Although EMA Policy 0070 states that PPD redactions are for the MAH/applicant to decide, EMA will comment and may even strongly recommend not redacting certain information if they do not agree. Ensure the justification of your redactions is clearly presented in the anonymisation report (AR).
In the case of CSR Section 14.3.4 (per ICH E3 “Abnormal Laboratory Value Listing [Each Patient])” – the policy and guidance at the time implied these were out of scope. The current implementation guidance V1.2 dated 12 April 2017 states that these are indeed out of scope of phase 1 of the policy, so they can be removed. However, we were told in late 2016 by the assigned EMA reviewers that these needed to be submitted, and so part-redaction of the data was agreed.
In my experience, we found that if reviewers felt that data utility was being compromised then they would try and push the MAH/applicant to re-think the PPD redactions. This would apply if the study was in a rare disease area with very small patient numbers. In such studies, you could potentially redact all the data/data subsets as the patient numbers are so low (sometimes as low as 2 to 3 patients). I would recommend establishing good working relationships with the EMA team at the start, to allow open discussions that help reach a balanced agreement in terms of data utility and patient confidentiality.
There are some inconsistencies between what EMA has allowed across different companies and these are visible in the published ARs. For example, the TransCelerate document recommends redaction of all vendor addresses, however when we submitted our proposal we were told that CRO addresses could not be redacted. To allow things to move forward, the Sponsor agreed with EMA’s request. Later the Sponsor discovered that another company was allowed to redact vendor addresses (detailed in the published AR). Therefore, it seems what is finally approved depends on which reviewers have assessed your AR.
I have also reviewed a few of the published ARs and noticed that in some cases although justifications for certain redactions have been documented, in the accompanying CSR the redaction is not applied (e.g. a threshold of 25 patients; redaction of site addresses). Does anyone have any thoughts on this? In relation to this point, I have a question for users of www.core-reference.org, which I have directed to EMA, and am awaiting their response. My question is: “If a Sponsor submits a variation to the current marketing licence, for example to add clinical data for the same product, should they revise the already published AR or submit a new AR for the new data?” If anyone has insight to share on this, please do so via: http://www.core-reference.org/disclosure-feedback/.Rukhsana Shaikh-Zaidi
Director, RSZ Scientific Limited