12
SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

Embed Size (px)

Citation preview

Page 1: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

SHARPn High-Throughput

Phenotyping (HTP)November 18, 2013

Page 2: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

Electronic health records (EHRs) driven phenotyping

• EHRs are becoming more and more prevalent within the U.S. healthcare system• Meaningful Use is one of the major drivers

• Overarching goal• To develop high-throughput semi-automated

techniques and algorithms that operate on normalized EHR data to identify cohorts of potentially eligible subjects on the basis of disease, symptoms, or related findings both retrospectively and prospectively

©2013 MFMER | slide-2

Page 3: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

DataTransformTransform

EHR-driven Phenotyping Algorithms – The Process

PhenotypeAlgorithm

Visualization

Evaluation

NLP, SQL

Rules

Mappings[eMERGE Network]

©2013 MFMER | slide-3

Page 4: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

Key lessons learned from eMERGE• Algorithm design and transportability

• Non-trivial; requires significant expert involvement• Highly iterative process• Time-consuming manual chart reviews• Representation of “phenotype logic” is critical

• Standardized data access and representation• Importance of unified vocabularies, data elements, and value sets• Questionable reliability of ICD & CPT codes (e.g., billing the wrong

code since it is easier to find)• Natural Language Processing (NLP) plays a vital role

©2013 MFMER | slide-4

[Kho et al. Sc. Trans. Med 2011; 3(79): 1-7]

Page 5: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

DataTransformTransform

Algorithm Development Process - Modified

PhenotypeAlgorithm

Visualization

Evaluation

NLP, SQL

Rules

Mappings

Semi-Automatic Execution

©2013 MFMER | slide-5

• Standardized representation of clinical data

• Create new and re-use existing clinical element models (CEMs)

• Standardized and structured representation of phenotype definition criteria

• Use the NQF Quality Data Model (QDM)

• Conversion of structured phenotype criteria into executable queries

• Use JBoss® Drools (DRLs)

[Welch et al., JBI 2012; 45(4):763-71]

Page 6: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs ©2013 MFMER | slide-6

[Thompson et al., AMIA 2012]

Page 7: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs ©2013 MFMER | slide-7

[Li et al., AMIA 2012]

Page 8: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

[Endle et al., AMIA 2012]

http://phenotypeportal.org

Page 9: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

Page 10: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

Phenotype Modeling and Execution Architecture (pheMA): New 4-year NIH R01

©2013 MFMER | slide-10

Page 11: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

Modeling

Reviewing

Evaluation

One algorithm modeled by two individual MAT & QDM experts

Measures reviewed by three individual domain experts(Comparison for the two versions of the measure)

QDM & MAT extension

Measure Authoring Tool

Gold Standards

eMERGE phenotypes:T2DM; Resistant Hypertension; Hypothyroidism; Cataracts; Diabetic Retinopathy; PAD; Dementia; VTE; Glaucoma; Ocular Hypertension Continuous variable phenotypes QRS duration from ECG; Lipids (inc. HDL); Height; RBC; WBC

Standards to validate and compare the created measures: • Measure how concise of the measure is (more concise is better)• Measure is true to the algorithm• Measure how much existing values sets and measures are re-used• Measure how much time it took to implement in MAT• Measure how many rules in the MAT version vs. Word document• Considerations:

how experienced the person was w/ MAT to start, and for ea. phenotype as gain more experience make note of it

how well the person knew the phenotype to start

Plan for Aim 1: Evaluation of Quality Data Model

Page 12: SHARPn High-Throughput Phenotyping (HTP) November 18, 2013

High-Throughput Phenotyping from EHRs

Plan for Aims 2 & 3: National Library of Computable Phenotyping Algorithms

©2013 MFMER | slide-12