Skin domain Pressure Ulcer Prevention Aanalysis Model

Approach

Concept

We will generate an HL7 Domain Analysis Model based on existing work. This model will later be constrained to a model (a Domain Information Model) that will support the generation of a variety of implementable artifacts—HL7 V2 and V3 messages, CDA documents, functional profiles, and any specification necessary for the capture, storage, interpretation, and exchange of information in the domain.

As the KP-VAdraft includes most of the LOINC Nursing Group content, and as the KP-VA project white paper indicates an intention to use standards development organizations to further the validation and adoption of the model, we will use the KP model as a baseline. We will assess differences between the two models to determine whether LNP approaches should be adopted in the KP model.

The conceptual scope includes two packages in the KP-VA model: Pressure Ulcer Risk Assessment and Pressure Ulcer Prevention Intervention. The name of the model has been changed from Skin Assessment to Pressure Ulcer Prevention.

The production of Detailed Clinical Models (DCMs)as a primary goal depends on an understanding of the product that is not possible at this point. Candidate DCMs may be identified in the DAM, time permitting.

The creation of a CDA specification is likely to be feasible in the near term. This project will support efforts for CDA analysis, but will not commit to including a CDA specification in the May ballot.

Key Gaps

The current model includes a great deal of detail in the information model, but because the process used to generate this detail depended heavily on the expertise of the participants, the proper scope is difficult to confirm. The primary objective in the early stages will be to model use cases that will provide scope to the information model, confirming the procedural and conceptual boundaries, level of detail, and metadata required to make the model useful.

Since much of the information model seems to have been generated from concept-harmonization discussions rather than process discussions, the second key activity will be to complete the information model with process-related information to support the use cases.

The third key activity will be to bring the model into conformance with a consistent metamodel by, e.g., assigning data types, transforming enumerations into coded values, creating smaller and more legible class diagrams, etc.

Finally, much of the information in the model supports other external requirements (quality instruments, research terms, etc.) but does so in terminology crafted to meet the needs of clinical nurses. This effort will, where possible, document the tracing to those source specifications.

Process

  1. Determine the deliverable metamodel
  2. Define use cases (candidates follow)
  3. Patient transfer between KP and VA using IHE e-Nursing Summary
  4. CMS Minimum Data Set (MDS)
  5. IHE profile nursing summaryOther quality measures
  6. Model use cases
  7. Define information artifacts
  8. At this point, CDA transformation may be drafted
  9. Harmonize class model with existing information sources
  10. Document tracing to source materials
  11. Update class model with LNP project changes as appropriate
  12. (optional) DCM
  13. Identify candidate DCM(s)
  14. Apply semantic constraints from LNP project

Metamodel

Dynamic model

Use caseswill be represented as model elements. Each model element will include a textual description and a trace to the source requirement (e.g., MDS).

Actorsare defined by use cases. Each actor will include a textual description.

Activity diagrams provide medium-level representations of the activities within use cases. Certain steps will require the use of information objects. Each diagram will represent one use case.

Information Objects define the information exchanged at identified steps in the activity diagrams.

Static Model

Classes represent the entities of interest. They include definitions.

Relationships indicate how classes are related, by generalization, where one class is a more specific type of another more general class, or by association, where the precise nature of the association is indicated by a textual label. The label also includes an arrow pointing from the subject to the predicate of the label. Associations also include cardinalities.

Properties will include properties other than coded concepts (e.g., times, measurements). They will be given definitions and assigned HL7 datatypes.

Properties may include tracings to source materials, modeled as constraints. A constraint type of Source Specification will be used.

Selected packages may be identified as candidate DCMs. These packages will be small and potentially reusable. Coded properties will be bound to specified sets of concept identifiers. The manner of binding is to be determined, but it may involve the use of constraints. We do not anticipate using enumerations, as enumerated values are more volatile than the models they inform: maintenance and synchronization work can be minimized by decoupling the model from the vocabulary specification.

Scope

Candidate Use Cases

  1. Kaiser-Permanente / VA record exchange for patient transfer
  2. As specified by project participants
  3. Using Optimum Data Set as defined in October
  4. Using resources as available from IHE e-Nursing Summary framework
  5. Is there a specification? A VistA template? Element list provided by Mona Baharestani?
  6. CMS Minimum Data Set (MDS) is part of a comprehensive Resident Assessment Instrument, moving from assessment (based on MDS) to decision-making, care plan development, care plan implementation, and evaluation.
  7. To what extent do we need to analyze downstream elements, e.g., triggers, care plan, etc?
  8. Other quality measurement instruments, as feasible

3.IHE e-Nursing Summary (ENS)

  1. The IHE case includes sections, but no details about information

Key Concepts

Represent the difference between planned and accomplished actions.

Represent the difference between observations documented explicitly, in a “narrative” style, and those documented implicitly, in an “exception-based” style.

Represent the model as clearly as possible, possibly with extensive descriptive front matter, in order to accommodate the needs of non-technical readers.

Reference literature establishing the prevalence and significance of pressure ulcers, possibly referring to efforts to establish this topic in the Meaningful Use framework.

1