Missing Data in COVID Adverse Events
Much of the discussion around adverse events (AEs) and serious adverse events (SAEs) in clinical trials through the COVID-19 pandemic has been focussed AEs associated with new therapies and vaccines for COVID-19 itself. But what about other clinical trials that have been impacted by COVID-19? We have already explored the many ways in which the COVID-19 pandemic has affected the conduct, analysis, and interpretation of clinical trials with respect to efficacy data. We have discussed how global quarantines and travel restrictions have resulted in site closures and interruptions to investigational product supply, which have subsequently resulted in issues such as missed assessments and treatment interruptions. But how might these issues also affect safety data?
Appropriate safety evaluations of investigational products are crucial to assess their benefit-risk ratio and to ensure that patient safety remains paramount in clinical research. An Independent Data Monitoring Committee’s (IDMC’s) duty is to ensure that the interests of patients entered onto trials are being well served. They need to ensure that the scientific integrity of the trial is maintained and that study participants are not exposed to unnecessary or unreasonable risks as a consequence of their trial participation. Virtually all clinical trials potentially pose some risk to patients under treatment, so it is of paramount importance that AEs are accurately collected and reported. The process for collecting, recording, analysing, and reporting AEs, however, is complex and can be less developed than those processes used when evaluating efficacy. This complexity can be magnified when trials are global and spread across multiple sites.
The COVID-19 pandemic has the potential to influence the reporting behaviour of AEs in clinical trials. More specifically, under-reporting of AEs could become apparent for drugs unrelated to COVID-19 as the opportunities for individual reporting of AEs may be affected due to COVID-19 restrictions, and the reporting behaviour of health professionals may also be affected. Reporting of AEs can be prone to biases and underreporting under the best of circumstances, but with the additional strain on health resources caused by the COVID-19 pandemic, it’s easy to see how these issues could be magnified.
If you know that you are likely to see an underreporting of AEs, is there anything that can be done about it? PHASTAR are involved in numerous IDMCs, many of which have been affected by the COVID-19 pandemic. Here, we present a case study of how we identified the problem in one such trial, and tried to quantify it, thus ensuring that the IDMC members had the most accurate information available on which to make their subsequent decisions regarding the future of the trial.
An understanding of the potential bias caused by the pandemic on AE reporting rates was needed in a non-randomised umbrella study, where ad-hoc comparisons for both efficacy and safety endpoints are made using historical controls and prior cohorts. A (planned) data cut off at the end of January 2020 was used to define ‘pre-‘ and ‘post-‘ COVID-19 periods. A crude incidence rate of (sum of number of adverse events/sum of total time on study during period) for all patients on-study during those periods was calculated. No adjustments were made for actual treatment exposure or time since day 0. The results revealed a clinically significant drop in reporting of adverse events in the ‘post-COVID’ period. This was seen consistently in every country, and at SOC level for all AEs, all SAEs and all Grade 3 and above AEs. The analysis helped guide discussion amongst IDMC members tasked with safety review.
Further statistical analysis is warranted, including investigation into the relationship of frequency of AEs and time since start of treatment (as it is understood that AE reporting tends to peak towards the start of a patient’s journey on a trial, particularly with the most sick patients), and also the potential relationship between AE reporting and efficacy, as the ‘post-COVID’ period coincided with cohorts displaying improved efficacy.