Why is data anonymization all of a sudden a buzz term in the clinical trials sector?
Sponsors today need to be able to publish their clinical trials data easily and quickly while complying with data privacy and transparency regulations. Both the European Union and the United States have taken steps to make clinical trial data more transparent in order to enhance research and increase patient safety. However, at the same time, regulations are now looking at protecting the personal data of citizens, which of course means patient identifiers as well.
The FDA is establishing guidelines for transparency through the use of de-identified and masked data. The hurdle is protecting data privacy of patients while protecting the commercial investments of Sponsors.
The European Union has been talking about data transparency for a few years now. With the approval of Clinical Trial Regulation 536/2014, which comes into effect in 2019, there will be an EU portal for publishing data once a Marketing Authorization Application is submitted regardless of the final decision. While many companies are supporting the idea of more data transparency, Sponsors are also struggling with how they are going to remove patient identifiers to make data acceptable and compliant. There is still a lack of experience in the industry regarding data anonymization.
Meanwhile the European Union has taken data protection to a new level. In an effort to protect citizen data and information, the EU passed the General Data Protection Regulation (GDPR) which is coming into effect in May 2018.
Of course, health data falls into this category. Clinical trial data is considered sensitive personal data, and therefore GDPR will call for tighter conditions for processing data. In terms of clinical trials, Sponsors will have to look at the process in which they carry out patient consent but also carry out a data protection impact assessment.
With GDPR, both Sponsors and vendors (i.e. CROs) are responsible for compliance when it comes to data handling.
Do you have a data anonymization process?
Sponsors should begin by verifying that their datasets are CDISC compliant. Sponsors should evaluate with their vendor if the biometrics team has a workflow and quality control check process in place for data de-identification and anonymization.
De-identification involves removing or recoding health information that could identify an individual such as patient identifiers, free text verbatim terms or references to dates. Subsequently, data anonymization involves destroying all links between the de-identified datasets and the original datasets.
A typical data anonymization workflow includes:
- SDTM/ADaM input
- Metadata and Functional specifications
- Data anonymization macro
- Anonymized SDTM/ADaM data
- Anonymized Procedure document
There are three levels of patient data:
- Level 1: direct identifiers such as biometrics information
- Level 2: indirect identifier such as date of birth or body measurements
- Level 3: Risk of data linking by combining more than one data point
CROS NT offers support for data anonymization compliance starting with patient-level data with a comprehensive process and expert team of statisticians, programmers, medical writers and regulatory professionals.
- Up to Level 3 data privacy protection
- Pre-validated macros less prone to errors
- Data format specifications required for EudraCT portal data upload
- Expert consultancy for data transparency strategy
- CDISC support as Gold Member
- ISO 27001 (Information Security Management System) certified for IT systems and processes