Medilink: Guidelines via Workflow

1
MedilinkGuidelines via Workflow

• Lucy Hederman
• Vinny Wade

2
Outline

• Clinical guidelines, representation, automation.
• Workflow,
• Role of workflow in the Medilink framework

3
Clinical Guidelines

• Recommended strategies for managing healthcare in
  specific clinical circumstances.
• Evidence-based best practice
• Largely text-based to date.
• Any flowcharts are usually ambiguous.
  
• Poorly and slowly adopted.

4
Guideline Automation

• Need guideline-based computersied decision
  support at the point of care
• Range of approaches
• Therapy-planning
• Process-oriented
• Need a guideline representation model

5
Guideline Tasks (Tu and Musen)

• Make decisions
• Sequence actions
• Set goals
• Maintain BP lt 130/85 for DM patients
• Interpret data
• Severity of asthma from freq, lung function
• Refine actions
• Operationalise an abstract intervention

6
Guideline Repn. Reqmts. Peleg

• Precise, unambiguos, executable
• Human readable
• Support dift. kinds of guideline
• Support difft. modes of use
• Sharable (to reuse effort) Boxwala
• Adaptable
• Integratable

7
Guideline Repn Models Survey Wang

• Primitives
• Decisions, actions
• Patient state of execution state or both
• Temporal constraint spec
• Flowcharts, state transition diagrams or
  constriant satisfaction graphs
• Nesting
• Abstraction, hiding complexity

8
Models Surveyed by Wang

• Arden Syntax (MLMs) single decisions
• DILEMMA
• EON - very knowledge-based
• PROforma - active decision support and workflow
  managment
• GLIF3
• Asbru - for AI-planning style system
• GUIDE
• PRODIGY

9
GLIF

• Intermeds GuideLine Interchange Format
• Aims to become a standard shared representation
• Tries to support spec. of guidelines that differ
  in various ways.
• GLIF3 recently released executability
• HL7 is likely to adopt GLIF

10
GLIF Levels of Abstraction

• Conceptual flowchart
• Computable/parsable specification
• Expression syntax, defns. of patient data items,
  clinical actions, flow of the algorithm.
• Consistency and completeness checks possible
• Implementable specification
• Incorporated into specific institutional systems
• Loses part of its sharability

11
GLIF3 Steps

• Patient state step
• E.g., possible plasma glucose emergency
• Action step
• Could be a sub-guideline (nesting)
• Decision steps
• Case step automatable branching
• Choice step physicians decision point
• Branching and synchronization steps

12
GLIF3 Computable level

• Interpretable by computer
• Logical consistency and completeness can be
  analysed.
• Uses Arden syntax for decision criteria and
  patient states.
• Domain object model
• Patient data items are defined by a controlled
  terminology

13
Diabetes diagnosis in GLIF
14
Subguideline DM diagnosed
15
Subguideline possible plasma glucose emergency
16
Over to Vinny
17
Workflow at a glance.

• defines, manages and executes processes .
• driven by a computer representation of the
  process logic.
• WfMC 2001

18
Process

• one or more linked procedures or activities
  which collectively realise a business objective
  or policy goal, normally within the context of an
  organisation stucture defining functional roles
  and relationships

19
Workflow goes mainstream ader

• Emergent within
• Enterprise Application Integration
• Active Databases
• Business Object Management Systems
• Process automation/management systems
• CASE Tools
• International Standardisation
• OMG, WfMC, W3C

20
Key Research Areas in Workflow

• Process Modelling Specification
• Process Automation Management

21
Process Representation

• Traditional workflow representation focused on
  low-level implementation mechanisms e.g.
  petri-nets etc.
• Need to move toward common modelling language
  (notation) for capturing reqd control data flow

22
Process Representation Reqment

• Precise
• Executable
• Support multiple workflow patterns
• Understandable by human readers

23
Process Representation in UML

• UML universal software development modelling
  language
• Consist of 11 different notations for different
  aspects of software development cycle e.g.
  business modelling, requirements modelling etc.
• Activity Graphs currently being enhanced for
  workflow representation
• UML v1.4, and in 2002/3 UML v2.0

24
Main diabetes diagnosis guidelines as a UML
activity diagram
25
Possible emergency sub-guideline in UML
26
Diabetes diagnosed sub-guideline in UML
27
Workflow Management Automation

• the management automation of a process (in
  whole or in part) duing which documents,
  information and tasks are passed from one
  participant to another for action according to a
  set of rules
• 
  WfMC2001

28
Business Process Integration Framework
29
EJB Framework
30
Serene workflow framework
31
Medilink Draft Architecture

• Synapses feeds data from DBs and components that
  provide patient data to components that need
  data.
• Care process (workflow) engine invokes activities
  in accordance with the process rule base.

32
Draft Architecture
Protocols, plans and logs
Order
Prescrib
Protocol Automation
CBR
Schedul
LIS
Care Process Knowledge
Synapses
Care Process Engine
PAS
Process Exec. Log
Other Pat. DB
Report
MODEL
SIMUL
Analyse
Other decision support components

invocation
data flow
33
Care Process Engine
Handles patient to process instance mapping
Workflow Engine (Serene)
Care Process Knowledge
Care Process Manager
Start/stoppatient on process
Process Exec. Log
Patient process data
Process instance data
34
Role of workflow

• Engine sequences activities
• Process knowledge may
• make decisions
• interpret data
• refine actions
• or these tasks may be embedded in components
  invoked by the engine.

35
Next Steps

• Refine the draft architecture
• Specify data and invocation interfaces
• Design scenarios which demonstrate Medilink
  integration
• Encode scenarios as processes
• Develop necessary prototype components

36
References

• Boxwala, A., S. Tu, Q. Zeng, M. Peleg, O.
  Ogunyemi, R. Greenes, E. Shortliffe, V. Patel,
  Towards a representation format for sharable
  clinical guidelines, Journal of Biomedical
  Informatics, 2001 (submitted). Available as
  Stanford Tech Report SMI-2001-0896. 
• Peleg, M., A. Boxwala, O. Ogunyemi, Q. Zeng, S.
  Tu, E. Bernstam, L. Ohno-Machado, E. Shortliffe,
  R. Greenes. GLIF3 The Evolution of a
  Guideline Representation Format. AMIA Annual
  Symposium, Los Angeles, CA, (20 Suppl)645-649.
  2000.
•  Wang, D. A., M. Peleg, S. Tu, E. H. Shortliffe,
  R. A. Greenes. Representation of Clinical
  Practice Guidelines for Computer-Based
  Implementations. Medinfo, London, UK 285-289.
  2001.
• Tu, Samson, Mark Musen, Representation
  Formalisms and Computational Methods for Modeling
  Guideline-Based Patient Care, EWGLP, Leipzig
  125-141. 2000.

37
UML Activity Diagrams

• Are special case of UML State Diagrams which in
  turn, are graphical representations of state
  machines
• State machines are transition machines whose arcs
  are labeled by ECA rules