ways of

doing things which need to be shared with novices.

Moreover, as vision and strategy change, new prac-

tices need to be implemented to ensure that medical

staff and users are able to function properly in the new

model of operation.

In recent years, ensuring patient satisfaction in

hospitals is becoming increasingly important, requir-

ing hospitals to develop means to continuously as-

sess and adapt their processes and practices. Patient-

centered planning is a way of thinking and doing

things that sees the patients in health and social ser-

vices as equal partners in planning, developing and

monitoring care to make sure the treatment meets

their needs. The patients and their families are put

at the center of decisions and seen as experts in their

own condition, working alongside professionals to

get the best outcome. Traditional planning technolo-

gies are typically designed to produce complete plans

that handle all situations given a well deﬁned set

of goals. This approach is not suitable for patient-

centered planning, which involve human agents, as it

leads to a combinatorial explosion if the actions of the

human agents are not predictable or limited. Social

practices provide a mean to deal with the complexity

by considering the social and physical constructs that

shape and constrain the interaction.

A hospital can be seen as the embodiment of on-

going social practices of many different forms (in-

cluding, for example, nursing, diagnosing, operating,

ofﬁce working, drug dispensing). It is also materi-

ally bound up in those practices (and their success-

ful performance), and in how these are changing over

time. For example, as patient-centered planning (in-

cluding how it is organized, managed and monitored)

becomes the practice in hospital management, exist-

ing processes and roles may become less functional

and need to adapt in order to stay ﬁt for that pur-

pose [Walker et al., 2014].

This work contributes with frameworks to help

with understanding of how patient-centered planning

relates to the changing practices of hospital work and

understandings of what constitutes a well-working

hospital. In particular we consider the emergency de-

partment and how the social practices that govern it

have consequences across the hospital. In the present

paper we propose our approach for applying a frame-

work for formalizing social practices of the emer-

gency department and how such a formalization can

Larsen, J., Dignum, V., Villadsen, J. and Dignum, F.

Querying Social Practices in Hospital Context.

DOI: 10.5220/0006580904050412

In Proceedings of the 10th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2018) - Volume 2, pages 405-412

ISBN: 978-989-758-275-2

405

be used for answering operational queries about the

expected outcome of operational actions. The emer-

gency department is also the focus of [Larsen and Vil-

ladsen, 2017] which proposes an approach for model-

ing it as a multi-agent organization.

2 SOCIAL PRACTICES

Social interactions are the focus of Social Sciences

and have also long been discussed in Multi-Agent

Systems (MAS) where there are many formalizations

and theories that describe speciﬁc aspects of inter-

actions [Van Den Broek et al., 2005, Dignum and

Dignum, 2011, K

ohler, 2007].

Social practices are accepted ways of doing

things, contextual and materially mediated, that

are shared between actors and routinized over

time [Reckwitz, 2002]. They can be seen as patterns

that are ﬁlled in by a multitude of single and often

unique actions. Through (joint) performance, the pat-

terns provided by the practice are ﬁlled out and repro-

duced. Each time a social practice is used, elements

of the practice, including know-how, meanings and

purposes, are reconﬁgured and adapted [Shove et al.,

2012]. Actors learn from the use of social practices,

and the more one uses a social practice with satis-

factory results, the more that actor will tend to use

that practice. Social practices are shared which means

that individuals playing a role in a social practice will

know what to do and will expect certain behavior and

reactions of the other participants. Success enforces

the social practice.

Related research on work practices and cognitive

architectures are the closest to our proposal. Work

practice research recognizes the inherent difference

between the work ﬂows as described and prescribed

by the organization and employee behavior. The work

practice model Brahms enables to deﬁne the behav-

ior of entities by means of activities and workframes,

amongst others [Sierhuis et al., 2009] but lacks learn-

ing capabilities to adjust priorities, and means to dis-

tinguish between context and action preconditions.

Cognitive architectures [Sun, 2009] use drives as ba-

sis to dynamically derive goals during agent interac-

tions. As such, it can capture the motivational com-

plexity of the human mind [Newell, 1994], but it takes

an agent perspective rather than a societal one, such as

the one we propose. Our approach with social prac-

tices also shows some resemblance to agent organiza-

tions (see e.g. [Dignum, 2004]) as they both provide

structure to the interactions between the agents. How-

ever, the structure provided by social practices arises

from the bottom up from components that deﬁne ex-

pected behavior in a given context rather than being

an imposed (top-down) organizational structure.

In [Dignum and Dignum, 2015] an initial formal

representation of social practices is presented that

allows for its incorporation in agent deliberation

architectures. It is based on modal logic and extends

work on agent organizations and landmarks. In

particular, it extends work on the Logic for Agent

Organizations (LAO) [Dignum and Dignum, 2011]

which formalizes notions of capability, stit [P

orn,

1974], attempt and responsibility. For the purpose of

this paper, we sufﬁce with the informal deﬁnitions of

the symbols in table 1 and the components of a social

practice as follows:

• Context

– Roles describe the competencies and expecta-

tions about a certain type of actors [Sunstein,

1996]. Thus a lecturer is expected to deliver

the presentation.

– Actors are all people and autonomous systems

involved, that have capability to reason and (in-

ter)act. This indicates the other agents that are

expected to fulﬁll a part in the practice.

– Resources are objects that are used by the ac-

tions in the practice such as — in case of a lec-

ture — seats, projector, screen, etc. So, they

are assumed to be available both for standard

actions and for the planning within the practice.

– Affordances are the properties of a resource, a

place or a context that indicate the action pos-

sibilities that are readily perceivable by an ac-

tor. E.g. any ﬂat surface at approximately knee-

length can afford the action ‘sit’. Affordances

permit social actions and depend on the match

between context conditions and actor character-

istics [Gaver, 1996].

– Places indicate where all objects and actors are

usually located relatively to each other, in space

or time.

• Meaning

– Purpose determines the social interpretation of

actions and of certain physical situations.

– Promotes indicate the values that are promoted

(or demoted, by promoting the opposite) by the

social practice.

– Counts-as are rules of the type “X counts as Y

in C” linking brute facts (X) and institutional

facts (Y) in the context (C) [Searle, 1995]. E.g.,

in a voting place, ﬁlling out a ballot counts as a

vote.

ICAART 2018 - 10th International Conference on Agents and Artiﬁcial Intelligence

406

• Expectations

– Plan patterns describe usual patterns of ac-

tions [Bresciani et al., 2004] deﬁned by the

landmarks that are expected to occur.

– Norms describe the rules of (expected) behavior

within the practice, using the ADICO grammar

proposed by Ostrom [Crawford and Ostrom,

1995]. ADICO statements are formed using

ﬁve components: Attribute (or Acting entity),

Deontic, aIm (or Intention), Condition, and Or

else (or sanction). Norms are statements that

include the acting entity, deontic, intention, and

condition (ADIC) and possibly also a sanction

(ADICO).

– Strategies indicate the possible activities that

are expected within the practice. Not all activ-

ities need to be performed! They are meant as

potential courses of action. Strategies are spec-

iﬁed as AIC statements in ADICO.

– Start condition, or trigger, indicating how the

social practice starts.

– Duration, or End condition, indicating how the

social practice ends.

• Activities

– Possible actions describe the expected actions

by actors in the social practice.

– Requirements indicate the type of capabilities

or competences that actors are expected to have

in order to enact roles and perform activities

within the social practice.

It is also important to note that social practices can

be described at different levels of abstraction. Ab-

stract practices are generic and do not give much de-

tails on the speciﬁcs of a context. Concrete social

practices apply to more speciﬁc domains, roles or ac-

tors and typically provide more details about the pos-

sible plans. E.g. an abstract social practice ‘greeting’

just indicates when a greeting gesture is expected,

but given a more speciﬁc context or actors it also de-

scribes how it should be performed; for example, by

‘bowing’ if context is Japan, or ‘hugging’ if actors are

good friends.

Finally, social practices relate to each other and

to the current situation. Depending on the features

of a situation individuals will choose (a set of) social

practices to explore. Moreover, social practices can

be composed or contain other social practices. E.g. a

handshake can be the start of a meeting. So it forms

part of the meeting social practice. On the other hand

the handshake is one type of greeting. So, handshake

is a specialization of greeting and a part of a meet-

ing. In fact, most classiﬁcation relations that apply to

processes (specialization, part-of, is-a, ...) can also be

applied to social practices.

3 PATIENT-CENTERED

PLANNING

Advance care planning is a process “whereby a pa-

tient, in consultation with health care providers, fam-

ily members and important others, makes decisions

about his or her future health care, should he or she

become incapable of participating in medical treat-

ment decisions” [Singer et al., 1996]. The process of

advance care planning informs and empowers patients

to have a say about their current and future treatment.

Patient-centered planning is a process in which

the hospital plans are made with needs of the patient

in center. A social practice approach to modeling

patient-centered planning can accommodate societal

requirements of patient-centered planning, including:

• respecting patients values and putting patients at

the center of planning.

• taking into account patients preferences and ex-

pressed needs.

• coordinating and integrating care planning.

• working together to ensure good communication,

information and education.

Social practices have both a social and a func-

tional purpose, which determines how they are used

by the different actors. To regard patient care as a so-

cial practice puts an appropriate emphasis on the rea-

sons why people take part in it and how they choose

to interpret the various expectations. Young doctors

value the hospital regulations and formal workﬂows

because they assume it serves efﬁciency and patient

care. Experienced doctors care for patients but often

less for formalisms. Phoning a specialist directly is

seen by senior doctors as a good social practice, be-

cause they know that the reaction of the specialist is

direct and will result in a quick treatment of the pa-

tient. Delays caused for other patients are not consid-

ered, because the overall hospital perspective is less

important for experienced doctors. Identifying and

understanding these differences in perceiving hospi-

tal policy is important for hospital managers and can

be used to redeﬁne processes and policies. In this

project, we will develop simulation models that en-

able hospital managers to analyze the consequences

of different social practices from the perspective of

patients and medical staff.

In the remainder of this section, we ﬁrst introduce

a simple scenario and then illustrate the use of so-

Querying Social Practices in Hospital Context

407

Table 1: Symbols of the social practice formal representation where O(γ) and F(γ) have standard dynamic deontic logic

semantics.

Symbol Meaning

Activities

Basic actions {α

, ..., α

} A ﬁnite number n of basic actions in the social

practice. Entails notion of a complex action γ.

Actors {a

, ..., a

} A ﬁnite number m of actors in the social practice.

Context {c

, ..., c

} A ﬁnite number n of special context constants. A

subset of these are social practice identiﬁers SP.

Capabilities Capability(a) A function that returns the set of actions that ac-

tor a is capable of performing.

Cap(a, α) True iff actor a is capable of performing action a

i.e. Cap(a, α) ≡ α ∈ Capability(a).

DO(a, α) Actor a performs action α next. Additionally as-

sumes that DO(a, α) → Cap(a, α)

Beliefs and Assumptions

purpose(a, α, c) = φ ≡

(Salient(a, α, c) ∧

DO(a, α) → Goal(a, φ) ∧

([α(a)]φ)

It is common belief among the agents in context

c that if actor a performs action α in context c

then φ is a goal of a and a believes it can achieve

it by performing α.

Meanings

play(a, r) Indicates that actor a enacts role r.

promote(sp, α(a), v) Indicates that in the social practice sp, the action

α(a) performed by actor a promote value v. For

a more formal characterization of the promotes

relation see [Weide, 2011].

Expectations

O(r, φ, γ) O(r, φ, γ) ≡ ∀a : play(a, r) ∧ B

(φ) → O(γ(a)).

F(r, φ, γ) O(r, φ, γ) ≡ ∀a : play(a, r) ∧ B

(φ) → F(γ(a)).

strategy(φ, DO(B, γ), sp) Indicates that if all actors (involved in the social

practice sp) believe that B believes the condition

φ then they all believe that all actors in B will

perform their part of γ next.

cial practices modeling to represent the situation de-

scribed in the scenario.

3.1 Scenario: Specialist Consults

In this paper, we consider the case of scheduling spe-

cialist consults for emergency ward patients. These

patients arrive at the hospital unscheduled and the at-

tending doctor may decide that a specialist opinion is

needed. However, these specialists follow a regular

consult plan and emergency patients will need to be

placed in between the regular patients. Several strate-

gies can be used for this, depending on the severity

of the case, the knowledge the emergency doctor has

about hospital procedures and informal strategies or

workarounds, and the overall culture of the hospital

location.

In particular, we will consider the following sce-

nario: A man arrives at the emergency ward with

very bad headache. The patient is not critically ill but

the doctor fears that his condition can deteriorate and

needs a second opinion from a specialist. Depending

on the seniority and expertise of the attending doctor,

two different plans will be followed:

• A junior doctor, not aware of the hospital’s infor-

mal processes and networks, will ﬁle a request for

a specialist consult through the formal channels

and wait for the specialist to make the time to at-

tend to this patient.

• A senior doctor, knowing how things work, will

directly call a specialist who will directly attend to

this request, delaying her consult with a scheduled

regular patient.

These two different practices lead to different re-

sults. In the ﬁrst case, the acute patient will need

to wait, which depending on his condition may have

serious health consequences. On the latter case, the

ICAART 2018 - 10th International Conference on Agents and Artiﬁcial Intelligence

408

scheduled patients will need to wait longer than their

appointed times, leading to irritation and possibly

complaints to the hospital administration.

As both of the situations above illustrate, the deci-

sions that the emergency ward doctors make depend

a lot on their experience with “how things work” in

the hospital. The junior doctor looks for an avail-

able specialist whereas the experienced doctor ﬁnds

a specialist who is scheduled for a checkup that they

agree can wait. The experienced doctor can do this

as they know that there is usually a specialist who has

an appointment with a patient who is otherwise do-

ing ﬁne. The experienced doctor is aware of how the

patient depends on the specialist; the junior doctor is

not. There is no clear cut “best” outcome as there are

both satisﬁed and dissatisﬁed patients in both cases

but it is clear that the interaction between the acute

doctors and the specialists inﬂuence the outcomes.

Overseeing the entire operation, the hospital man-

agement people are tasked with ensuring that there is

a balance between taking care of the scheduled pa-

tients and the acute patients. They are capable of ex-

ecuting operational actions that have either an imme-

diate effect during operation or long term actions with

effects in future operations.

Depending on the local culture and social expecta-

tions, delays to regular consults may require the hos-

pital to reschedule all patients to another day or offer

some kind of compensation, rather than expect those

regular patients to wait patiently to their consult. De-

pending on the frequency and number of emergency

ward patients, long term adaptation of existing proto-

cols may be required which demands well informed

management, as discussed in the next section.

3.2 Scenario as Social Practice

As described in Section 2, social practices can be de-

scribed at several levels of abstraction, resulting in a

graph structure that represents differentiation in terms

of context specialization, inclusion and salience. In

the case of the Specialist Consults scenario, practices

vary for senior and junior doctors and also for the so-

cietal expectations on how to deal with delays. Figure

1 gives an overview of these relations between social

practices.

We provide a basic formal representation of the

scenario as an abstract social practice as follows.

• Context

– Roles are identiﬁed from the actors that are

mentioned in the scenario: the junior doctor,

the senior doctor, the regular patient, and the

acute patient. We consider each actor as a role.

We also consider both the acute patient and the

regular patient as a patient role:

jd, sd, d, rp, ap, p : ap, rp ∈ p ∧ jd, sd ∈ d

– Actors are deﬁned by the play-predicate. We

use it to state that four people have the roles

above:

play( j, jd), play(s, sd), play(r, rp), play(a, ap)

– Resources include the room in which the doc-

tor diagnoses the patient, the scheduling sys-

tem, and other objects that we leave vague on

purpose for this abstract social practice:

, ..., o

– Places are the three main areas of the scenario:

the hospital as a whole, an MRI room (where

the specialists carry out their diagnosis), and

the emergency ward:

hospital, MRI-room, ER

• Meaning

– Purpose is deﬁned by purpose-predicate. We

use it to state that the purpose of the consult

action for acute patient a is to get a diagnosis,

and for a junior doctor j is to learn and gain

experience:

purpose(a, consult, sp) = diagnosis(a)

purpose( j, consult, sp) = learn( j)

– Promoted Values are deﬁned by the promote-

predicate. We use it to state that for the junior

doctor, tending patients promotes learning. We

also state that for the patient, getting a treatment

promotes awareness about their condition and

staying long in the ER demotes timeliness:

promote(sp, tendPatient( j), Learning)

∀p : promote(sp, diagnosis(p), Awareness)

∀p : promote(sp, longStay(p), ¬Timeliness)

• Expectations

– Plan Patterns are the sequences of abstract ac-

tions that models the general procedure as de-

scribed in the scenario. We use α

for the ab-

stract actions in the sequences. The sequence

models the two different approaches for the ab-

stract social practice to take place depending on

which of the two doctors take care of the acute

patient, as described in the scenario:

(arrive); α

(checkup); (α

(get2Op)+

(α

(plan2op); α

(wait);α

(get2Op)))

Querying Social Practices in Hospital Context

409

emergency

patient consult

society

doctor

find specialist wait specialist wait reschedule compensate

context context

abstract SP

concrete SP concrete SP concrete SP concrete SP concrete SP

senior

junior

Figure 1: Contextual variations on social practices.

start

end

checkup

get 2

plan 2

arrive

wait @home

Figure 2: Plan pattern.

– Norms are stated as deontic logic predicates.

We state that junior doctors are obliged to treat

acute patients ﬁrst, and that junior doctors are

forbidden from giving a diagnosis:

O( jd, plays(a, ap), prioritize( jd, a))

F( jd, plays(a, ap), diagnose( jd, a))

– Strategies are deﬁned by the strategy-predicate.

We use strategies to state that the patient waits,

that a senior doctor calls for help from a spe-

cialist by phone, and that a junior doctor waits

for help from a specialist:

strategy(>, DO(p, wait), sp)

strategy(diagnose(d, p),

DO(senior(d), phone), sp)

strategy(diagnose(d, p),

DO( junior(d), wait), sp)

– Start Condition & Duration/End are respec-

tively that the acute patient arrives and that the

acute patient has a treatment plan. As we only

consider this particular abstract social practice

we do not state them explicitly.

• Activities

– Possible Actions includes waiting, calling on

phone, and stating a diagnosis, among others:

wait(), phone(), diagnose(X, p), ...

– Requirements are stated with formulas using

the cap-predicate. We use it to state that the

doctors have medical expertise, and that pa-

tients wait:

∀a

, play(a

, d) : cap(a

, medicalexpertise)

∀a

, play(a

, p) : cap(a

, wait)

In Table 2 we provide a summary of the above. We

provide a more detailed description of the play-, pur-

pose-, promote-, strategy-, and cap-predicates. We

are working on the full speciﬁcation of the patient-

centered scenario as a validation of this formal repre-

sentation language.

4 OPERATIONAL QUERIES

Operational actions are generally considered risky

in the sense that they can have far reaching effects

and that it can be difﬁcult to estimate the conse-

quences. Immediate actions compromise the sched-

ule greatly and can turn up much more expensive than

ﬁrst thought. Long term actions that change the pro-

tocols may lead to immediate dissatisfaction among

staff but can be beneﬁcial in the long run. As an ex-

ample in the Specialist Consults scenario, the man-

agement may choose to have the specialists prioritize

acute patients over regular patients so that the junior

doctors will have an easier time ﬁnding an available

specialist. Such a change can cause dissatisfaction in

the specialized wards and it is unclear if it would actu-

ally be beneﬁcial in the long run. For that reason they

need insight into the expected outcome of the current

situation, and insight into how an operational action

can change the expected outcome. We consider an

approach for giving such insight by the way of opera-

tional queries that can answer questions such as:

• How many acute patients do we expect in the near

future?

• How long will the average waiting time be for

acute patients?

• What are the expected behavior of the senior doc-

ICAART 2018 - 10th International Conference on Agents and Artiﬁcial Intelligence

410

Table 2: A social practice application.

Social Practice A: Hospital SP (informal) B: Hospital SP (formal)

Context

Roles

junior doc, senior doc, regular

patient, acute patient, patient

jd, sd, d, rp, ap, p : ap, rp ∈ p ∧ jd, sd ∈ d

Actors

j, s, r, a

play( j, jd),play(s, sd), play(r, rp),play(a, ap)

Resources

rooms, scheduling system...

, ..., o

Places

hospital, MRI-room, ER... hospital, MRI-room, ER

Meaning

Purpose

diagnosis of patient purpose(a, consult, sp) = diagnosis(a)

teach junior doctor

purpose( j, consult, sp) = learn( j)

Promoted Values

tending patients promotes learn-

ing

promote(sp, tendPatient( j), Learning)

getting a diagnosis promotes

awareness

∀p : promote(sp, diagnosis(p), Awareness)

a long stay demotes timeliness ∀p : promote(sp, longStay(p), ¬Timeliness)

Expectations

Plan Patterns

the plan pattern graph in ﬁgure 2.

(arrive); α

(checkup); (α

(get2Op)+

(α

(plan2op); α

(wait);α

(get2Op)))

Norms

acute patients must be prioritized O( jd, plays(a, ap), prioritize( jd, a))

junior doctors cannot give diag-

nosis

F( jd, plays(a, ap), diagnose( jd, a))

Strategies

patients wait strategy(>, DO(p, wait), sp)

senior doctors ﬁnd available spe-

cialist

strategy(diagnose(d, p), DO(senior(d), phone), sp)

junior doctors wait for available

specialist

strategy(diagnose(d, p), DO( junior(d), wait), sp)

Start Condition

acute patient arrives

Duration/End

patient has plan

Activities

Possible Actions

wait, phone, diagnosis, ...

wait(), phone(), diagnose(X, p), ...

Requirements

doctor: medical expertise

∀a

, play(a

, d) : cap(a

, medicalexpertise)

patient: wait ∀a

, play(a

, p): cap(a

, wait)

tors?

• What are the expected behavior of the junior doc-

tors?

These queries are hard to answer analytically, but by

simulating the situation with artiﬁcial agents that act

according to our formalization of the social practices,

we can provide sufﬁciently accurate answers about

the expected outcome. For the simulation we con-

sider having the agents include their knowledge about

social practices when they decide on an action. The

operational query is then answered by simulating the

actions of the agents within the immediate future, re-

turning a list of pairs of outcomes together with their

expectancy. In this way, operational queries and sim-

ulation can provide insight into how decisions and

changes to hospital work practices inﬂuence expected

developments and potential bottlenecks.

5 CONCLUSION AND FUTURE

WORK

The social practice approach shows promise as a way

to augment agents with social reasoning. The hos-

pital context provides a good scenario for validating

this approach. The social practice approach supports

decision making. In the scenario we considered, the

operational queries can trigger long term operational

actions that change the protocols. If social context is

known then protocols can be designed to be in line

with social expectations.

In the future we would like to investigate to which

extent operational queries can trigger short term op-

erational actions with an immediate effect. Our next

step towards an evaluation is to implement the for-

mal representation in an agent simulation framework

Querying Social Practices in Hospital Context

411

that shows how changes to the social practice model

inﬂuence agent decisions. Furthermore, we intend to

investigate the work in the recent PhD thesis by Chris-

tian Michel Sørup on a generic performance mea-

surement model for an emergency department [Sørup,

2015]. We consider that work highly relevant as it

investigates performance measurement and decision

support in the emergency department scenario and

was done in close collaboration with the emergency

department at one of the hospitals in the Danish cap-

ital region. We hope that our approach using social

practices can add to that work.

ACKNOWLEDGEMENTS

This work is part of the Industrial PhD project Hos-

pital Staff Planning with Multi-Agent Goals between

PDC A/S and Technical University of Denmark. We

are grateful to Innovation Fund Denmark for funding

and the governmental institute Region H, which man-

ages the hospitals in the Danish capital region, for be-

ing a collaborator on the project. We would like to

thank PDC A/S for providing feedback on the ideas

described in this paper. We would also like to thank

Anders Schlichtkrull for comments on a draft.

REFERENCES

Bresciani, P., Perini, A., Giorgini, P., Giunchiglia, F., and

Mylopoulos, J. (2004). Tropos: An agent-oriented

software development methodology. Autonomous

Agents and Multi-Agent Systems, 8(3):203–236.

Crawford, S. E. S. and Ostrom, E. (1995). A grammar

of institutions. American Political Science Review,

89(03):582–600.

Dignum, V. (2004). A Model for Organizational Interac-

tion: based on Agents, founded in Logic. SIKS Dis-

sertation Series 2004-1. Utrecht University. PhD The-

sis.

Dignum, V. and Dignum, F. (2011). A logic of agent orga-

nizations. Logic Journal of IGPL.

Dignum, V. and Dignum, F. (2015). Contextualized plan-

ning using social practices. In Ghose, A., Oren, N.,

Telang, P., and Thangarajah, J., editors, Coordination,

Organizations, Institutions, and Norms in Agent Sys-

tems X, pages 36–52. Springer.

Gaver, W. W. (1996). Situating action II: Affordances for

interaction: The social is material for design. Ecolog-

ical Psychology, 8(2):111–129.

ohler, M. (2007). A formal model of multi-agent organi-

sations. Fundamenta Informaticae, 79(3-4):415–430.

Larsen, J. B. and Villadsen, J. (2017). An approach for

hospital planning with multi-agent organizations. In

Polkowski, L., Yao, Y., Artiemjew, P., Ciucci, D., Liu,

D., Slezak, D., and Zielosko, B., editors, Rough Sets -

International Joint Conference, IJCRS 2017, Olsztyn,

Poland, July 3-7, 2017, Proceedings, Part II, volume

10314 of Lecture Notes in Computer Science, pages

454–465. Springer.

Newell, A. (1994). Uniﬁed theories of cognition. Harvard

University Press.

orn, I. (1974). Some Basic Concepts of Action, pages 93–

101. Springer Netherlands, Dordrecht.

Reckwitz, A. (2002). Toward a theory of social practices.

European Journal of Social Theory, 5(2):243–263.

Searle, J. R. (1995). The construction of social reality. Si-

mon and Schuster.

Shove, E., Pantzar, M., and Watson, M. (2012). The Dy-

namics of Social Practice. Sage.

Sierhuis, M., Clancey, W. J., and van Hoof, R. J. J. (2009).

Brahms: An agent-oriented language for work prac-

tice simulation and multi-agent systems development.

In El Fallah Seghrouchni, A., Dix, J., Dastani, M.,

and Bordini, R. H., editors, Multi-Agent Program-

ming: Languages, Tools and Applications, pages 73–

117. Springer.

Singer, P. A., Robertson, G., and Roy, D. J. (1996).

Bioethics for clinicians: 6. advance care plan-

ning. CMAJ: Canadian Medical Association Journal,

155(12):1689.

Sørup, C. M. (2015). Development of a Generic Perfor-

mance Measurement Model in an Emergency Depart-

ment. PhD thesis, Department of Management Engi-

neering, Technical University of Denmark.

Sun, R. (2009). Motivational representations within a com-

putational cognitive architecture. Cognitive Computa-

tion, 1(1):91–103.

Sunstein, C. R. (1996). Social norms and social roles.

Columbia Law Review, 96(4):903–968.

Van Den Broek, E. L., Jonker, C. M., Sharpanskykh, A.,

Treur, J., et al. (2005). Formal modeling and analysis

of organizations. In International Conference on Au-

tonomous Agents and Multiagent Systems, pages 18–

34. Springer.

Walker, G., Shove, E., and Brown, S. (2014). How does

air conditioning become ‘needed’? A case study of

routes, rationales and dynamics. Energy Research &

Social Science, 4:1–9.

ICAART 2018 - 10th International Conference on Agents and Artiﬁcial Intelligence

412