Lab Data Modeling Options
Lab LOINC MeetingJune , 2015
Indianapolis, Indiana
CIMI and HSPC
• Clinical Information Modeling Initiative (CIMI)– International organization for making “logical”
models for interoperability• Healthcare Services Platform Consortium (HSPC)– Defining service APIs to support truly interoperable
software
IsoSemantic Models – Example of Problem
e.g. “Suspected Lung Cancer”
(from Dr. Linda Bird)
Data Comes in Different Shapes and Colors
Finding – Suspected Lung Cancer
Finding – Suspected CancerLocation – Lung
Finding – CancerLocation – LungCertainty – Suspected(Let’s say this is the preferred shape)
Data Standardized in the Service
Shape and color of data in the local database
Shape and color translation
Application
Data in preferred shape and color
Application and User
Partial Interoperability
TermTranslators
Standard Terms(Non-standard Structure)
Application and User
Application
Local databases,CDA, HL7 V.2, etc.
Preferred Strategy – Full Interoperability
Local databases,CDA, HL7 V.2, etc.
Term andStructureTranslators
Application
Standard StructureAND Standard Terms
(As defined by CIMI Models)
Application and User
Requ
irem
ents
Reasons to do it on the server side
• Person writing the translation is most likely to understand the meaning of the data in their own database.
• The person writing the translation only has to understand their own data and the preferred model.– They can optimize query execution for their own system
• The query for the data is simpler. If the application has to write a query that will work for all shapes, the query will be inefficient to process by every system.
Reason for Modeling
• Models are defined so that when a program needs to access patient data it knows the precise logical structure of the data, and the exact codes that are allowed.
Request from CIMI to LOINC
• Create “methodless” LOINC codes where they don’t exist for the top 2000 lab LOINC codes
• This would be about 500 new LOINC codes
Options for Data Representation(Glucose in blood by manual test strip)
Option 1PreCoordinatedGlucoseModel name (focus): 2341-6 (Glucose [Mass/volume] in Blood by Test strip manual) data.value: 120 mg/dL
Option 2PostCoordinatedGlucoseModel 1 name (focus): 2341-6 (Glucose [Mass/volume] in Blood by Test strip manual) method (qual): 1234 Test strip manual data.value: 120 mg/dL
Option 3PostCoordinatedGlucoseModel 2 name (focus): 2339-0 (Glucose [Mass/volume] in Blood) method (qual): 1234 Test strip manual data.value: 120 mg/dL
Options
Option 4PostCoordinatedGlucoseModel 3 name (focus): (2341-6 OR 2339-0) --- Either a pre or a post coordinated code is allowed method (qual): 1234 Test strip manual data.value: 120 mg/dL
Option 5PostCoordinatedGlucoseModel 4 name (focus): (2341-6 OR 2339-0) --- Either a pre or a post coordinated code is allowed method (qual): 1234 Test strip manual component: Glucose specimen: BLD property: MCNC scale: QN data.value: 120 mg/dL
Raw LOINC CodesCode Component Prop System Scale Method 2339-0 Glucose MCnc Pt Bld Qn mg/dL Glucose
[Mass/volume] in Blood
2341-6 Glucose MCnc Pt Bld Qn Test strip manual
mg/dL Glucose [Mass/volume] in Blood by Test strip manual
2340-8 Glucose MCnc Pt Bld Qn Test strip automated
mg/dL Glucose [Mass/volume] in Blood by Automated test strip
QN Lab “Universal” - 1Example 1 – Quantitative lab tests with method and challengeAbstract Model
PostCoordinatedQuantLabUnivModel name (focus): [value set of QN Lab LOINC items where method and challenge is not specified] method (qual): [value set of all valid quantitative lab methods] challenge (mod): [value set of all valid challenge concepts] data.value.units: [value set of all quantitative lab units of measure]
Item to be represented: Glucose [Mass/volume] in Blood
Leaf level post coordinated QN lab model (no method) – CIMI preferred modelPostCoordinatedGlucoseModel name (focus): LN 2339-0 (Glucose [Mass/volume] in Blood) data.value.units: SCT 258797006 (mg/dL)
QN Lab “Universal” - 2Example 1 – Quantitative lab tests with methodAbstract Model
PostCoordinatedQuantLabUnivModel name (focus): [value set of QN Lab LOINC items where method and challenge is not specified] method (qual): [value set of all valid quantitative lab methods] challenge (mod): [value set of all valid challenge concepts] data.value.units: [value set of all quantitative lab units of measure]
Item to be represented: Glucose [Mass/volume] in Blood by Test strip manual
Leaf level post coordinated QN lab model and method – CIMI preferred modelPostCoordinatedGlucoseModel – (model meaning binding, id code at the model level, should be the same for both the pre and post coordinated models) name (focus): LN 2339-0 (Glucose [Mass/volume] in Blood) method (qual): CIMISCT 1111 Test strip manual (117021008 Test strip method) data.value.units: SCT 258797006 (mg/dL)
Leaf level pre coordinated QN lab model and method – CIMI iso-semantic modelPreCoordinatedGlucoseModel name (focus): LN 2341-6 (Glucose [Mass/volume] in Blood by Test strip manual) data.value.units: SCT 258797006 (mg/dL)
QN Lab “Universal” - 3Example 1 – Quantitative lab tests with challengeAbstract Model
PostCoordinatedQuantLabUnivModel name (focus): [value set of QN Lab LOINC items where method and challenge is not specified] method (qual): [value set of all valid quantitative lab methods] challenge (qual): [value set of all valid challenge concepts] data.value.units: [value set of all quantitative lab units of measure]
Item to be stored: Glucose [Mass/volume] in Serum or Plasma --1.5 hours post 0.5 g/kg glucose IV
Leaf level post coordinated QN lab model and challenge – CIMI preferred modelPostCoordinatedGlucoseChallengeModel name (focus): LN 2339-0 (Glucose [Mass/volume] in Blood) challenge (qual): CIMISCT 2222 (1.5H post 0.5 g/kg glucose IV) data.value.units: SCT 258797006 (mg/dL)
Lead level pre coordinated QN lab model and challenge – CIMI iso-semantic modelPreCoordinatedGlucoseChallengeModel name (focus): LN 1492-8 (Glucose [Mass/volume] in Serum or Plasma --1.5 hours post 0.5 g/kg glucose IV) data.value.units: SCT 258797006 (mg/dL)
QN Lab “Universal” - 4Example 1 – Quantitative lab tests with mehtod and challengeAbstract Model
PostCoordinatedQuantLabUnivModel name (focus): [value set of QN Lab LOINC items where method and challenge is not specified] method (qual): [value set of all valid quantitative lab methods] challenge (qual): [value set of all valid challenge concepts] data.value.units: [value set of all quantitative lab units of measure]
Item to be stored: Glucose [Presence] in Urine by Test strip --1.5 hours post 75 g glucose PO
Leaf level post coordinated QN lab model method and challenge – CIMI preferred modelPostCoordinatedGlucoseChallengeAndMethodModel name (focus): LN 2349-9 (Glucose [Presence] in Urine) method (qual): CIMISCT 1111 Test strip manual (117021008 Test strip method) challenge (qual): CIMISCT 3333 (1.5H post 75 g glucose PO) data.value.code: [value set binding for Present/Absent]
Lead level pre coordinated QN lab model method and challenge – CIMI iso-semantic modelPreCoordinatedGlucoseChallengeAndMethodModel name (focus): LN 6763-7 (Glucose [Presence] in Urine by Test strip --1.5 hours post 75 g glucose PO by Test Strip) data.value.code: [value set binding for Present/Absent]
Rationale for Request
• To support the CIMI preferred way of representing data we need “methodless” LOINC codes
• These codes would sometimes be useful for ordering
Post Coordinated Specimen Lab “Universal” - 1
Common situations where specimen type is XXX: microscopic exams, antigens, antibodies, DNA, RNA, cultures (need to verify that there is an automated method for determining these classes of items)
Example 1 – Lab test with post coordinated specimen and methodAbstract Model
PostCoordinatedSpecimenLabUnivModel name (focus): [value set of labs where test is one of the types listed above] specimen (qual): [value set of all valid specimen types] method (qual): [value set of all valid lab methods] data.value.code: [value set binding]
Item to be stored: Arthropod identified in Unspecified specimenLeaf level post coordinated lab model specimen and method – CIMI preferred model
PostCoordinatedArthropodIdentifiedModel name (focus): 10644-3 (Arthropod identified ) data.value.code: [value set binding for Arthropod Answers]
Post Coordinated Specimen Lab “Universal” - 2Example 1 – Lab test with post coordinated specimen and methodAbstract Model
PostCoordinatedSpecimenLabUnivModel name (focus): [value set of labs where test is one of the types listed above] specimen (qual): [value set of all valid specimen types] method (qual): [value set of all valid lab methods]
Item to be stored: Arthropod identified in Unspecified specimen by Permanent mount
Leaf level post coordinated lab model specimen and method – CIMI preferred modelPostCoordinatedArthropodIdentifiedModel name (focus): 10644-3 (Arthropod identified) method (qual): CIMISCT 3333 (Permanent mount) data.value.code: [value set binding for Arthropod Answers]
Leaf level precoordinated lab model specimen and method – CIMI isosemantic modelPreCoordinatedArthropodIdentifiedPermMountModel name (focus): 14787-6 (Arthropod identified in Unspecified specimen by Permanent mount)
data.value.code: [value set binding for Arthropod Answers]
Post Coordinated Specimen Lab “Universal” - 3Example 1 – Lab test with post coordinated specimen and methodAbstract Model
PostCoordinatedSpecimenLabUnivModel name (focus): [value set of labs where test is one of the types listed above] specimen (qual): [value set of all valid specimen types] method (qual): [value set of all valid lab methods]
Item to be stored: Chlamydia trachomatis DNA [Presence] in Cervix by Probe and target amplification method
Leaf level post coordinated lab model specimen and method – CIMI preferred modelPostCoordinatedCTrachomatisDNACervixProbe name (focus): CIMISCT 1234 (Chlamydia trachomatis DNA [Presence]) specimen (qual): SCT 110949001 (Cervix cytology material) method (qual): CIMISCT 5555 (Probe and target amplification) data.value.code: [value set binding for Present/Absent]
Leaf level precoordinated lab model specimen and method – CIMI isosemantic modelPreCoordinatedCTrachomatisDNACervixProbe name (focus): 21190-4 (Chlamydia trachomatis DNA [Presence] in Cervix by Probe and target amplification method)
data.value.code: [value set binding for Present/Absent]
Top Related