Page tree
Skip to end of metadata
Go to start of metadata

Short Description

The track will focus on extracting patient medication data from an EHR FHIR server and generating an SDTM dataset to submit to FDA's submission Gateway for validation. This is an evolution from the January track and new steps will cover the assessment of the confidence in the quality of the data and addition of a FHIR based algorithmic approach to consolidation of medication data for a patient. 

Long Description

Clinical Research studies currently require the redundant entry of clinical data that already typically reside in certified EHR/health IT systems.  It would be substantially better to utilize EHR source data to directly populate clinical research data capture systems wherever feasible. At this time, the data standard used for submission of tabulated clinical study data for pharmaceutical and biologicals to the FDA is SDTM (Study Data Tabulation Model), created by the Clinical Data Interchange Standards Consortium (CDISC) standards development organization, which is optimized for clinical research and regulatory uses. 

This track is the third in a series evolving an approach to collect concomitant medication data for a patient from one or more EHR systems and to aggregate this into a summary record that meets regulatory standards for accuracy and trustworthiness and which can then be converted into a format for submission.  

The January 2021 considered a series of fictional patients, and retrieved retrieve medications relevant dates from the following resources: MedicationStatment;  MedicationRequest.   

In May 2021, we expanded to allow retrieval from additional resources: MedicationDispense and MedicationAdministration and began to explore how best to represent a complete medical record set utilizing something more efficient than a FHIR bundle.  Additionally explored how to represent additional supplemental Medication information in SDTM where such information is not normally stored in the core CM domain.

For September we aim to expand on this to allow interrogation of source data with CapabilityStatement which should help all downstream decision-making for a reference implementation, improve data retrieval and consolidation of cross-resource medication information into a refined FHIR profile, achieve a level of deduplication, and improve the ultimate SDTM output to fully utilize the SUPPxx domain methodology.


Test the design of a set of resources, Test an Implementation Guide

Submitting Work Group/Project/Accelerator/Affiliate/Implementer Group  

Vulcan Accelerator

Track Lead(s)

Scott Gordon, Lauren McCabe

Track Lead Email(s)

Related Tracks

FHIR Version

FHIR R4 is current primary focus, also US Core.  However, since many other implementations exist, we are not limiting this exclusively.  Further, as R5 moves forward in balloting, we will keep an eye on relevant changes.

Specification(s) this track uses

Aim to have a 1st Draft Implementation Guide for this Connectathon

Artifacts of focus 

Expected participants

Vulcan members, Regulators, EHR Vendors, EDC Vendors, Academic Medical Centers, Pharma Companies.  We averaged 10-15 participants in January and May - should be similar.

We particularly invite EHR vendors to provide test systems to work with and Developers to address the medication record consolidation part of the system.

Zulip stream


Kick-off Call

Tuesday September 7, 2021, 11:00 AM - 12:00 PM Eastern Time

Whova Agenda and Webinar Link:

Track Details

System Roles:

  • FHIR EHR Server - An EHR server pre-populated with synthetic or fully anonymized records. The connectathon will not make any changes to the data on this server.
  • Consolidation engine
  • FHIR-to-SDTM Mapping Capability - This capability will be developed for the connectathon. It will convert medication-related FHIR resources for a given patient to CDISC SDTM Concomitant Medication dataset, i.e. CM domain dataset. This capability will also convert the Patient FHIR resource for the same patient to CDISC SDTM Demographics dataset, i.e. DM domain dataset. For the purposes of the January connectathon other necessary CDISC SDTM domain datasets, e.g. TS/ TE/ TA, will be created manually, especially since the EHR Server is less likely to contain clinical trials related data.
  • Validation of SDTM messages - The FDA Gateway will receive the CDISC SDTM datasets and validates it. Note: Pinnacle 21 will serve to validate CDISC SDTM datasets. 


The September scenario for a patient in the EHR FHIR server by a patient identifier

  1. Query EHR/Server with CapabilityStatement (focus on MedicationXxxxx Resources)
  2. Retrieve patient’s medication records for a given period - including all relevant attributes to relay as much information about the record
    1. ie: It is a prescription, administration, patient reported only, etc.
  3. Display medication records on the UI
  4. Consolidate the records into single document (potentially utilizing MedicationStatement profile as destination)
  5. Perform any relevant deduplication (if achievable with given data set)
  6. Show the FHIR XML/JSON representation of the medication records, consolidated record, and deduplication set
  7. Convert the final FHIR data set to CDISC SDTM datasets, CM and DM domains
    1. including SUPPCM data file to contain information not typically captured in the SDTM CM domain
  8. Validate the created SDTM datasets 

Test Script(s):

The wireframe below shows the January/May reference implementation.  For September it may be similar

  1. The user enters a patient number in block A and clicks [Search]
  2. The system will retrieve the patient from the FHIR EHR server and display basic patient demographics information in block B, and the patient’s medications record in block C.
  3. The user can click [Display Results (XML/JSON)] to display the XML or JSON FHIR representation of the medications record
  4. The user can click [Export SDTM Dataset] and the system creates the DM (Demographics) and CM (Concomitant Medication) CDISC SDTM domain files to store on the user’s local machine.

Note: Other CDISC SDTM files necessary for the submission will be manually created, i.e. TS (Trial Summary), TE (Trial Elements), TA (Trial Arms), and define.xml. Since the server will have EHR clinical data, and no clinical trial data, the clinical trial related variable values used in these SDTM domain files will be synthetic and constant for any patient used in the prototype.

Security and Privacy Considerations:

The data used for this connectathon will be synthetic or fully anonymized.  There may be security requirements for access tot the FDA gateway for testing.


Files and links during Connectathon

Kick-off session Slides (September 7, 2021): 

Concomitant query and bundle creation Module : 

Demo - Vulcan Medication Bundle (

Transformation Module (transform FHIR bundle to SDTM)

Transform Bundle - FHIR 2 SDTM (

Day 1 (Sept 14) 8am Slides:

FHIR links for medication status filter activity: remove medication if the status tells us it was not or will not be taken.

Draft of decision algorithms based of Medication[x] Status field


Draft identifying all elements of all Medication[x] resources which we recommend be implemented and utilized consistently (where relevant) in all EHRs in order to make concomitant medication determination more reliable.  (There are absolutely strong pure healthcare justifications as well for these elements being present).

MedicationX Resources and ConMed.docx

1 Comment

  1. Useful links for next connectathon:


    LZZT Study protocol

    This is the LZZT protocol, Lilly donated it years ago.  There is a database of the SDTM on GitHub - Transcelerate - not sure who (?); CDISC has it on the website years ago but it is no longer there.


    Brought to you by: xml4pharma

    Software workup created by Jozef Aerts, covers many aspects of EHR -→ SDTM

    The semantic mappings and the API for it can be found at:

    Demonstration videos


    presentation part 1:
    presentation part 2 (containing the actual demo):

    CDISC Interchange 2020: generating COVID-19 SDTM datasets from EHRs: (also includes a demo).