Fra Office Automation via BPM til tvær-organisatorisk Case Management og Citizen Development af...

IT UNIVERSITY OF COPENHAGEN

Thomas Hildebrandt, [email protected]

Thomas Hildebrandt

Lektor & leder af Process and System Models gruppen vedIT Universitetet i København

Forbedrede finansielle processer med Business Process Management ?

Get F’IT - 25. September 2012

Fra Office Automation via BPM til tvær-organisatorisk Case Management og Citizen Development

Tuesday, September 25, 12

mailto:[email protected]



Get F’IT, BPM Seminar, 25. September, 2012 Thomas Hildebrandt, [email protected]

www.itu.dk/research/models

kort om mig• PhD i datalogi (processbeskrivelser), Århus, 2000

• Forsker og underviser på IT Universitetet, 1999-

• Forskningsprojekter i it-støttede mobile og tværorganisatoriske forretningsprocesser og arbejdsgange (MDCC, Resultmaker, Exformatics)

• Leder af Proces og System modeller forskningsgruppe, interessegruppe for Processer og IT, og videngruppe for digitalisering og procesorientering (VidenDanmark)

• Nyt europæisk netværk for “Behavioral Types”2




http://www.itu.dk/research/models





Overblik• State-of-the-art:

Fra Office Automation til BPMN 2.0

• Problemer/udfordringer:

Planlægningsoptimisme, fleksibilitet, distribution

• State-of-the-future:

Case Management Modeling Notation (CMMN)

Behavioral/Session Types

Citizen Development3









State-‐of-‐the-‐art

4

• Process-Aware Information Systems

• A shift from data orientation to process orientation: Explicit process models

• Business Process Model and Notation (BPMN) 2.0 - OMG Standard, 2011 (www.omg.org)

However, the focus is not on data but on process-related information (e.g., theordering of activities). Process mining is also related to monitoring and businessintelligence [41].

8 ConclusionProcess-aware information systems (PAISs) follow a characteristic life-cycle. Fig-ure 13 shows the four phases of such a life-cycle [7]. In the design phase, theprocesses are (re)designed. In the configuration phase, designs are implementedby configuring a PAIS (e.g., a WFMS). After configuration, the enactment phasestarts where the operational business processes are executed using the system con-figured. In the diagnosis phase, the operational processes are analyzed to identifyproblems and to find things that can be improved. The focus of traditional work-flow management (systems) is on the lower half of the life-cycle. As a result thereis little support for the diagnosis phase. Moreover, support in the design phase islimited to providing an editor while analysis and real design support are missing.

Figure 13: PAIS life-cycle.

In this article, we showed that PAISs support operational business processesby combining advances in information technology with recent insights from man-agement science. We started by reviewing the history of such systems and thenfocused on process design. From the many diagramming techniques available, wechose one particular technique (Petri nets) to show the basics. We also emphasizedthe relevance of process analysis, e.g., by pointing out that 20 percent of the morethan 600 process models in the SAP reference model are flawed [24]. We also

26






http://www.omg.org

http://www.omg.org




• Kundepakke: Forretningsprocess for salg af pakkeløsning: kreditkort, netbankadgang, ...

• 10 forskellige systemer, 30 del-processer, 200 service-kald/menneskelige opgaver.

• Fra 200 pakker om dagen til 1800

• 80% automatisering

5

Steen Brahe, Industrial PhD, Danske Bank & IT University of Copenhagen“BPM on Top of SOA: Experiences from the Financial Industry”, BPM 2007

BPM eksempel









• Kundepakke: Forretningsprocess for salg af pakkeløsning: kreditkort, netbankadgang, ...

• 10 forskellige systemer, 30 del-processer, 200 service-kald/menneskelige opgaver.

• Fra 200 pakker om dagen til 1800

• 80% automatisering

5

Steen Brahe, Industrial PhD, Danske Bank & IT University of Copenhagen“BPM on Top of SOA: Experiences from the Financial Industry”, BPM 2007

BestIndustry Paper

BPM eksempel









ORDE

R PR

OCES

SING

Log

Requ

est

Type

Ord

er

Send

Ord

er

Rece

ive

Ord

er

Broces

s I I J I I • :1.

\ 1

Custo

mer

Re

ques

t A

rriv

al lL

-""

/ I I r • r J /. I I A / I

" Ord

er

Form

'" I +0

I

, Cu

s tam

er

J j

Fi le

I

II

Bill

ing

File

I I I J I J I I I I I I

I /

I ,

lOut

"

1\:;;1

tJ .

lOu

t \.

I t

Form

'-

---

----

/--"

"

',---

- ... _

----

--_

.. _-.

.,.-

'

F.ig

ure

2

BPM anno 1979

• Office automation: Zisman & Hammer ’77, IBM Business Definition Language, Information Control Net (Ellis ’79 Xerox)

6

z c :=

Computer Science and Office Information Systems

By Clarence A. Ellis and Gary J. Nutt









ORDE

R PR

OCES

SING

Log

Requ

est

Type

Ord

er

Send

Ord

er

Rece

ive

Ord

er

Broces

s I I J I I • :1.

\ 1

Custo

mer

Re

ques

t A

rriv

al lL

-""

/ I I r • r J /. I I A / I

" Ord

er

Form

'" I +0

I

, Cu

s tam

er

J j

Fi le

I

II

Bill

ing

File

I I I J I J I I I I I I

I /

I ,

lOut

"

1\:;;1

tJ .

lOu

t \.

I t

Form

'-

---

----

/--"

"

',---

- ... _

----

--_

.. _-.

.,.-

'

F.ig

ure

2

BPM anno 1979

• Office automation: Zisman & Hammer ’77, IBM Business Definition Language, Information Control Net (Ellis ’79 Xerox)

6

z c :=

Computer Science and Office Information Systems

By Clarence A. Ellis and Gary J. Nutt

By considering the specification language, the internal representation, and the design of a prototype

system using one unified model, Zisman has been able to study the office as a system rather than

simply as a collection of isolated tasks and pieces of equipment. Although Zisman suggests the

language and the model need refinement, his basic notions will probably have great impact on the

office of the future.








www.itu.dk/research/models 7

Business Process Model and Notation, v2.0 47

Figure 7.8 - An example of a stand-alone Process (Orchestration) diagram

(Figur fra BPMN 2.0 Final Specification, 2011)

BPM anno 2011: BPMN2.0









Hvad skete der i de 30 år?

8









Hvad skete der i de 30 år?• Sene 70‘ere/tidlige 80‘ere: Optimisme, men ikke

success - manglende standarder, primitive brugergrænseflader og for rigide processer...

8











• Fra 90‘erne til nu: Service og process-standarder (WfMC, Oasis, OMG), intuitive brugergrænseflader, optimisme....

8











• Fra 90‘erne til nu: Service og process-standarder (WfMC, Oasis, OMG), intuitive brugergrænseflader, optimisme....

8

men stadig for rigide process-modeller...

“Good standards for business process modelling are still missing and even today’s

WFMSs are too rigid”Process-Aware Information Systems:Design, Enactment, and AnalysisWil M.P. van der AalstDepartment ofMathematics and Computer Science, Eindhoven University of Tech-nology, P.O. Box 513, NL-5600 MB Eindhoven, [email protected]

Abstract. Process-aware information systems support operational business pro-cesses by combining advances in information technology with recent insightsfrom management science. Workflow management systems are typical examplesof such systems. However, many other types of information systems are also“process aware” even if their processes are hard-coded or only used implicitly(e.g., ERP systems). The shift from data orientation to process orientation has in-creased the importance process-aware information systems. Moreover, advancedanalysis techniques ranging from simulation and verification to process miningand activity monitoring allow for systems that support process improvement invarious ways. This article provides an overview of process-aware informationsystems and also relates these to business process management, workflow man-agement, process analysis techniques, and process flexibility.

Keywords: Process-Aware Information Systems, Workflow Management, Busi-ness Process Management, Petri Nets, Process Mining, Process Verification, Sim-ulation

1 IntroductionInformation technology has changed business processes within and between enter-prises. More and more work processes are being conducted under the supervisionof information systems that are driven by process models. Examples are work-flow management systems such as FileNet P8, Staffware, WebSphere, FLOWerand YAWL and Enterprise Resource Planning (ERP) systems such as SAP andOracle. Moreover, many domain specific systems have components driven by(process) models. It is hard to imagine enterprise information systems that areunaware of the processes taking place. Although the topic of business processmanagement using information technology has been addressed by consultants

1









Eksempel

• Arbejdsgangsbanken.dk - KL og KOMBIT

• Næsten 900 gode (best practice) og korrekte (lovmedholdige) arbejdsgange modelleret i BPMN og gjort frit tilgængelige

9









Nogle udfordringer....

10









Nogle udfordringer....• Processer, opgavefordeling og

lovgivning ændrer sig konstant og er afhængig af lokale forhold

10

[Magnus Nilsson, ITU TR-2011-140]










• Den samme hændelse optræder i flere processer/love - med forskellige navne

10











• Den samme hændelse optræder i flere processer/love - med forskellige navne

• Ingen formel sammenhæng mellem lov og process

10

• Lov om Aktiv beskæftigelsesindsats

(LBK nr 1428 af 14/12/2009)

• Lov om Aktiv socialpolitik

(LBK nr 946 af 01/10/2009)

• Lov om Arbejdsløshedsforsikring

(LBK nr 574 af 27/05/2010)

• Lov om Integration af udlændinge

(LBK nr 1062 af 20/08/2010)

• Lov om Sygedagpenge

(LOV nr 563 af 09/06/2006)

• Retssikkerhedsloven

(LBK nr 1054 af 07/09/2010)

• Datagrundlag

(BEK nr 418 af 23/04/2010)









Computer Supported CooperaUve Work (CSCW)

• Nyt forskningsområde introduceret i 80‘erne - fokus på at forstå hvordan folk samarbejder som basis for at designe it-systemer der kan støtte

• Afhængigheder mellem og koordinering af opgaver

• Konkluderede allerede i ’83 at office automation systems “do not deal well with unanticipated conditions” (Barber) og “were automating a fiction” (Sheil)

11

[Schmidt & Bannon: Taking CSCW Seriously: Supporting Articulation Work, 1992]









Mål

• Processstøtte

• Effektivitet, korrekthed, vedligeholdbarhed

• “Elasticitet” og fleksibilitet på tværs af organisationer

• Bygge bro mellem regulationer/policies og processer

12









Mål

• Processstøtte

• Effektivitet, korrekthed, vedligeholdbarhed

• “Elasticitet” og fleksibilitet på tværs af organisationer

• Bygge bro mellem regulationer/policies og processer

12

office information systems. In the seventies, people like Skip Ellis [10], AnatolHolt [11], and Michael Zisman [12] already worked on so-called office informa-tion systems, which were driven by explicit process models. It is interesting tosee that the three pioneers in this area independently used Petri-net variants tomodel office procedures. During the seventies and eighties there was great op-timism about the applicability of office information systems. Unfortunately, fewapplications succeeded. As a result of these experiences, both the application ofthis technology and research almost stopped for a decade. Consequently, hardlyany advances were made in the eighties. In the nineties, there again was a hugeinterest in these systems. The number of WFMSs developed in the past decadeand the many papers on workflow technology illustrate the revival of office infor-mation systems. Today WFMSs are readily available. However, their applicationis still limited to specific industries such as banking and insurance. As indicatedby Skip Ellis it is important to learn from these ups and downs. The failures inthe eighties can be explained by both technical and conceptual problems. In theeighties networks were slow or not present at all, there were no suitable graphicalinterfaces, and proper development software was missing. However, there werealso more fundamental problems: a unified way of modeling processes was miss-ing and the systems were too rigid to be used by people in the workplace. Most ofthe technical problems have been resolved by now. However, the more conceptualproblems remain. Good standards for business process modeling are still missingand even today’s WFMSs are too rigid.

One of the great challenges of PAISs is to offer both support and flexibility.Today’s systems typically are too rigid, thus forcing people to work around thesystem. One of the problems is that software developers and computer scientistsare typically inspired by processes inside a computer system rather than processesoutside a computer. As a result, these engineers think in terms of control systemsrather than support systems. This explains that few of the existing WFMSs allowfor the so-called implicit choice, i.e., a choice resolved by the environment ratherthan the system.

To summarize we would like to state that, although the relevance of PAISs isundisputed, many fundamental problems remain to be solved. In the remainder ofthis article, we will try to shed light on some of these problems.

7

Process-Aware Information Systems:Design, Enactment, and AnalysisWil M.P. van der AalstDepartment ofMathematics and Computer Science, Eindhoven University of Tech-nology, P.O. Box 513, NL-5600 MB Eindhoven, [email protected]

Abstract. Process-aware information systems support operational business pro-cesses by combining advances in information technology with recent insightsfrom management science. Workflow management systems are typical examplesof such systems. However, many other types of information systems are also“process aware” even if their processes are hard-coded or only used implicitly(e.g., ERP systems). The shift from data orientation to process orientation has in-creased the importance process-aware information systems. Moreover, advancedanalysis techniques ranging from simulation and verification to process miningand activity monitoring allow for systems that support process improvement invarious ways. This article provides an overview of process-aware informationsystems and also relates these to business process management, workflow man-agement, process analysis techniques, and process flexibility.

Keywords: Process-Aware Information Systems, Workflow Management, Busi-ness Process Management, Petri Nets, Process Mining, Process Verification, Sim-ulation

1 IntroductionInformation technology has changed business processes within and between enter-prises. More and more work processes are being conducted under the supervisionof information systems that are driven by process models. Examples are work-flow management systems such as FileNet P8, Staffware, WebSphere, FLOWerand YAWL and Enterprise Resource Planning (ERP) systems such as SAP andOracle. Moreover, many domain specific systems have components driven by(process) models. It is hard to imagine enterprise information systems that areunaware of the processes taking place. Although the topic of business processmanagement using information technology has been addressed by consultants

1









State-‐of-‐the-‐future• Adaptive Case Management

• Communication-centered computing & Behavioral types (protokoller/kontrakter)

• Brugerinddragelse (Generativity, citizen developers)

13









AdapUve Case Management

14

http://www.xpdl.org/nugen/p/adaptive-case-management/public.htm

“Adaptive Case Management (ACM) is information technology that exposes structured and unstructured business information (business data and content) and allows structured (business) and unstructured (social) organizations to execute work (routine and emergent processes) in a secure but transparent manner.”

Keith Swensson: Mastering the Unpredictable










14



BPM











14



BPM


ACM

(case files)










Case Management Modelling

• OMG Request for Proposals (RFP) September 2009 Case Management Process Modeling (CMPM) udvidelse til BPMN 2.0

(bmi/2009-09-23, www.omg.org/public_schedule)

• BizAgi, Cordys, IBM, Oracle, SAP og Singularity:

Case Management Modelling Notation (CMMN) V1.0

(stærkt inspireret af Guard-Stage-Milestone paradigme udviklet ved IBM Research)

15




http://www.omg.org/public_schedule

http://www.omg.org/public_schedule




Case Management RFP

• A situation/case and desired outcome

• A subject: person, insurance claim, ...

• Actions/events to achieve the outcome

• Human judgment/decision-making

• No predefined sequence

• Guidance, prompts, constraints & planning support

16

http://www.omg.org/public_schedule/)






http://www.omg.org/public_schedule/

http://www.omg.org/public_schedule/




Case Management Modelling

17







Case Management Modelling• Afhængigheder mellem handlinger - ikke

eksplicit process-flow

17









17

Hold meeting with DA

Open case

U LO

14 Days

Update Case

Betingelse(condition)

Opfølgning (response)









• Tillad dynamiske ændringer

17


Open case

U LO

14 Days

Update Case












17

Agree on date

LO & DA

3 Days Hold meeting with DA

Open case

U LO

14 Days

Update Case












17

Propose dates-LO

AcceptLO

Lo

Propose dates-DA

AcceptDA

Lo

DA DA+

+

%

%

3 Days

1 Day

%

%

1 Day


Open case

U LO

14 Days

Update Case












17

Propose dates-LO

AcceptLO

Lo

Propose dates-DA

AcceptDA

Lo

DA DA+

+

%

%

3 Days

1 Day

%

%

1 Day

Dynamic Condition Response (DCR) GraphsCross-organizational Case ManagementGuard-Stage-Milestone Paradigm, IBM Research, NY

Ph.d-projekt, Tijs Slaats og afhandling af Rao R. Mukkamala

(se www.itu.dk/people/rao og www.trustcare.dk)

Trustworthy Pervasive Healthcare Services project

Event-based Process Modelling and Programming withDynamic Condition Response Graphs

Thomas Hildebrandt1, Raghava Rao Mukkamala1, Tijs Slaats1,2 and Francesco Zanitti1(hilde,rao,tslaats,[email protected])

1 IT University of Copenhagen, IT University of Copenhagen, Rued Langgaardsvej 7, 2300 Copenhagen, Denmark2 Exformatics A/S, Lautrupsgade 13, 2100 Copenhagen, Denmark

Dynamic Condition Response Graphs (DCR Graphs) [3]have been developed in the TrustCare [7] research projectas a declarative formal foundation for event-based, adapt-able and flexible pervasive workflow processes and health-care services.

As exemplified by the simple treatment process illustratedin the figure below, a DCR Graph is essentially a setof repeatable events, whose execution order can be con-strained by five relations. The condition (!•) denotesthat an event is blocked until some other events happensat least once, the response (•!) denotes that an event isrequired to happen after some other event, the milestone(!⇧) denotes that some event can not happen while an-other event is required to be done, the exclude (!%) re-lation denotes that an event removes an event (togetherwith it’s constraints) from the workflow and it’s dual,the include (!+) relation can be used to add an eventback into the workflow. The example shows the visualnotation for DCR Graphs, inspired by Declare [1], an-other declarative language for modelling flexible work-flows formalized in Linear-time Temporal Logic (LTL).

prescribe medicine sign

give medicine

don't trust

+

D D

NN+

%%end

treatment

N Condition Relation

Include Relation

Response Relation

+Exclude Relation %

Milestone Relation

%

During the last year DCR Graphs have been extendedwith nested sub graphs [4] and we developed a formaltechnique for safe distribution of DCR Graphs [5]. Asexemplified below, the distribution provides for any (cov-ering) division of the events, a set of corresponding localgraphs obtained as projections.


don't trust

D D

end treatment

%


give medicine

don't trust

+ NN+

%%end

treatment

N

%

ND

In our most recent work, DCR Graphs have been ex-tended with data [8] and support for model checking ofsafety and liveness properties in SPIN [9]. A first versionof a declarative Process-oriented Event-based Program-ming Language (PEPL) [6] based on DCR Graphs andaimed for programming context-sensitive reactive serviceshas been developed and prototyped.

A PEPL program for the example process is shown be-low.Treatment(patient:num, medicine:string, qty:num) {

init prescribesign {include give include distrust}must give when condition sign {exclude distrust}sign {include give include distrust}distrust when condition sign {response sign exclude give}final endtreatment when milestone give

}

We have applied DCR Graphs as a modelling languagefor healthcare processes and case management systems [4].As exemplified below, Exformatics, which co-financesthe industrial PhD of the third author, has adopted DCRGraphs within their case management and documenthandling system.

Currently we are relating our work to other formalismssuch as Declare, Petri-nets and the Guard-Stage-Milestonemodel [2], which is proposed as formal foundation forthe coming OMG Case Management Standard. We arealso developing techniques for adaptive DCR Graphs andcontinuing the work on distribution by investigating be-havioral type systems for distributed DCR Graphs.

References:1. Van der Aalst, Pesic.: DecSerFlow: Towards a truly declarativeservice flow language. WS-FM 2006.2. Hull et al.: Business Entities with Guard-Stage-Milestone Life-cycles: Managing Entity Interactions with Conditions and EventsDEBS 20113. Hildebrandt and Mukkamala: Distributed Dynamic Condition Re-sponse Structures. PLACES 20104. Hildebrandt, Mukkamala and Slaats: Designing a Cross-organizationalCase Management System using Dynamic Condition Response Graphs.EDOC 20115. Hildebrandt, Mukkamala and Slaats: Safe Distribution of Declar-ative Processes. SEFM 20116. Hildebrandt and Zanitti: A Process-oriented Event-based Pro-gramming Language DEBS20127. Trustworthy pervasive healthcare processes (TrustCare) researchproject. Funded by The Danish Council for Strategic Research (grant#2106-07-0019) http://www.trustcare.dk/8. Mukkamala. A Formal Model For Declarative Workflows - Dy-namic Condition Response Graphs. PhD thesis.9. http://tinyurl.com/dcrveri


Open case

U LO

14 Days

Update Case






http://www.itu.dk/people/rao

http://www.itu.dk/people/rao

http://www.trustcare.dk

http://www.trustcare.dk




Behavioral Types• Tværorganisatoriske/globale

processer kan beskrives som koreografier

• Men hvordan sikres at de interne/lokale processer følger koreografien?

• Behavioral Types: Beskriver protokollen for hver lokal process der typecheckes før de køres

18

SimpleTreatmentChoreorgraphy

Test1

Nurse

PatientDoctor

Test2

Nurse

PatientDoctor

Prescribe

Doctor

NursePatient

Discharge

result1

result2

prescript ion

not o

k

ok

Associate Professor Thomas Hildebrandt 1 of 1 23.05.2012

38 Business Process Model and Notation, v2.0

Merging BPMN uses the term :merge; to refer to the exclusive combining of two or more paths into one path (also known as an OR-Join).A Merging Exclusive Gateway is used to show the merging of multiple Sequence Flows (see upper figure to the right).If all the incoming flow is alternative, then a Gateway is not needed. That is, uncontrolled flow provides the same behavior (see lower figure to the right).

Looping BPMN provides 2 (two) mechanisms for loop-ing within a Process.

See Next Two Figures

Activity Looping The attributes of Tasks and Sub-Processes will determine if they are repeated or per-formed once (see page 197). There are two types of loops: Standard and Multi-Instance. A small looping indicator will be displayed at the bottom-center of the activity.

Sequence Flow Looping Loops can be created by connecting a Sequence Flow to an :upstream; object. An object is considered to be upstream if that object has an outgoing Sequence Flow that leads to a series of other Sequence Flows, the last of which is an incoming Sequence Flow for the original object.

Table 7.2 - BPMN Extended Modeling Elements

38 Business Process Model and Notation, v2.0

Merging BPMN uses the term :merge; to refer to the exclusive combining of two or more paths into one path (also known as an OR-Join).A Merging Exclusive Gateway is used to show the merging of multiple Sequence Flows (see upper figure to the right).If all the incoming flow is alternative, then a Gateway is not needed. That is, uncontrolled flow provides the same behavior (see lower figure to the right).

Looping BPMN provides 2 (two) mechanisms for loop-ing within a Process.

See Next Two Figures

Activity Looping The attributes of Tasks and Sub-Processes will determine if they are repeated or per-formed once (see page 197). There are two types of loops: Standard and Multi-Instance. A small looping indicator will be displayed at the bottom-center of the activity.

Sequence Flow Looping Loops can be created by connecting a Sequence Flow to an :upstream; object. An object is considered to be upstream if that object has an outgoing Sequence Flow that leads to a series of other Sequence Flows, the last of which is an incoming Sequence Flow for the original object.

Table 7.2 - BPMN Extended Modeling Elements

Figure 1. Workflow as BPMN 2.0 Choreography

corresponding Process Matrix spreadsheet. We then demonstrate in Sec. 2.2 howthe process matrix workflow processes can be described compactly in multipartysession types with assertions and symmetric sum types. Finally we overview theprototype implementation in Sec. 2.3.

2.1 Example Workflow as Choreography and Process Matrix

A simple CPG workflow involving three participants is described in Fig. 1 as aChoreograpy diagram in the Business Process Modelling Notation (BPMN) 2.0.The described workflow is activated, when a patient is admitted (indicated by thestart event). Then two tests are executed in parallel by a nurse. Note that eachactivity box is a communication between the three participants with one initiator(indicated in the white ribbon) and two receivers (indicated in the shaded ribbons).Thus, the test results are sent by the nurse to both the patient and the doctor.Each test may be repeated, as indicated by the repeating subprocess arrow, e.g. ifthe test failed or the result was not clear. Then, depending on the results of thetests, either the patient is discharged directly, or the doctor prescribes a drug to thepatient, sending the prescription to both the patient and the nurse. The workflowis ended when the patient is discharged. The described workflow is a standardparadigm in CPGs; that is, first a set of tests are performed and, depending on theresults, either more tests are performed, the patient is discharged, or a treatmentis executed. In this workflow the treatment consists of simply prescribing a drugto the patient.

For our demonstrator we do not use BPMN 2.0 choreography diagrams. In-stead we use a simplified version of the declarative Process Matrix representationdeveloped by our industrial partner Resultmaker (http://www.resultmaker.com/)in the TrustCare research project. The process matrix corresponding to the chore-ography in Fig. 1 is shown in Fig. 2 below, using three bolean data fields (pre,result1, and result2) explained below.

3

Betty: EU COST Action









Brugerinddragelse

19

The model-view-controller architecture of apims supports fully flexible inter-face design without compromising the trustworthiness of correct execution of thespecified workflows. Furthermore, the communication between clients and apimsis mediated through the interface definition language framework Thrift [23], whichsupports multiple language bindings. Altogether, this supports flexible client de-sign for usability in a pervasive context; e.g., a simple approach to secure ande�cient inputting on a tablet computer has been by using QR-codes scannedthough the tablet’s camera. This enables a user to scan the drug name and thedose from physical objects, minimising the amount and attendant risks of manualtyping. We have only superficially touched upon the technical and usability chal-lenges of developing user interface clients for tablet computers in comparison toconventional PC clients, however.

3 Experiment: An End-user Developed Workflow

Id Name D N S AN OPN Seq Log Condition1.1.1 Nurse evaluation R W R R R1.1.2 Patient History W R R R R1.1.3 Extended history W R R R R abnorm1.1.4 Preoperative treatment W R R R R cyto1.1.5 Objective W R R R R1.1.6 Extended objective W R R R R abnorm21.1.7 Ultrasound W R R R R1.1.8 Formalia W R R R R1.1.9 Extended formalia W R R R R abnorm31.2.0 Information for the patient R W R R R 1.1.1 - 1.1.91.2.1 Schedule for OP R R W R W 1.2.0

Figure 9. End-user developed workflow (Flow).

To test the developed software, we performed a simple experiment with thehelp of a physician: Dorthe Furstrand Lauritzen (DFL). The motivation behindthe experiment was to get first hand impressions from a domain expert, to evaluatethe current implementation and set goals for future development. Although DFLis a physician and not a computer scientist, she has experience with use of IT andin particular implementation of CPGs. However, she had never seen any of thetechniques used in the demonstrator before, in particular the declarative processmatrix notation was completely new to her.

The main component of the experiment was to put DFL in the role of theworkflow designer, letting her use the spreadsheets to formalise a simple self-chosenmedical workflow, which can be run on the Android tablets. There are severalaspects of the experiment:

– Letting a medical professional come with a self-chosen workflow, tests theexpressiveness of the system.

– Letting a new user interact with the workflow creation tool tests the usabilityof the tool.

10

Citizen Development/Generativity:Så nemt som at bruge et regneark...

– Letting a medical professional use the tool, tests the hypothesis: the domainexpert can implement simple workflows, leading to a simpler and more flexibledevelopment process, e.g.• The domain experts might be able to make simple changes directly without

involving the development team.• The domain experts can use simple workflows to communicate more di-

rectly and e�ciently with the development team.

Id Input Action1.1.1 name height weight bP1.1.2 cave ever birth healthy if ! healthy then set(abnorm);

if cave then set(abnorm);if ! ever birth then set(cyto)

1.1.3 cavetx healthtx1.1.4 rp cytotec1.1.5 gU ia stet c et p ia if ! gU ia then set(abnorm2);

uterus retroflekteret if ! stet c et p ia then set(abnorm2)1.1.6 sttx gutx1.1.7 fHR cRL gA1.1.8 clamydiatested clamydia negative if ! clamydiatested then set(abnorm3);

rhesus negative signed form A if ! clamydia negative then set(abnorm3);under 18 gA under 12 if rhesus negative then set(abnorm3);

if under 18 then set(abnorm3);if ! signed form A then reset(1.1.8);if ! gA under 12 then reset(1.1.8)

1.1.9 rp antibiotics rp anti D signed form B1.2.0 pt informeret samtykke1.2.1 op tid gA ved op

Figure 10. End-user developed workflow (Data).

3.1 The Experiment

The experiment, which took a single day, was set-up as follows: DFL had access toa computer where the server, the code generator and example spreadsheets wereavailable. To simplify the interface, all spreadsheets were placed on the desktopand batch commands performed the code generation and server start. To learn thesyntax, DFL did a small exercise under instruction by one of the authors.

The workflow chosen by DFL model how a healthy woman gets an abortion,according to DFL was “a simplification of the simplest workflow I could find”.

The developed workflow is shown in Fig. 9 and Fig. 10. The roles are: Doctor(D), Nurse (N), Secretary (S), Anaesthesiologist (AN) and operation nurse (OPN).

An example screenshot from the running Android client is shown in Fig. 11.

3.2 Evaluation

Generally the experiment turned out very successfully: DFL was easily able to usethe spreadsheets to build her own workflow. The instructing author only had to

11

2nd International Symposium on Foundations of Health Information Engineering and Systems, 2012









Brugerinddragelse

20


– Letting a medical professional use the tool, tests the hypothesis: the domainexpert can implement simple workflows, leading to a simpler and more flexibledevelopment process, e.g.• The domain experts might be able to make simple changes directly without

involving the development team.• The domain experts can use simple workflows to communicate more di-

rectly and e�ciently with the development team.

Id Input Action1.1.1 name height weight bP1.1.2 cave ever birth healthy if ! healthy then set(abnorm);

if cave then set(abnorm);if ! ever birth then set(cyto)

1.1.3 cavetx healthtx1.1.4 rp cytotec1.1.5 gU ia stet c et p ia if ! gU ia then set(abnorm2);

uterus retroflekteret if ! stet c et p ia then set(abnorm2)1.1.6 sttx gutx1.1.7 fHR cRL gA1.1.8 clamydiatested clamydia negative if ! clamydiatested then set(abnorm3);

rhesus negative signed form A if ! clamydia negative then set(abnorm3);under 18 gA under 12 if rhesus negative then set(abnorm3);

if under 18 then set(abnorm3);if ! signed form A then reset(1.1.8);if ! gA under 12 then reset(1.1.8)

1.1.9 rp antibiotics rp anti D signed form B1.2.0 pt informeret samtykke1.2.1 op tid gA ved op

Figure 10. End-user developed workflow (Data).

3.1 The Experiment

The experiment, which took a single day, was set-up as follows: DFL had access toa computer where the server, the code generator and example spreadsheets wereavailable. To simplify the interface, all spreadsheets were placed on the desktopand batch commands performed the code generation and server start. To learn thesyntax, DFL did a small exercise under instruction by one of the authors.

The workflow chosen by DFL model how a healthy woman gets an abortion,according to DFL was “a simplification of the simplest workflow I could find”.

The developed workflow is shown in Fig. 9 and Fig. 10. The roles are: Doctor(D), Nurse (N), Secretary (S), Anaesthesiologist (AN) and operation nurse (OPN).

An example screenshot from the running Android client is shown in Fig. 11.

3.2 Evaluation

Generally the experiment turned out very successfully: DFL was easily able to usethe spreadsheets to build her own workflow. The instructing author only had to

11










Brugerinddragelse

21


Figure 7. Screenshot from Android client logged in as Doctor role.

the values shown on the right-hand side once the Doctor has entered those. Thelast guivalue in line 24 is used to pass the execution state of the action to theclient. This results in a small green checkmark for the action, so the user knowsthat it has been performed. In the screenshot the execution state is false, so thecheckmark is not green.

It is important to stress that every participant uses the same generic Androidclient. The GUI manager uses the generated code to make sure that the Androidclient used by the Doctor presents only the local process corresponding to theworkflow relevant to the doctor, and the Android client used by the Nurse presentsonly the local process relevant to the Nurse. An example screenshot of the Androidclient running the example as the nurse role is shown in Fig. 8. It shows theworkflow in a state where the nurse has performed the lung test with a negativeresult and still needs to perform the throat test.

Figure 8. Screenshot from Android client logged in as Nurse role.

9










Konklusion(er)• State-of-the-art tillader at beskrive procedurer som

brugere og forretningsfolk kan forstå

• I praksis bryder videnarbejderen ofte med den idéelle procedure og processer ændres før de tages i brug...

• Ringe support for compliancetjek, brugerinddragelse, deling & decentral udvikling af distribuerede processer

• Forskere og OMG arbejder på fremtidige standarder og teknologier for case management modelling, distribuerede processer og brugerinddragelse

22









Konklusion(er)• State-of-the-art tillader at beskrive procedurer som

brugere og forretningsfolk kan forstå

• I praksis bryder videnarbejderen ofte med den idéelle procedure og processer ændres før de tages i brug...

• Ringe support for compliancetjek, brugerinddragelse, deling & decentral udvikling af distribuerede processer

• Forskere og OMG arbejder på fremtidige standarder og teknologier for case management modelling, distribuerede processer og brugerinddragelse

22

Næste møde i videngruppe for digitalisering og processorientering (VidenDanmark.dk) 24. oktober - tema: Brugerinddragelse

Næste seminar i interessegruppen for Processer og IT (Infinit.dk) 13. november - Usabilityudfordringer i

offentlige digitaliseringsprocesser

(håber at se nogle af jer igen :-)






Fra Office Automation via BPM til tvær-organisatorisk Case Management og Citizen Development af...

Technology

Transcript of Fra Office Automation via BPM til tvær-organisatorisk Case Management og Citizen Development af...