Optimizing Our Data and Time

http://www.w3.org/2015/Talks/1019-shex-egp/

Eric Prud'hommeaux, <eric@w3.org>
- Sanitation Engineer, W3C

Data Value

Data utility depends on whether it

Directs our actions
- improves research
- saves money
- improves environment
- improves lives
- saves lives
Affects our confidence
Costs less than the value it provides

Data dimensions

Size of population
Coverage of population
Phenotypic breadth

Size/Structure (false) dichotomy

Data Availability

Permitted to see it -- license/policy
Physically reachable -- publication/distribution
Understandable -- published model/semantics
Machine-readable -- consistent, unambiguous representation
Temporally available -- use it before the sun burns out
Interconnectable -- consistent with other data

Lower barrier to investigation.
Data for far-sighted decisions.
Democratize innovation.

Data scale

Growing data size:

imaging: PET scans, CT scans, FMRI
molecular diagnostics
genome (3 billion base pairs)
biome (300 billion base pairs)

Growing data vectors:

next-generation (post-Sanger) sequencing. (name stuck to Moores law?)
online genomics services: Navigenics, 23andMe and Helix Health
online support communities: CureTogether, patientslikeme
virtual visits: American Well, Cisco HealthPresence, RelayHealth, Teladoc, Medfusion Virtual Office
remote clinical devices: RP-7i (robo-nurse), iPill
commodity clinics: minute clinic (CVS), take care clinic (Walgreens), RediClinic (H-E-B), Target Clinic, The Little Clinic
consumer devices: asthma-charter, USB chip for STDs, glucose meter, urinalysis

Semantic standards:

Reduce the n² pressure.

hundreds of partially-overlapping formats.
no forum for collaborative mapping.
many mappings are use-case-dependent.
reduce risk to entrepreneurs.
foster interesting innovation.

Semantic Web standards:

further reduce the n² pressure.

homogeneous medium for different standards.
enable distributed, progressive standardization.
reduce cost of standards.
tools and conventions apply to al standards.
- query - SPARQL
- description logic - OWL
- rules - RIF, SWRL
- validation - ShEx/SHACL

Clinical Informatics: connected models

information model
(Concepts and Relationships)
- Observations
- AdverseEvents
- Substance Administrations
on patient Isabella Jones
terminology model
(Terms and Relationships)
- Pneumonia a Pulmonary Disease
- Serum Sodium a lab test
- Penicillin an antibiotic
value sets
- Pneumonia, SLE, CHF
- mild, medium, severe
datatypes
- dates
- times
- intervals

Clinical model

Data dictionary, SDTM Domain, Case Report Form, ...

Simple (as possible) models of clinical knowledge.
Not generic - "systolic" instead of "observation opaque code"
Reflect causes and factors of disease.
Largely a rigorous model of common wisdom imparted in med school.
Developed in initiatives like CIMI.

FDA-TA

A Therapeutic Areas is a disease or family of diseases with similar mechanisms, targets or treatments.

Objective: Combine studies to look for small safety and efficacy signals.

Approach: Unify submission models for e.g. consistent variables and study info.

mission

map domain models to RDF
using BRIDG
inventing as little as possible
constraining to clinically significant
(where possible)

process

interview domain experts
construct concept mapts
encode as BRIDG
test pictorially

Mapping requirements

Moving data/queries between isosemantic models.

Clinical model to observation model:
- Describe coding practice/implementation guidelines.
- Demonstrate EMR use.
- Map clinical queries to EMR.
- Map app (à la SMART) queries.
Observation model to clinical model:
- Materialize study data.
- Materialize app (à la SMART) data.
Different observation models:
- Patient portability.
- Intra-department exchange.
- Warehousing/AHRQ registries.

Recognition models

Direct query - SPARQL/SQL
- + Very adaptable - fits any model with query facility.
- - Must enumerate representations.
- + Complexity: pretty low.
Description logic - OWL
- + Infer common representation.
- - Only extracts individuals -- need extra step.
- - Complexity: high.
Synthetic - SPARQL over OWL
- + Infer representation.
- ~ Report some attributes.
- - Little free tooling.
- - Complexity: high
Grammar validation - Shape Expressions
- + Higher-level than SPARQL
- - Low expressivity
- + Complexity: low
- + Can be reallized as SPARQL (higher complexity).

ShExMap

BP FHIR

<BPunitsDAM> {
  :systolic {
    :value xsd:float
        %Map{ bp:sysVal %},
    :units xsd:string
        %Map{ bp:sysUnits %}
  },
  :diastolic {
    :value xsd:float
        %Map{ bp:diaVal %},
    :units xsd:string
        %Map{ bp:diaUnits %}
  }
}

BP FHIR

<BPfhir> {
    a (fhir:Observation)?,
    fhir:coding { fhir:code (sct:Blood_Pressure) },
    fhir:related { fhir:type ("has-component"), fhir:target @<sysBP> },
    fhir:related { fhir:type ("has-component"), fhir:target @<diaBP> }
}
<sysBP> {
    a (fhir:Observation)?,
    fhir:coding { fhir:code (sct:Systolic_Blood_Pressure) },
    fhir:valueQuantity {
        a (fhir:Quantity)?,
        fhir:value xsd:float %Map{ bp:sysVal %},
        fhir:units xsd:string %Map{ bp:sysUnits %}
    },
}
<diaBP> {
    a (fhir:Observation)?,
    fhir:coding { fhir:code (sct:Diastolic_Blood_Pressure) },
    fhir:valueQuantity {
        a (fhir:Quantity)?,
        fhir:value xsd:float %Map{ bp:diaVal %},
        fhir:units xsd:string %Map{ bp:diaUnits %}
    },
}

Meta-mapping layers

Tools, languages and protocols for leveraging metadata:

11179 four-corners -- separate codes and concepts
UML, MOF
21090 -- model for code representation
SNOMED-CT upper level schema -- separate measurements from phenomena
FHIR-RDF add-ons -- infer only what you should from codes
Metadata Vocabulary for Tabular Data
SQL warehouse ETL, ShEx maps,