Adverse Events – PHUSE COVID-19 Guidance for Data Scientists
As the COVID-19 pandemic severely impacts many facets of human activity around the world, the pharmaceutical industry is being presented with significant challenges related to clinical trial research and development activities. Regulatory authorities have released guidance documents focusing on the impacts to study start-up activities, changes to ongoing study procedures, and items considered urgent safety matters during this pandemic.
This blog considers the impacts of COVID-19 on the collection and analysis of adverse events, offering guidance from industry experts on what clinical data scientists  (i.e. those who analyse data collected in clinical trials) can expect in the short term relating to the studies impacted by the COVID-19 pandemic, as well as consideration of future implications.
In the current, constantly evolving atmosphere sponsors have been faced with urgent decisions regarding each clinical study in their portfolios, regardless of development stage. Challenging decisions based on mitigating the risks to study participants are being made on whether to delay initiation of new clinical studies, temporarily halt or terminate ongoing clinical studies, or update study protocols as necessary. These decisions must be based on the current regulatory guidance issued relevant to the COVID‑19 pandemic, including those published by the US Food and Drug Administration (FDA) , the European Medicines Agency (EMA)  and many other countries (including individual member states within the EU). Given the scope and breadth of the guidance documents, as well as the multiple changes required to be considered for any given study, it is essential to collaborate with all relevant team members to ensure all adverse event updates are managed appropriately. It will also be beneficial to discuss proposed approaches with health authorities and individual review divisions as part of taking decisions for individual scenarios. Involvement in a COVID-19 task force will be helpful when communicating these types of challenges and developing decisions tailored to each study.
The statements “Ensuring the safety of trial participant is paramount”  and “The ability to confirm eligibility and to conduct key safety assessments and trial evaluations is of particular importance”  both highlight the importance of safety monitoring of ongoing studies during the COVID-19 pandemic. How sponsors achieve this will differ, and different trials may require different approaches depending on the indication or the mode of action of the investigational or comparator drug(s). A number of considerations are raised in both the FDA guidance  and the EMA guidance  relating to the conduct of the study. These include, but are not limited to, the conversion of physical visits into phone or video visits or utilising alternative locations for assessments, changes to the study visit schedule, pausing trial recruitment or withdrawing subjects from trials. Any steps taken by a clinical study sponsor relating to the operational aspects of a clinical trial in order to guarantee the safety and wellbeing of patients may have an impact on the safety analysis of the study. Clinical data scientists will need to consider the actions taken by the sponsor when preparing to perform these analyses.
It is best practice to focus on the statistical analysis plan when designing clinical studies and in defining how and what data is collected. This ensures the necessary data is collected and helps to mitigate the risk of excess data collection and the need to update data collection design during study conduct to address analysis needs . While most studies impacted by the COVID-19 pandemic will be in the study conduct phase, the same principles should be applied to any updates made to the data collection design. It will be important that any updates made to the operational aspects of the study and to the data collection design are properly reflected in protocol amendments.
All COVID-19 infections and positive tests should be collected as adverse events in a study. It is also important to note that when collecting adverse events that there has been, in many jurisdictions, limited access to testing. This may make the identification of confirmed cases of COVID-19 infection challenging. Symptoms that could appear to be COVID-19 infection related do not always follow with a positive test. Sponsors will need to exercise caution to ensure that all COVID-19 related adverse events are correctly identified.
On 19 April, MedDRA MSSO, with the approval of the MedDRA Management Committee, released an updated version of MedDRA 23.0 with new COVID-19 terms and revisions included to allow organisations to capture, share and analyse scientific and medical information appropriately. Sponsors were urged to implement this re-released version by 4 May to facilitate this . Consequently, clinical data scientists should ensure, as per normal practice, that all adverse events are coded to the latest MedDRA version. The new list of MedDRA terms contains values for positive and negative COVID-19 tests and for asymptomatic COVID-19. Where tests are performed, it is important for sponsors to consider where this data is captured.
Where sponsors do not recode to the latest version of MedDRA, it will be necessary to identify COVID-19 related adverse events. This may be requiring fixed text to be included in the verbatim, for example a prefix of COVID-19 or by updating the data collection to add an indicator that the event is COVID-19 related. Where sponsors do choose to update the CRF, CDISC has provided guidance on where to include this data in SDTM as part of the Guidance of Ongoing Studies Disrupted by COVID-19 Pandemic , in this case the non-standard variable AEEPRELI (Epi/Pandemic Related Indicator) has been proposed.
Given that the list of COVID-related terms added to MedDRA v23.0  provides a search criteria for COVID-19, many sponsors will choose not to update the data collection for adverse events to flag COVID-19 related adverse events as these terms can be identified as part of data analysis activities using customised queries. By using search criteria to identify COVID-19 related terms, sponsors may avoid the need to update implementation of EDC systems for studies, which is particularly relevant where multiple concurrent studies are being run. An additional benefit of this approach is that search criteria will likely evolve as scientific understanding of the virus develops.
Due to the pandemic, both the FDA  and EMA  have highlighted that different approaches may be needed to monitor both the completeness and quality of data to ensure patient safety. This may include the “optimizing use of central and remote monitoring programs to maintain oversight of clinical sites” . Much of the impact of COVID-19 on a clinical study will be identified by protocol deviations and protocol amendments. Missing data and protocol deviations will be discussed in separate blogs but both of these will potentially be important to establish if the impact on adverse events occurring during a clinical study is balanced across treatment arms.
The basis for adverse event analysis is described in the PHUSE White Paper “Analysis and Displays Associated with Adverse Events: Focus on Adverse Events in Phase 2-4 Clinical Trials and Integrated Summary Documents” . As noted in Nilsson et al. , if the impact of COVID-19 affects treatment arms equally, analyses of adverse events from controlled data can largely remain unchanged. Where an imbalance occurs, different analytical approaches may be needed.
For serious adverse events of COVID-19, these should be reported as such and narratives should be included in the CSR as per usual practice. Sponsors’ existing processes will generally be sufficient to gather additional details (e.g. treatments, procedures and tests) relating to the adverse event.
It is important to understand the prevalence of COVID-19 related adverse events in a clinical study and to investigate whether there is an imbalance in the distribution of these events between treatment arms. To enable this, study teams will need to review data for COVID-19 related adverse events. Depending on the number of events, this may be achievable by reviewing existing tables and listings for adverse events, serious adverse events and deaths or by generating new outputs for COVID-19 related adverse events.
COVID-19 related adverse events are highlighted in the list of COVID-19 updates to MedDRA v23.0 developed by MSSO . Some sponsors may choose to create a customised query for these terms based on this list to simplify the identification of such events. To support COVID-19 analyses in the future, it would be beneficial for a MedDRA SMQ to be developed to consistently identify these events across sponsors. Alternatively, some sponsors will have updated their data collection to allow investigators to flag COVID-19 related adverse events; in this case, additional tables, listings and figures will be required to report this subset of events. Where verbatim terms have not been coded to MedDRA v23.0 some sponsors will have provided instructions on how to report COVID-19 related adverse events by use of key phrases in the event description. This would require manual review and/or programming logic across free text fields, which will create additional challenges due to inconsistencies in data entry.
Depending on the number of cases the investigation into potential treatment arm imbalances could be achieved with interactive data review, listings or, if sufficient numbers, summary tables. As not all search terms are explicitly related to COVID-19 (e.g. “Quarantine” or “Patient Isolation”), some sponsors may choose to only include events that occur after the known start of the pandemic in their review or customised query. When submitting analyses based on customised queries it is critical that the list of terms used is submitted to the health authorities. Where data is submitted, it is beneficial to submit this as a data file and to document this in the Analysis Data Reviewer’s Guide (ADRG)  to support the reproducibility of analysis.
Where imbalances in the number of COVID-19 related adverse events between treatments are suspected or identified between treatment arms, it may be necessary to present additional tables, figures and listings for adverse events. These could include extra analyses of deaths, study discontinuations and adjustments to study treatment (e.g. discontinuation or delays to administration) due to COVID-19 adverse events. Per the FDA guidance , sponsors may also need to consider further subgroup analyses to assess comorbidities.
Some sponsors may pre-plan analyses on the treatment effect of COVID-19 related events if there is the risk that the mode of action of the investigation product(s) may lead to an imbalance of COVID-19 related events between treatment arms. In these instances, it may be necessary to consider instances of COVID-19 as a safety topic of interest. Detailed considerations on this topic are available in the PHUSE White Paper “Analysis and Displays Associated with Safety Topics of Interest: Focus on Phase 2-4 Clinical Trials and Integrated Summary Documents” . They may also choose to conduct further analyses on events that occur concurrently with a COVID-19 adverse event. This would allow sponsors to identify any additional signs and symptoms of COVID-19 reported separately from the COVID-19 adverse events based on scientific understanding of the virus and the symptoms caused at the time of the collection of the adverse event.
While we have focussed on the identification of imbalances in the distribution of COVID-19 related adverse events between treatment arms, it may also be necessary to evaluate the impact of COVID-19 on some safety topics of interest. In these instances, per Nilsson et al. , it may also be necessary to implement more complex statistical methodologies to evaluate this. It will be necessary to take into account both the number of COVID-19 related adverse events and the timing of these events in any analysis.
It is important that data is carefully reviewed to understand the prevalence of COVID-19 related adverse events in a clinical study and to investigate whether there is an imbalance in the distribution of these events between treatment arms. While many studies may not require additional analysis to report any imbalances, where an imbalance is suspected of occurring, additional analyses may be required for both individual studies and across all studies in a compound to evaluate both the prevalence of COVID-19 related adverse events and the impact of COVID-19 on patient safety. To identify these adverse events, it is good practice to use the search criteria identified for MedDRA published as part of MedDRA v23.0 until such a time that an SMQ is available to identify these events. Likewise, if the impact of COVID-19 affects treatment arms equally, analyses of adverse events from controlled data can largely remain unchanged. Similarly, where an imbalance occurs, different analytical approaches may be required.
 PHUSE Data to Knowledge
 FDA Guidance on Conduct of Clinical Trials of Medical Products during COVID-19 Public Health Emergency: Guidance for Industry, Investigators, and Institutional Review Boards – March 2020, Updated on June 3, 2020
 Guidance on the Management of Clinical Trials During the COVID-19 (Coronavirus) Pandemic – Version 3 (28 April 2020)
 End-to-End Data Flow Today: Current Processes and Challenges – Chris Price (Roche) & Praveen Garg (AstraZeneca) on behalf of the PHUSE Executive Summit (April 2019)
 MedDRA Management Committee Announcement about COVID-19 terms in MedDRA (16 April 2020)
 Guidance for Ongoing Studies Disrupted by the COVID-19 Pandemic
 COVID-19 Related Terms MedDRA 23.0 Update Spreadsheet
 Analysis and Displays Associated with Adverse Events: Focus on Adverse Events in Phase 2-4 Clinical Trials and Integrated Summary Documents
 Clinical Trial Drug Safety Assessment for Studies and Submissions Impacted by COVID-19 – Mary Nilsson, Brenda Crowe, Greg Anglin, Greg Ball, Melvin Munsaka, Seta Shahin, Wei Wang
 Analysis and Displays Associated with Safety Topics of Interest: Focus on Phase 2-4 Clinical Trials and Integrated Summary Documents
 Analysis Data Reviewer’s Guide (ADRG)
Posted by Kayley Phillpott on
23 June 2020 at 12:00 AM
Working Groups News