ihtsdo document · web viewwe wish to thank raj dash, mary kennedy, alexis carter and ted carithers...

SNOW SHRINE metadata release notes

A service of the Greater Plains Collaborative Research Network

Date 20180330Version 20181001

Amendment History

Version Date Editor Comments

20180330 20180330James CampbellJay Pedersen

First release of SNOW SHRINE metadata

20181001 20181207James CampbellJay PedersenYeshwanth Narayana

Second release of SNOW SHRINE metadata; revisions to ICD-10-CM, RxNorm, NDC, SNOMED CT, LOINC

Review Timetable

Review date Responsible owner Comments

20180330 James R. Campbell

© University of Nebraska Medical Center Nebraska Medicine

SNOW SHRINE metadata release notes of 16

Table of Contents

1 Purpose..................................................................................................................................5

2 Audience of interest for this document...................................................................................5

3 Content scope and history......................................................................................................5

4 Quality assurance, versioning and maintenance....................................................................6

5 Description of Release Datasets and Instructions for use......................................................7

5.1 Reference copy of SCILHS 3.1b; renamed Accessible Research Commons for Health (ARCH) Model Ontology...................................................................................................................7

5.2 SNOW SHRINE metadata files; FULL_YYYYYMMDD release (versions 20180330 and 20181001)............................................................................................................................................7

5.2.1 Pcornet_Demo_oracle.zip: ARCH (SCILHS 3.1b) metadata for DEMOGRAPHICS.............................75.2.2 Pcornet_diag.zip: ARCH (SCILHS 3.1b) metadata for DIAGNOSES; modified v20181001 to include 124503 SNOMED CT conditions, 93522 ICD-10-CM codes from 20181001 release; SNOMED CT metadata searching supported via transitive closure tables............................................................................................75.2.3 Pcornet_enc.zip: ARCH (SCILHS 3.1b) metadata for ENCOUNTERS..................................................75.2.4 Pcornet_enroll.zip: ARCH (SCILHS 3.1b) metadata for ENROLLMENT................................................75.2.5 Pcornet_lab.zip for LABORATORY and PATHOLOGY..........................................................................75.2.6 Pcornet_med.zip for MEDICATIONS.....................................................................................................75.2.7 Pcornet_proc_nocpt.zip: ARCH (SCILHS 3.1b) metadata for PROCEDURES......................................85.2.8 Pcornet_vital.zip: ARCH (SCILHS 3.1b) metadata for VITALS..............................................................85.2.9 KUMC_NACCR.zip:............................................................................................................................... 85.2.10 AdapterMappings.csv:..........................................................................................................................8

5.3 SNOW SHRINE metadata files; DELTA_YYYYMMDD release.....................................................85.3.1 Pcornet_lab_delta_{Oracle/Sqlserver/Postgres}.zip:..............................................................................85.3.2 Pcornet_med_delta_{Oracle/Sqlserver/Postgres}.zip:...........................................................................85.3.3 README_for_{Oracle/Postgres/Sqlserver}.txt.......................................................................................85.3.4 AdapterMappings.csv............................................................................................................................. 8

6 Deployment Instructions.........................................................................................................8

6.1 About I2b2 metadata sets; for the beginner...................................................................................8

6.2 FULL release step-by-step.............................................................................................................96.2.1 Installing i2b2 client metadata................................................................................................................96.2.2 Installing metadata to support SHRINE................................................................................................106.2.3 Java memory parameters for the TomCat server which runs SHRINE................................................11

6.3 DELTA release.............................................................................................................................116.3.1 SNOW_SHRINE_20180330 procedures..............................................................................................11

7 Acknowledgements..............................................................................................................12

7.1 Grant Support...............................................................................................................................12

7.2 SNOMED CT; IHTSDO Observables project...............................................................................12

7.3 College of American Pathologists................................................................................................13

7.4 LOINC..........................................................................................................................................13

7.5 National Library of Medicine.........................................................................................................13


7.6 Grant Funding..............................................................................................................................13

8 Appendix A Medication metadata.......................................................................................14


1 Purpose

This document summarizes the datasets, use cases and deployment of i2b2 metadata to support interoperation of the SNOW SHRINE research data network and in service to the larger community of SHRINE affiliated datamarts.

2 Audience of interest for this document

Interested parties may include: SNOW SHRINE and collaborating SHRINE datamart information managers,

programming staff and administrators I2b2 user community members with interest in interoperation of queries with others

employing Office of the National Coordinator (ONC) for Health IT terminology standards

Research network data managers evaluating resources for use by their collaborators

3 Content scope and history

Dr Isaac Kohane, Dr Shawn Murphy and a team at Harvard University led the project which developed i2b2. Staff at Harvard developed the protocols of the SHRINE network which allows i2b2 datamarts to share queries across research networks. In order for those queries to interoperate, i2b2 concept metadata and concept coding of observation facts must be aligned or extensively mapped at every site supporting a datamart. In view of the fact that a large fraction of i2b2 research data originate in the electronic health record (EHR), the semantic interoperability framework imposed by ONC pragmatically defines a major portion of the terminology standards that could contribute to rapid development of interoperability in any US research data network.

Harvard promoted a metadata set for SHRINE networks which it named in reference to their information management scheme “Scalable Collaborative Infrastructure for a Learning Healthcare System” (SCILHS). In fall of 2017 the Greater Plains Collaborative (GPC) network evaluated SCILHS metadata for support of interoperation within the SNOW SHRINE network it was deploying with hub at the University of Wisconsin, Madison. GPC identified opportunities for enhancing SCILHS 3.1b metadata content for medications, laboratory results and diagnoses(problems/conditions) to assure comprehensive use of ONC data standards. GPC consulted with Harvard in 2017 to propose an expanded metadata set which UNMC would support on a regular maintenance cycle in order to maintain currency with ONC terminology standards evolution and provide versioning and synchronization of SHRINE networks for interoperation.


UNMC began publishing SNOW_SHRINE metadata 20180330 and supports FULL implementations consisting of pipe delimited TXT or comma delimited CSV which can build all version-specific tables. First release is also accompanied by a DELTA dataset which consists of TXT files of SQL commands that may be executed to convert an existing deployment of (previous) version to the new release version. The first DELTA to be published 20180330 consisted of version changes to move from SCILHS(ARCH) 3.1b to SNOW_SHRINE_20180330.

Publication date

Version Content

3/30/2018 20180330 SCILHS 3.1b content AND enhancements which include added content:

a) MEDICATIONS includes historically comprehensive RXNORM (IN/SCD/GPCK/SBD/BPCK) and NDC codes indexed by ingredient

b) LABORATORY with expanded LOINC 2.61 lab metadata covering 2484 tests in lab, anatomic and molecular pathology

12/7/2018 20181001 New content development:a) ENCOUNTER DIAGNOSES: ICD-10-CM updated

with additional 1485 concepts from CMS to 20181001; 30 code revisions this release

b) PROBLEM LIST: SNOMED CT conditions added with US extension 20180930

c) LABORATORY expanded to include 218 more molecular pathology and microbiology

d) MEDICATIONS updated with 1461 new RxNorm and 9781 NDC from NLM release files 20181001

4 Quality assurance, versioning and maintenance

Relevant International and ONC data standards included within each SNOW SHRINE version follows:

SNOW SHRINE version ONC Reference Standard versions20180330 LOINC 2.61; RXNORM 20180130; ICD-10-CM 2017; ICD-9-

CM 20141001; HCPCS(includes CPT4® but discussion ongoing with American Medical Association)

20181001 LOINC 2.61; RXNORM 20181001; NDC 20181001; ICD-10-CM 20181001; ICD-9-CM 20141001


5 Description of Release Datasets and Instructions for use

5.1 Reference copy of SCILHS 3.1b; renamed Accessible Research Commons for Health (ARCH) Model Ontology

May be accessed at https://github.com/ARCH-commons/arch-ontology

5.2 SNOW SHRINE metadata files; FULL_YYYYYMMDD release (versions 20180330 and 20181001)

Contains all metadata tables, adapter mappings and TABLE_ACCESS modifications to install SNOW_SHRINE i2b2 node including the NAACCR data tables developed at KUMC.

5.2.1 Pcornet_Demo_oracle.zip: ARCH (SCILHS 3.1b) metadata for DEMOGRAPHICSMetadata for DEMOGRAPHICS in Oracle and SQL Server formats includes: a)Pcornet_demo_{Oracle/SQLServer_metadata.txt: metadata for DEMOGRAPHICS table, b)Pcornet_demo_Oracle_schemes.txt, c)Pcornet_demo_oracle_table_access.txt: additions to Table_Access for including demographics in deployment scheme

5.2.2 Pcornet_diag.zip: ARCH (SCILHS 3.1b) metadata for DIAGNOSES; modified v20181001 to include 124503 SNOMED CT conditions, 93522 ICD-10-CM codes from 20181001 release; SNOMED CT metadata searching supported via transitive closure tables

5.2.3 Pcornet_enc.zip: ARCH (SCILHS 3.1b) metadata for ENCOUNTERS

5.2.4 Pcornet_enroll.zip: ARCH (SCILHS 3.1b) metadata for ENROLLMENT

5.2.5 Pcornet_lab.zip for LABORATORY and PATHOLOGYExpanded laboratory metadata set which includes 1880 common tests for laboratory, anatomic and molecular pathology. 20181001 release file is I2B2_OBSERVABLES.txt for lab metadata.

5.2.6 Pcornet_med.zip for MEDICATIONSExpanded metadata schema for orderable, dispensed and administered medicinal substances meant to be comprehensive for the history of clinical drug use in the US pharmacopoeia. The schema offers classification by VA drug therapeutic classes as well as alphabetically by ingredient. The classifications are linked to orderable medications and then to NDC package codes at the base level. This release includes 109154 RXCUI( of type = IN,MIN,PIN,SCD,SBD,GPCK or BPCK) and 527401 NDC, each representing a unique combination of manufactured product and package.


5.2.7 Pcornet_proc_nocpt.zip: ARCH (SCILHS 3.1b) metadata for PROCEDURES

5.2.8 Pcornet_vital.zip: ARCH (SCILHS 3.1b) metadata for VITALS

5.2.9 KUMC_NACCR.zip: University of Kansas development of NAACCR tumor registry datasets; includes a) Shrine_ont.naaccr_ontology.zip: metadata files in csv format, b) Adapter_mappings_naaccr_1to1.zip: adapter mappings for NAACCR metadata

5.2.10 AdapterMappings.csv: Adapter mappings for all other metadata types required for deployment of SHRINE network query functionality using SNOW SHRINE metadata, available only for the 20180330 release

5.3 SNOW SHRINE metadata files; DELTA_YYYYMMDD releaseThis first DELTA release consists of SQL executables in TXT files which can update the implementation of the SCILHS 3.1b i2b2 metadata to SNOW_SHRINE_20180330.

5.3.1 Pcornet_lab_delta_{Oracle/Sqlserver/Postgres}.zip: SQL command file which incrementally modifies the LAB table from SCILHS 3.1b to current version SNOW_SHRINE

5.3.2 Pcornet_med_delta_{Oracle/Sqlserver/Postgres}.zip: SQL command file which incrementally modifies the MED table from SCILHS 3.1b to current version SNOW_SHRINE

5.3.3 README_for_{Oracle/Postgres/Sqlserver}.txt Contains a brief procedural description for deploying the delta release in each of three common operating systems

5.3.4 AdapterMappings.csv Contains a copy of full adapter mappings for deployment of SNOW SHRINE

6 Deployment Instructions

6.1 About I2b2 metadata sets; for the beginnerMetadata tables, sometimes referenced by users as ‘ontologies’, support i2b2 aggregate query functionality in the user client and across SHRINE networks. They do this by organizing code references stored in OBSERVATION_FACT.CONCEPT_CD in hierarchical arrays in the metadata tables where the code reference is contained in the attribute C_BASECODE. The hierarchical structure is configured in the attribute C_FULLNAME which specifies all the hierarchical paths which may lead to the concept of interest. Steps in the path which may be queried as ‘parent’ concepts –


supporting aggregation queries - are called ‘folders’ in the client and are the ontological supertypes of all nodes descendant in the ontology.

Metadata are deployed in the i2b2 metadata schema as a set of relational tables, one for each data type stored in i2b2, such as demographics, diagnoses, medications, etc. Each metadata row in the table specifies one path to a particular C_BASECODE concept reference and includes C_METADATAXML which supports query functionality for the concept as well as details of aggregate searching in C_COLUMNNAME, C_OPERATOR and C_DIMCODE.

The metadata deployment for a datatype is enabled when a row is created in TABLE_ACCESS specifying the table name in C_TABLE_NAME and the root of the ontology in C_FULLNAME. Installing an i2b2 metadata schema therefore, consists of creating a metadata table for each datatype to be supported, adding the table name and metadata details to TABLE_ACCESS and then running a utility to create the CONCEPT_DIMENSION table which supports client searching. Metadata tables may specify the use of certain code sets for indexing OBSERVATION_FACT.CONCEPT_CD. The code sets must be identified as coding systems in a SCHEMES metadata table.

Metadata tables may also contain rows which define modifiers for the datatype. Modifiers are an additional i2b2 search tool which support query of subsets of concepts, effectively enforcing a data model in i2b2. Modifier paths are specified in C_BASECODE and are applied to hierarchies in CONCEPT_DIMENSION as specified in the attribute M_APPLIED_PATH. A table row with M_APPLIED_PATH not equal to ‘@’ indicates that that row is a modifier definition. Once the metadata table is installed, modifiers are deployed in the schema by running a utility which creates the table LOAD_MODIFIER_DIM.

Sites which are participating in SHRINE networks have an additional requirement for installing adaptor mappings, which support mapping a network query to the specific paths and structures of the site metadata.

6.2 FULL release step-by-step

6.2.1 Installing i2b2 client metadataExtract the zipped file SNOW_SHRINE_FULL_YYYYMMDD to expose a folder ‘MetadataFiles’ and a zipped file ‘AdapterMappings.zip’. Within the ‘MetadataFiles’ folder you will find zipped sets of TEXT files named pcornet_<type>,datestamp.zip, where type is lab, med, etc. The following table describes the names and purpose of the enclosed files.File Descriptionpcornet_<type>_metadata.txt CSV or ‘|’ delimited file containing the entire

contents for one metadata table. This should be used to replace any existing content in the metadata table.

pcornet_<type>_table_access.txt Text file containing the row entries to update the


TABLE_ACCESS entry for the table pcornet_<type>.

pcornet_<type>_schemes.txt Text file containing entries for the SCHEMES table in the i2b2 ONT cell associated with the pcornet_<type> table.

The following should be done if the pcornet_<type> table is not currently defined in the TABLE_ACCESS table in the i2b2 ONT cell, otherwise omit this step. Insert the entry for the pcornet_<type> table into TABLE_ACCESS by executing the SQL in the pcornet_<type>table_access.txt file.The SNOW_SHRINE_20180330 FULL release replaces the contents of the SHRINE metadata tables in the i2b2 ONT cell. The pcornet_<type>_metadata.txt file contains the full content for the metadata table for that type. It is a pipe-delimited, text-in-quotes file. Replace the contents of the metadata table by truncating the table and loading this file into the table C_TABLE_NAME attribute in the associated TABLE_ACCESS row entry. Do this for demographics, diagnoses, encounters, enrollment, laboratory, medications, procedures and vitals.

The ‘KUMC_NAACCR.zip’ metadata set is tumor registry metadata for SNOW SHRINE datamarts but is free for use by any site wishing to use it. As you create each metadata table, load the associated code system reference data into a SCHEMES table that you create.

Note that empty strings should be imported as NULLs otherwise m_exclusion_cd='' breaks the modifiers!

Your i2b2 system should have utilities: ‘LOAD_CONCEPT_DIM’ and ‘LOAD_MODIFIER_DIM’. These utilities should be run now to create ‘CONCEPT_DIMENSION’ and ‘MODIFIER_DIMENSION’ tables after metadata refresh. They should be rerun everytime any change in made to ontology metadata or modifier metadata respectively.

6.2.2 Installing metadata to support SHRINESHRINE datamarts have a separate scheme for SHRINE functionality. The release includes a compressed file in FULL\MetadataFiles.shrine.zip. The compressed file is ‘shrine.csv’ and should be extracted and installed as a replacement for the corresponding table in the SHRINE scheme.

The compressed file FULL\AdapterMappings.zip contains the file AdapterMappings.csv. Extract the ‘AdapterMappings.csv’ file and use it to replace the existing AdapterMappings.csv file in the tomcat/lib folder on your Shrine Spoke machine. This is the library folder associated with the TomCat web server software which runs the SHRINE software. For example, the path to the file is /opt/shrine/tomcat/lib/AdapterMappings.csv at a known site. After installing the file, the Tomcat server software should be restarted to cause the SHRINE software to load the updated mappings in the AdapterMappings.csv file.


6.2.3 Java memory parameters for the TomCat server which runs SHRINEThe SNOW_SHRINE_20180330 release significantly increases the number of paths defined in the AdapterMappings.csv configuration file used by the SHRINE software. This may increase the amount of memory needed by the TomCat server. We suggest that the server should be allowed at least 2 gigabytes of memory.

Within the bin directory of the TomCat server (e.g. /opt/shrine/tomcat/bin), the file setenv.sh defines Java configuration parameters for the TomCat server, including memory settings. The following is an example (sentenv.sh) script that sets the memory allocation for the TomCat server to 2 gigabytes, it does this through the “-Xms” parameter.

export CATALINA_OPTS=" -XX:MaxPermSize=256m -Dakka.daemonic=on -Xms2g”

6.3 DELTA release

6.3.1 SNOW_SHRINE_20180330 proceduresFiles in the MetadataFiles folder which relate to the incremental updating of SHRINE metadata tables from SCILHS 3.1b to SNOW_SHRINE_20180330 include:

#1. pcornet_lab_delta_oracle.zip -- for updating lab metadata (pcornet_lab table).#2. pcornet_med_delta_oracle.zip -- for updating medications metadata (pcornet_med table).

There are README.txt files in the release folder which have instructions for installing the DELTA in systems with Oracle, SQL Server or Postgres. The command structures are somewhat different for each but the instructions below ONLY RELATE TO ORACLE by way of example.

Example: Update the metadata in the pcornet_lab table using the DELTA dataset(NOTE: An analogous procedure would update the MED metadata, replace "lab" by "med" for this)

Unzip the file pcornet_lab_delta_oracle.zip in the MetadataFiles folder and copy pcornet_lab_delta_oracle.sql file to the machine where Oracle resides. This file contains the SQL commands to modify a 3.1B version of the pcornet_lab metadata table for ARCH to SNOW_SHRINE_20180330 version.

The SQL assumes that the lab metadata table is named pcornet_lab. It is possible that the table at your site has a version number in the name, such as pcornet_lab_2_1_2. If that is the case, then you can still update your metadata table using this SQL by temporarily renaming the table to pcornet_lab, applying the SQL, and then renaming to the original name. The following SQL commands would rename a table named pcornet_lab_2_1_2 to pcornet_lab, and shows analogous SQL would be used to restore the original name after applying the SQL file.

>alter table pcornet_lab_2_1_2 rename to pcornet_lab


At this point, the SQL can be applied to update the pcornet_lab table to the SNOW_SHRINE_20180330 version. After that, the following could be used to restore the original table name:

>alter table pcornet_lab rename to pcornet_lab_2_1_2

The following procedure applies the SQL in pcorent_lab_delta_oracle.sql to the pcorent_lab table:

Start the Oracle sqlplus utility, using the Oracle account associated with SHRINE table maintenance. The following command sequence starts the sqlplus utility on a Linux system running Oracle:

sqlplus "<type Oracle account associated with SHRINE metadata>"Enter password: <type password for account>You are prompted by: SQL>Enter the following sequence of commands: SQL> set define off SQL> set escape off SQL> start pcornet_lab_delta_oracle SQL> commit; SQL> exitExecution may take some time but will update the lab metadata file to SNOW_SHRINE_20180330 metadata.

7 Acknowledgements

7.1 Grant SupportWe gratefully acknowledge support from University of Nebraska Medical Center Departments of Internal Medicine, Pathology and Emergency Medicine; Nebraska Medicine and the Research Information Technology Office (UNMC) in managing this work.

7.2 SNOMED CT; IHTSDO Observables project This material includes SNOMED Clinical Terms® (SNOMED CT®) which is used by permission of the International Health Terminology Standards Development Organisation (IHTSDO). All rights reserved. SNOMED CT®, was originally created by The College of American Pathologists. “SNOMED” and “SNOMED CT” are registered trademarks of the IHTSDO.

We wish to thank Farzaneh Ashrafi, Suzanne Santamaria, Daniel Karlsson and all members of the project team for their assistance and advice in developing this material.


7.3 College of American PathologistsWe wish to thank Raj Dash, Mary Kennedy, Alexis Carter and Ted Carithers as well as members of the International Pathology and Laboratory Medicine (IPaLM) Special Interest group of the IHTSDO for their work and advice in developing this material.

7.4 LOINCThis material contains content from LOINC® (http://loinc.org). The LOINC table, LOINC codes, LOINC panels and form file, LOINC linguistic variants file, LOINC/RSNA Radiology Playbook, and LOINC/IEEE Medical Device Code Mapping Table are copyright © 1995-2016, Regenstrief Institute, Inc. and the Logical Observation Identifiers Names and Codes (LOINC) Committee and is available at no cost under the license athttp://loinc.org/terms-of-use.

We wish to thank Daniel Vreeman, Katie Allen and Swapna Abhyankar of Regenstrief Institute for the hours they have spent creating LOINC term sets and advising us on this development.

7.5 National Library of Medicine

We wish to thank Vivian Auld, Susan Roy and James Case of the NLM for their kind advice and support in this project. We are grateful for support of publication of the Nebraska Lexicon© by reference on the UMLS Terminology Server.

7.6 Grant Funding

NIH 5U01HG009455-02 Deploying ONC National Standards in Support of Metadata for Big Data Research Warehouse ManagementPCORI CDRN-1501-26643 CDRN 04:The Greater Plains CollaborativeNIH 1U54GM115458-0, Subproject ID 5422 Great Plains Idea-Ctr-Biomedical Informatics and Cyberinfrastructure Enhancement Core


8 Appendix A Medication metadata

In the US domain, the medicinal product to encode medication orders and prescriptions are mandated to be recorded with RxNorm medication codes. The manufactured medicinal products that are approved by the FDA for medical treatment are packaged by the pharmaceutical manufacturers and registered with the FDA as NDC codes. There is an NDC code for every dosage form and package preparation. NDC codes are therefore employed by all pharmacies for inventory control and medications dispensed from hospital or community pharmacies under a legal medication order or prescription authorized by a licensed medical professional are billed and reported with the NDC code of the packaged product employed. When medications are administered to patients in clinics or within hospitals as part of their treatment, the packaged product employed in the treatment is recorded in the EHR by the nurse with the NDC code from the pharmacy, although it may also be linked back to the RxNorm code chosen by the clinician to populate the order authorizing the treatment. Within the University of Nebraska Medical Center data warehouse, all medication orders and prescriptions are extracted from the EHR with RxNorm codes (TTY = SCDC, SBDC SCD, GPCK, SBD or BPCK) and all medications dispensed or administered are encoded with NDC codes.

RxNorm codes were developed by the National Library of Medicine in conjunctions with HL7 and the Veterans Administration in order to provide unqie and reproducible naming and encoding for clinical drugs in use in the US. RxNorm is not an ontology of medications but is rather a controlled terminology employing a number of semantic types called TTYs. In the RxNav and RxClass APIs, NLM supports browsing and monthly publication of the full historical RxNorm and NDC code systems which are also summarized and current codes are reported in the UMLS Metathesaurus three times yearly. UMLS reports only codes in active use and so the informatician must process historical releases of UMLS in order to get a comprehensive listing of codes ever used in the US. RxClass API also publishes and supports mapping of RxNorm codes(RXCUIs) to a number of drug classifications/ontologies to support various use cases for drug selection by a data user.

RxNorm semantic types(TTYs) are visible on the RxNav browser website and demonstrated in the screen shot included as figure . The relationships linking RXCUIs of different TTYs into a semantic network for RxNorm are supported in the RxClass API and in the UMLS release for current drugs.

Medication ingredients (TTYs=IN-ingredient, MIN-=multi-ingredient combination, PIN=precise ingredient such as salt of the ingredient drug and BIN=branded ingredient) are the base level and may be employed to identify all RXCUIs for clinical drugs containing one or more active components.

Drug components (TTYs = SCDC semantic clinical drug component and SBDC=semantic branded clinical drug component) are formulations including dose that are employed in manufactured drug products.


Dose form groups (TTYs = SCDG semantic clinical dose form group, DFG=dose form group and SBDG=semantic branded dose form groups) are dose forms for use in aggregate queries of RXCUIs of a specific type of dosing for a clinical drug. TTYs of ingredient classes, drug components, and dose form groups are NOT orderable as clinical drugs in the US EHR because they do not contain all formulation, dose strength and product information required to uniquely inform the pharmacists what the clinician is ordering. Only

Clinical and branded drugs (TTYs of SCD=semantic (generic clinical drug, GPCK=generic clinical drug pack, SBD=semantic branded drug and BPCK=branded drug packs) all represent packages of manufactured products that are orderable because the RXCUIs correspond to FDA approved medicinal ingredients of specified formulation and dosing that can be ordered and dispensed in the US. Figure: RxNav website

UNMC prepares i2b2 medication metadata developed from mining the RESTful services of the RxClass website and refreshes that metadata quarterly. The use cases supported for case retrieval include: 1) all medication events by active ingredient, 2) all orders/prescription events by clinical or branded drug, 3) all dispense or administration events by ingredient, clinical drug or manufactured product, 4) all medication events by VA drug class, 5) all orders/prescription events by VA clinical or branded drug or 6) all dispense or administration events by manufactured products in the VA formulary. For these retrieval use cases, two metadata (ontologies) are supported for medications:

Alphabetic drug ingredient-RXCUI\Clincal or branded drug or pack-RXCUI\Manufactured product-NDCOrVA drug class\ingredient-RXCUI\Clincal or branded drug or pack-RXCUI\Manufactured product-NDC

In the figure below the metadata for Piperacillin (RXCUI=8339/TTY=IN) is pictured. The first clinical drug employing piperacillin (RXCUI=884313\TTY=SBD) was retired by NLM in August 2015. Notice that 15,680 patients had orders\prescriptions for this drug, emphasizing the importance of historically


complete and comprehensive metadata. Likewise the first drug package listed has NDC=00206886201and was retired at that time but there are 14048 patients who had dispense\administration events recorded in the EHR. In total, there are five NDC package codes recorded in the EHR with events corresponding to this clinical drug. If the site extracts administration events encoded with the RXCUI from the order, these cases will still be retrieved by the metadata. Note that retired RXCUI and NDC codes are listed in parentheses in the metadata along with the retirement dates.

Figure: i2b2 metadata for Piperacillin


ihtsdo document · web viewwe wish to thank raj dash, mary kennedy, alexis carter and ted carithers...

Documents