Translating Medical Research for Patient Findability
Transcript of Translating Medical Research for Patient Findability
©2016 MFMER | slide-1
Translating Medical Research for Patient FindabilityMichael Panzer | Mike Hatton
Taxonomy Boot Camp 2016, Washington D.C.14 November 2016
©2016 MFMER | slide-2
Professional Background
• Large-scale classification systems• Semantic web technology• Long-time involvement with the
Dewey Decimal Classification• Editor-in-Chief until 2015• Notorious linked data evangelist
• Currently• Manager-Ontologist at Mayo Clinic• 9 person team of (Sr) Ontologists
dewey.info
©2016 MFMER | slide-3
Mayo ClinicLocations
©2016 MFMER | slide-4
©2016 MFMER | slide-5
Outline
1. Situation/Background• Status quo ante for the publication of
clinical trials
2. Designing a metadata framework• Leveraging standard clinical vocabularies• Developing a clinical study ontology
3. Evaluating outcomes• Findability, interoperability• Desiderata
©2016 MFMER | slide-6
What is a Clinical Trial (Clinical Study)?
• “A research study using human subjects to evaluate biomedical or health-related outcomes. Two types of clinical studies are Interventional studies (or clinical trials) and Observational studies.”
©2016 MFMER | slide-7
Publication of Clinical Trials
• Situation: People couldn’t find trials, study coordinators couldn’t find them either!
• Background: Antiquated browse system based on tags assigned by developers
• Focus on benefactors, not study purpose• Started using MeSH (only UF)• Trial summary was written for specialists
• Clinical trial management part of larger reworking of comprehensive research management
©2016 MFMER | slide-8
External Web
epiCenterCore Study Data
Mayo sponsored studies -More information to populate registration, including SNOMED terms
Clinicaltrials.gov
Sponsored studies
Ontologist
Core Study Data Capture Process Flow
KCMS
©2016 MFMER | slide-9
Outline
1. Situation/Background• Status quo ante for the publication of
clinical trials
2. Designing a metadata framework• Developing a clinical study ontology• Leveraging standard clinical vocabularies
3. Evaluating outcomes• Findability, interoperability• Desiderata
©2016 MFMER | slide-10
Career Opportunity
isLocatedIn LocationisLocationOf Research
Study
Organization
Person
Information Resource
Clinicial Trial
Subclass of
isLocationOf
isLocatedIn
isLocationOf
isLocatedIn
conducts
isRelatedResourceOf
relatedResource
conductedBy
offersisOfferedBy
Research Web Domain Model
©2016 MFMER | slide-11
Clinicial TrialprimaryCondition
secondaryConditionintervention
SNOMED Concept
skos:closeMatch
Consumer Health
Vocabulary Concept
associatedBodySystem Anatomical Structure
Medical Specialty
associatedMedicalSpecialty
OrganizationisConductedBy
associatedProcedureassociatedCondition
RxNorm Concept
investigatesDrugIntervention
Annotation Design
©2016 MFMER | slide-12
Annotation Design
Annotation Property Required
Repeatable
Vocabulary Provide to
epiCenter?
SNOMED synonyms needed?
Primary Condition N Y SNOMED CT
Y Y
Secondary Condition N Y SNOMED CT
Y Y
Intervention N Y SNOMED CT
Y Y
Drug Intervention N Y RxNorm Y YAssociated Condition
N Y CHV Y N
Associated Procedure
N Y CHV Y N
Medical Specialty N Y SNOMED > Healthcare specialty
Y N
©2016 MFMER | slide-13
Vocabulary selection
• SNOMED CT• Comprehensive clinical health terminology• ~300,000 concepts• Synonymy• Mappings to CHV
• CHV (Consumer Health Vocabulary)• Set of consumer-oriented concepts of
conditions, procedures, symptoms, devices, anatomy
• ~5,000 concepts, rich set of relationships
©2016 MFMER | slide-14
Vocabulary selection
• RxNorm• Normalized names for clinical drugs and
links to many drug vocabularies• Generic and branded drugs• Part of UMLS, edited by NLM
©2016 MFMER | slide-15
Vocabulary application
• Primary subject captures the primary condition the study is researching to finest granularity
• Secondary subject captures other conditions• “Diabetic neuropathy in patients with
untreated diabetes”
©2016 MFMER | slide-16
©2016 MFMER | slide-17
Questions that can be answered by the design• What SNOMED codes are annotated to this study?• What CHV terms are annotated to this study?• What organizations are associated to this study?• What medical specialty terms are associated to this
study?• What body systems are associated to this study?• What are the SNOMED synonyms associated with
this study?
©2016 MFMER | slide-18
Outline
1. Situation/Background• Status quo ante for the publication of
clinical trials
2. Designing a metadata framework• Leveraging standard clinical vocabularies• Developing a clinical study ontology
3. Evaluating outcomes• Findability, interoperability• Desiderata
©2016 MFMER | slide-19
Interoperability benefits
• Integration for trials with health information content
• Condition and procedure topic pages• Search
• Clinical study header service• Integration with clinical departments
• Emphasis on recall rather than precision• Example: Endocrinology
• Integration with research centers
©2016 MFMER | slide-20
mayoclinic.org mayo.edu
Integration overview
©2016 MFMER | slide-21
Examples
• mayoclinic.org: Condition• mayoclinic.org: Procedure• mayoclinic.org: Department• mayo.edu/research: Research organization
©2016 MFMER | slide-22
Outcomes
• Cancer Center saw increase from 3500 inquiries a year to 5000 after improvements went live
• Inquiry volume continues to increase• Deeper dive into a random sample of calls from
Mar. 2015–Mar. 2016• 1921 inquiries• 20% resulted in appointments• 25% of those resulted in signed consents (5%
overall)
©2016 MFMER | slide-23
Questions & Discussion
©2016 MFMER | slide-24
Translating Medical Research for Patient FindabilityMichael Panzer
Taxonomy Boot Camp 2016, Washington D.C.14 November 2016