One of the most time and resource intensive tasks in clinical research is data entry.
Clinical research coordinators spend countless hours transferring clinical data from a patient’s Electronic Health Record (EHR) into a study specific Electronic Data Capture (EDC) system.
This is annoying because the research coordinator could instead spend those hours on recruiting patients.
A direct EHR to EDC “connection” could easily solve this issue.
In fact such a solution does exist and the brain child behind this endeavor is our guest today, Dr. Hugh Levaux, founder and CEO of Protocol First.
Hugh is a clinical research expert and digital transformation leader. He is also a data nerd (in a good way), which I really appreciate and like about him.
Hugh has over 20 years of dedicated experience in clinical trial design and operations, as well as web-based technologies.
Most notably, Hugh has held the roles of SVP at United BioSource Corporation (UBC), VP of Product Strategy at Medidata Solutions, CEO of Ninaza, a web-based EDC provider, and SVP at Quintiles (currently known as IQVIA). Hugh has developed and commercialized multiple eClinical technologies in EDC, IVR/RTSM, ePRO/eCOA and trial-related technologies.
Prior to his career in the clinical trials industry, he worked as an analyst at the RAND Corporation, where he worked for US government clients (US Air Force and intelligence community) and private sector clients (oil and aerospace). Before RAND, he worked in Japan at Panasonic.
Hugh holds a Ph.D. in Policy Analysis from the RAND Graduate School and has also received master’s degrees in international economics and international relations from SAIS at Johns Hopkins University and international politics from Université Libre de Bruxelles in Belgium. He is fluent in English, French, Spanish and Japanese.
Please join me in welcoming Hugh to the Clinical Trial Podcast.
Florence Healthcare: Transform your Clinical Trials with the #1 Platform for Electronic Document Workflows and Remote Site Access.
Learn how Florence Healthcare helps over 10,000 research sites, sponsors, and CROs across 44 countries take their day back from paper
[4:38] Electronic Health Record (EHR) systems at large academic centers and small community sites also serve as the source data for clinical trials
[6:20] What does the future of Electronic Data Capture (EDC) look like if researchers have direct access to source data from the EHR?
- Single study level builds will likely disappear
- Replacement of EDC system with “data lakes”
- Need to establish mechanism for communication with sites about missing data, exchange queries, and address data entry errors, if were to have an EDC-less world
[8:03] Limitations of relying solely on EHR data for clinical trials
- Technology does not allow us to have 100% of the data in the EHR. There is data created outside the EHR system the cannot be pulled in
- Data exist in the EHR but cannot be easily retrieved
[9:23] Organizational stakeholders involved in EHR-EDC integrations
- Principal Investigator (PI)
- Research Coordinator (RC)
- Data Manager
- Senior Leadership
[10:09] EHR-EDC integration questions from PI and RC
- How fast can we get the system implemented at our site?
- We do not have everything in our EHR (ex: referral letters). How is this handled?
- What type of data is being transferred?
- What support is available?
[11:55] Cost of implementation to a research institution (resources)
- Security review
- IT support
Sponsor pays for the service and site user can use it for free
[12:43] Regarding Protocol First (Remote Monitoring, EDC) and Clinical Pipe (EHR to EDC connector)
[14:14] Standard to connect to EHR systems is Fast Healthcare Interoperability Resources (FHIR)
[18:34] Selecting EHR systems
- Connectivity is critical (FHIR API)
- APIs the EHR system supports
[19:38] EHR-EDC integration questions from Sponsors
- Global presence
- What EHR systems that they connect to? What percent of the market do you cover?
- When, where, and how will you be able to connect to the remaining EHR systems?
- What data can be transferred – data types, for what reason
- What data we cannot transfer – data types, for what reason
[22:28] Data Types
- Clinical lab data
- Structured data elements – quantitative value entered in a box or field
- Logs to collect concomitant medication and adverse events
- Unstructured data elements – progress notes, narratives, impressions from physicians and nurses
[25:07] Unstructured data is exploratory for sponsors to use – the next frontier
- FDA has been supportive of leveraging EHR data. FDA guidance document focuses on structured data elements – source data and does not need to be source data verified (SDV’d)
- For many of the notes in EHR are actually XML files and machine readable i.e. can be fed to a Natural Language Processing (NLP) machine that will recognize the words. This data can be exported straight into the EDC
- For the PDF documents that are uploaded in EHR, Optical Character Recognition (OCR) technology can be used to break down documents into every letter and every word. NLP can recognize some key terms and medical events, present them to the user (research coordinator), and then submit it to EDC
[29:22] What does Sponsor leadership care about the most?
- Speed, scalability, and product roadmap are some of the most important considerations
- People do really care about the speed at which we can deliver medicine to the patients
- There is a strong impetus to accelerate timelines and get new medicines to patients sooner than later
- How fast can we extend implement the EHR-EDC integration and speed at which we can go global
[31:42] Does EHR-EDC system integration require IRB approval?
- There is control on the data that is transferred
- There is much higher certainty of protection of the patient’s privacy through the technology
[33:25] Companies that do not want to adopt EHR-EDC integration
- Large sponsor organizations and CROs are inundated with initiatives and projects such as decentralized clinical trials, digitizing the clinical trial space
- EDC system that is not compatible
- Medidata RAVE is configured in such a way that prevents data transfer
- Sponsors may have established processes about data cleaning, and data loading and data coding, etc. They try to bypass the EDC and obtain the data from the EHR system and load it in the backend of the EDC system. It is challenging to prove the chain of custody of data and source provenance and source attributability is difficult. It is also difficult to resolve queries and communicate with the sites
[27:22] Most difficult part of EHR-EDC data transfer
- Site and Sponsor Education
- Ensuring sponsors and sites appreciate the nuances (ex: hybrid studies where some sites are using the technology and others are doing manual data entry)
- Technical aspect is relatively straightforward but not trivial
- Computer system validation – frequent product enhancements and ensure sites are in a validated state
[40:50] Protocol First provides product validation, release notes, updated validation documents etc. with Sponsors. Initially, there is a vendor selection process and regulatory audit
[42:17] Sites and sponsors need to stop thinking about the world they used to
- Manual data transcription
- Think about how the source data actually generated
[44:32] Local labs and lab ranges – past versus present
- Think about how data is changing at point of care
[46:52] Study Data Tabulation Model (STDM) standard allows standardization the data
- Many EDC systems do not natively support the standard
- STDM allows us to define and look at data in a consistent fashion
- Take an existing annotated case report form and perform after-the-fact mapping to SDTM
- Value of reusing edit checks
[52:08] Why are there so many EDC systems?
- Redcap lacks Part 11 compliance. Difficult to track data changes or version updates which is very common in industry studies and far less in academic research
- EDC systems for certain therapeutic areas
- There may be 50 companies with EDC systems
- The same EDC system at Sponsor A may be different than the one at Sponsor B or Sponsor C. There are as many EDC systems as the number of Sponsor companies using them
[56:21] How did Hugh become interested in EDC and EHR technology?
[58:36] Electronic Patient Reported Outcomes (ePRO) and Electronic Clinical Outcomes Assessment (eCOA)
- ePRO – populated and entered by the patient
- eCOA – Clinicians answer a question for the patient on behalf of the patient
[1:03:31] CDISC SDTM standard, HL7 standard and Transcelerate Biopharma
Athena Cloud EHR
Redcap EDC system, a dominant player in academic research