Preparing for EMA policy 0070
How to strike the right balance between protecting clinical data privacy and transparency
Introduction
In September 2023 the European Medicines Agency (EMA) relaunched its Policy 0070 regarding the publication of clinical data, for all new active substances and COVID products. This world-leading initiative aimed at increasing transparency and improving access to scientific evidence is once again expanding its remit in 2025. Pharmaceutical companies will be required to publish detailed clinical trial data for all new marketing authorization applications (MAAs) including negative opinions, withdrawn applications, line extensions and major clinical Type II variations (extension of indications). These changes are designed to enhance public trust, improve healthcare decision-making, and foster innovation by making clinical trial results more accessible.
At first glance protecting both patient privacy and sponsors’ intellectual property while maintaining data transparency may seem to be conflicting objectives. However, it is possible to protect sensitive data while sharing it in ways that assist further research. Doing so requires careful planning, standardised processes, and clear roles. Data utility is the extent to which researchers can analyse and make meaningful interpretations from information. In this blog, we briefly outline data anonymisation methods that ensure adherence to new changes to the EMA Policy 0070 and other regulatory requirements around the world.
Common data anonymisation measures
Two of the most commonly used methods of data anonymisation are redaction and transformation. Each has its own advantages and limitations which we outline below, along with considerations for sponsors when deciding which method to adopt.
Redaction
Redaction involves concealing information by placing a box over direct or indirect identifiers. Most common softwares can do this, either manually or semiautomatically. It may be perceived as a good, cost-effective option. For short documents with little personal information (PI) or Protected Personal Data (PPD) it may be the logical choice. However, because all of the information is obscured it has little to no data utility. In addition, deciding what to redact can be subjective and as a data anonymisation tool it is a blunt instrument.
Transformation
Transformation is the process of generalising or pseudonymising, participants’ direct or indirect information. Participant information such as full name, subject numbers, phone number, email address or a government ID number are direct identifiers. Indirect identifiers include city, state, demographics and sensitive information. In isolation indirect identifiers may not result in reidentification but combined with other information they could be used to identify an individual. In transformation, instead of revealing a participant’s age, the participants’ ages would be banded into groups. Validated software tools can assess the risk of reidentification and put in place transformation strategies and anonymisation techniques. Sponsors can save time by automatically anonymising multiple clinical datasets and documents in this way. When combined with robust quality control steps, this quantitative approach can ensure confidence in the outputs.
Reidentification risk
There is a significant risk of patient reidentification if data are insufficiently anonymised. In a 2006 court case between the Southern Illinoisan and The Department of Public Health an expert witness reidentified 18 out of 20 individuals in a data set from the Illinois cancer registry.1 For the remaining two patients the witness was able to suggest one of two alternative names. There have been other high profile reidentification cases. A Canadian national broadcaster aired a report on the death of a 26 year old student who took a particular drug. The student was reidentified using information from the adverse drugs reaction database released by Health Canada.2 Reidentifications like these can result in lawsuits and pay-outs as well as causing considerable distress for the individuals reidentified, or their grieving families.
Strategies to minimise risk and maximise research success
Additional resources may be necessary to address regulatory requirements and information-sharing requests. Putting appropriate strategies in place can ensure that resources are used most effectively and that already busy teams are not overstretched.
1. Plan in advance
Data anonymisation and transparency requires careful, proactive planning just as other regulatory requirements do. Develop a plan in advance to deal with information requests and disclosure activities from the start of the study. Regulators seek justification for both PPD and CCI anonymising, so be advance planning will save time when writing reports later on.
2. Standardise processes
Agreeing on the anonymisation method to be used and using validated software tools can ensure that processes are standardised. This can remove some of the subjectivity when deciding what information to disclose or withhold.
3. Define roles
A small team composed of well-informed cross-functional people is vital for data sharing initiatives. Team members’ roles should be defined and responsibilities clarified from the outset. It is possible to outsource many tasks related to document preparation and data anonymisation to experienced personnel. This can ease the burden on stretched clinical teams.
The challenges of voluntary data sharing
In recent years several voluntary data sharing programs have been established and many sponsors already share data beyond the regulatory requirements. The aims of these programs differ. Some are motivated by a desire to improve patients’ access to information while others are focussed on benefitting future research. Due to the varied objectives, decisions about which information to share is subjective and may be inconsistent. Different sponsors and stakeholders may have different views on what gets shared, with whom, and for what purpose. Voluntary data sharing also requires a change in corporate planning, resourcing and processes to deal with information requests.
Regulatory requirements
Regulatory authorities around the world have different requirements which can be as detailed as specifying which colour to use to overlay PPD, confidential business information (CBI) or company confidential information (CCI). The Food and Drug Administration (FDA) and National Institute of Health (NIH) allow information to be redacted that the sponsor deems necessary to safeguard PPD and CCI. In Japan and Europe regulatory bodies require justifications for redactions. The European Medicines Agency allows redactions of CCI on an extremely limited basis. Each proposed redaction must have a written justification which may be rejected. During the first year of the implementation of Policy 0070, of 1.3 million pages submitted just 134 pages, or 0.01% of the total pages published were accepted. Similarly, Japan’s Pharmaceuticals and Medical Devices Agency (PMDA) requires justification for CCI and PPD redactions which may be rejected if the justification is deemed unacceptable. Health Canada’s Public Release of Clinical Information (PRCI) policy provides public access to clinical information to support independent analysis of data and supports new scientific research directions. The recent joint initiative between the EMA and Heath Canada will allow for symbiosis in timelines, anonymisation report and in-scope document lists. This collaboration effort will bring benefits to applicants and researchers alike as having one redaction proposal package prepared and submitted to a lead agency and one outcome applicable for both jurisdictions leads to efficiencies for the MAH and faster access to data for researchers.
Conclusion
For pharmaceutical companies, staying informed about updates to EMA policy 0070 and other global data sharing initiatives is crucial. Non-compliance could result in reputational damage, regulatory delays, or restrictions on product approvals. There is a growing emphasis on transparency in the industry, and companies must ensure their processes and data management systems are aligned with the upcoming requirements. Proactive planning from the beginning is vital to manage disclosure activities. Regulators demand justification for anonymising PPD and CCI, so preparation in advance can save time when writing reports. Validated software tools and experienced personnel can ensure sponsors avoid accidental disclosures while contributing to advancing research.
1 Southern Illinoisan v Illinois Department of Public Health, (Supreme Court of Illinois 2006).
2 El Emam K, Dankar FK. Protecting Privacy Using k-Anonymity. Journal of the American Medical Informatics Association. 2008;15(5):627-637.