DIAGNOSTIC CRITERIA AND CLINICAL GUIDELINES STANDARDIZATION TO AUTOMATE CASE CLASSIFICATION
Mélanie Courtot, Ph.D. candidate Terry Fox Laboratory, BC Cancer Agency International Conference on Biomedical Ontology, July 8th 2013
Key points
• The Adverse Event Reporting Ontology (AERO) is to be used for
(1) Encoding guidelines (2) Adverse event reports based diagnosis (3) Data integration
(1) AERO for encoding guidelines
AND
AustralasianSociety of Clinical Immunology
and Allergy
National Institute for Health and Clinical
Excellence
BrightonCollaboration
(Level 1)
OR ONE OF
World Allergy Organization
AND/OR
AND/OR
VARIOUS COMBINATIONS OF
generalized urticaria or generalized erythema finding
angioedema finding
generalized pruritus with skin rash finding
clinical diagnosis of uncompensated shock
respiratory distress diagnosis
bilateral wheeze finding
stridor finding
upper airwayswelling finding
skin and mucosalchanges
involvement of the skin and/or mucosal tissue
(e.g. generalized hives, itching or flushing,
swollen lips-tongue-uvula)
persistent dizzinesscollapse
difficulty talkinghoarse voice
wheeze orpersistent cough
difficult/noisy breathing
swelling of the tongueswelling/tightness in throat
pale and floppy(young children)
circulation problem (hypotension
and/or tachycardia)
breathing problem(bronchospasm
with tachypnoea)
problems involving the airway
(pharyngeal or laryngeal)
Reduced BP or symptoms of end-organ dysfunction such as hypotonia, incontinence
Respiratory symptoms such as shortness of breath,
wheeze,cough, stridor, hypoxemia
sudden gastrointestinal syndromes such as crampy abdominal
pain, vomiting
DERM
OTO
LOG
ICAL
MUC
OSA
LCA
RDIO
VASC
ULAR
RESP
IRAT
ORY
OTH
ERS
OR
OR
OR
OR
OR
AND
OR
OR
OR
OR
measured hypotensionOR
Clinical guidelines in AERO • Surveillance Goals
• Provide a pattern to encode guidelines for adverse event reporting following immunization
• Make this pattern applicable to any type of clinical guideline
• Provide a means for the reports to be annotated with diagnosis according to a specific guideline (and keep track of which)
• Implementation Goals • Encode the guideline in OWL • Be able to infer correct classification (i.e., perform
accurate diagnosis)
Current status • Pattern for anaphylaxis clinical guideline according to the Brighton Collaboration has been implemented in OWL
• Colleague Dr. Jie Zheng has modeled WHO malaria clinical guidelines using the same pattern
Jie Zheng University of Pennsilvania
Brighton Collaboration: https://brightoncollaboration.org
(2) AERO for adverse event report based diagnosis
VAERS dataset • VAERS = Vaccine Adverse Event Reporting System
• Administered by the Centers for Disease Control and Prevention (CDC) and the Food and Drug Administration (FDA) in the United States
• A spontaneous reporting system • spontaneous reporting systems have issues with
underreporting and quality • MedDRA (Medical Dictionary of Regulatory Activities) is used to represent clinical findings
Free text partof the report
MedDRA encodedstructured data
Example VAERS report
Working with classified VAERS data • Unclassified files available publicly • Classified dataset available only upon request
• FDA provided dataset of classified adverse events following H1N1 immunization in winter 2009-2010
• FDA classified reports according to the Brighton case definitions
A test of ontology-based method 1. Map the current Brighton terms in AERO
to their MedDRA counterpart 2. Use a reasoner to classify the MedDRA-
annotated reports using the Brighton criteria
3. Compare results with FDA classification done by medical experts
Current status
• Created MedDRA –> Brighton mapping, in OWL, covering anaphylaxis guideline
• Tested classification of VAERS reports
(3) AERO for data integration
The semantic web • From a web of documents to a web of data • HTML pages can’t be understood by machines; humans have to manually follow hyperlinks
• Semantic web uses standard for data representation, querying, vocabularies to link data behind the scenes
• Use of Uniform Resources Identifiers (URIs) and Resource Description Framework (RDF)
VAERS as linked data • Transform the VAERS dataset in RDF to enable better integration with existing linked data
• Avoids typical need to worry about resources’ structure (CSV, databases, XML)
• Approach • VAERS reports are OWL individuals • RDF is generated using FuXi (python) from a relational
database I constructed from VAERS flat files
Link to ontology terms
A simple example: Change state code in VAERS to state URI in DBPedia Query against DBPedia to help prepare call to Google visualization API
Potential uses of query across linked data sets
• Using the VAERS dataset • Are there differences in the type of adverse events
between a live attenuated flu vaccine and a trivalent inactivated one?
• Using another dataset: DrugBank • Link to DrugBank based on drug mentions in text (e.g.
“Benadryl”) • Retrieve therapeutic class from DrugBank • In cases where therapeutic class is anti-allergic agents
infer that the patient may have had an allergic reaction.
Acknowledgements • Alan Ruttenberg, Ryan Brinkman • Jie Zheng, Chris Stoeckert • Julie Lafleche, Lauren McDonald, Robert Pless,
Barbara Law, Jan Bonhoeffer, Jean-Paul Collet • Oliver He, Yu Lin, Lindsay Cowell, Barry Smith, Albert
Goldfain
Mélanie Courtot [email protected] http://purl.obolibrary.org/obo/aero