Authors: Mary Nilsson, Aimee Basile and the PHUSE Adverse Event Collection Recommendations Project Team

There have been several publications suggesting broad alignment that the investigator assessment of causality for non-serious adverse events lacks value (2010 FDA Clinical Reviewer Template (1); 2005 CIOMS VI (2); PHUSE Adverse Event white paper (3)). Its collection is not required by regulations, and not required per the CDISC SDTM Implementation Guide (4). We agree with this sentiment; yet, its collection is predominant in practice (5). What does it take to change predominant practice?

As discussed in the FDA Clinical Review Template (1), the relatedness assessment made by the investigator generally is not considered useful as it tends to reflect what is known about the study drug at the time of the clinical trial. In the CIOMS VI report (2), there is a recommendation only to collect relatedness for serious events. The authors of the PHUSE Adverse Event white paper (3) concur with the discussion provided in the CIOMS VI report but recommend its collection to enable the ability to include both percentages of treatment-emergent events and events considered related by the investigator, which is typical in Japan labels. Discussion would be needed with PMDA as to whether it would be acceptable to include only percentages of treatment-emergent events. Recently, adding to the challenge, authors of a guidance document for Plain Language Summaries (6) have proposed using investigator relatedness data for summary tables. Plain language summaries will be required per EU regulations for all clinical trials starting January 2022, and are intended to summarise clinical trial results for the layperson audience.

Is there sufficient value to justify such broad collection? Is there value in early phase, but not in later phases? Is there an opportunity for simplification? Thoughtful discussion of the idea to only collect relatedness for serious events seems warranted. Who are all the stakeholders for these conversations? Having multiple publications with this idea hasn’t led to any action or change in practice. How can we move from an idea to action on this topic? Please share ideas through workinggroups@phuse.global.

References

(1) US Food and Drug Administration. (2010). Good review practice: Clinical review template. Manual of policies and procedures [serial online]. Retrieved 31 August 2021, from https://www.fda.gov/media/72472/download.

(2) CIOMS Working Group VI. (2005). Management of safety information from clinical trials. Geneva: Council for International Organizations of Medical Sciences.

(3) PHUSE Analysis & Display White Papers Project Team. (2017). Analysis and Displays Associated with Adverse Events: Focus on Adverse Events in Phase 2-4 Clinical Trials and Integrated Summary Documents (Doc ID: WP-003) Version 1.0.

(4) CDISC Submission Data Standards Team. (2021). Study Data Tabulation Model Implementation Guide: Human Clinical Trials Version 3.4.

(5) PHUSE Safety Analytics Working Group. (2020). Adverse Event Collection and Treatment Emergence (Doc ID: WP-052) Version 1.0. Retrieved 31 August 2021, from https://phuse.s3.eu-central-1.amazonaws.com/Deliverables/Safety+Analytics/Adverse+Event+Collection+and+Treatment+Emergence.pdf.

(6) Clinical Trials Expert Group. (2021, October). 2021 Good Lay Summary Practice. Retrieved 12 October, 2021, from https://ec.europa.eu/health/sites/default/files/files/eudralex/vol-10/glsp_en.pdf.