BEHAVIOR PIPELINE LIBRARY

• Smoking Status Detection (Smoking_status)

This is dictionary and rule based pipeline to extract smoking status from notes. It will recognize all smoking mentions and then category them into ‘NonSmoker’, ‘PastSmoker’ or ‘Smoker’;

Smoking-comp
Smoking-status Pipeline Components
Smoking-out
Smoking-status Pipeline Output

DISEASE_SYMPTOM PIPELINE LIBRARY

• Bleeding Mentions (Bleeding_extraction)

This pipeline will extract bleeding related concept and then map them to the UMLS CUI;

bleed-comp
Bleeding_extraction Pipeline Components
• Colorectal Cancer Concept Extraction (Colorectal_cancer)

This pipeline will extract colorectal cancer mentions and then map them to the UMLS CUI;

colorectal-comp
SColorectal_cancer Pipeline Components
• Cancer Pathology Pipeline (PathoPipelineFinalML)

TThis pipeline will extract Site, Procedure, Histology etc. entities and relations among them from pathology notes.

ppfml-comp
PathoPipelineFinalML Pipeline Output

GENERAL PIPELINE LIBRARY

• General Clinical Concept Extraction (Clamp-ner)

This is the CLAMP’s default named entity recognition pipeline. It will recognize ‘problem’, ‘treatment’ and ‘test’ from clinical notes and negation information of each concept. Users can add the UMLS encoder component to the end of this pipeline to get the UMLS CUI concept id (UMLS account is required for this component).

Clamp-ner-comp
SClamp-ner Pipeline Components
Clamp-ner-out
Clamp-ner Pipeline Output
• General Clinical Concept and Attribute Extractione (Clamp-ner-attribute)

This is the default named entity recognition and relation extraction pipeline. The primary entities include ‘problem’, ‘treatment’ and ‘test’. For each entity, the pipeline will recognize its attribute as well.
• Problem with: subject, condition, negation, severity, location, uncertainty;
• Lab test with: test value;
• Medicine with: dose, form, route, frequency, duration, necessity;

Ner-att-comp
Clamp-ner-attribute Pipeline Components
Ner-att-out
Clamp-ner-attribute Pipeline Output
• De-identification Pipeline (Deid_pipeline)

This is the de-identification pipeline. It will recognize all PHI information and then replace them with placeholder strings that are defined by the users. It contains 3 sub pipelines which are Disease-attribute, Lab-attribute and Medication-attribute.

Deid-comp
Deid Pipeline Components
• Disease Attribute Extraction (Disease-attribute)
disease-att-comp
Disease-attribute Pipeline Components
• Lab Attribute Extraction (Lab-attribute)
lab-att-comp
Lab-attribute Pipeline Components
lab-att-out
Lab-attribute Pipeline Output
• Medication Attribute Extraction (Medication-attribute)
med-att-comp
Medication-attribute Pipeline Components
• Temporal Phrase Extraction (Temporal-attribute)
temp-att-comp
Temporal-attribute Pipeline Components

Contact US

The CLAMP System was developed by Dr. Hua Xu’s team from the School of Biomedical Informatics at the University of Texas Health Science Center in Houston.
For technical issues, please contact: Jingqi.Wang@uth.tmc.edu .
For any other issues, please contact: Anupama.E.Gururaj@uth.tmc.edu .

Contact Us

Center for Computational Biomedicine

School of Biomedical Informatics

The University of Texas Health Science Center at Houston

7000 Fannin St, Houston, TX 77030

Research Coordinator

Anupama E. Gururaj, PhD

anupama.e.gururaj@uth.tmc.edu

713-500-3619

Technical Support

Jingqi Wang, MS - Analyst Programmer

jingqi.wang@uth.tmc.edu

713-500-3620

Follow Us


follow me on twitter follow me on facebook follow me on LinkedIn