The COVID-19 vaccine approvals were expedited as an emergency response to the global pandemic. As with any other drug, the vaccines were required to undergo post-authorisation safety surveillance to monitor safety and efficacy data and collect additional information for vaccine use in wider populations, including pregnant women, persons with autoimmune disorders or comorbidities, etc. The pharmacovigilance plans for the COVID-19 vaccines faced several additional, compounded challenges: large-scale data collection, analysis and management and difficulty determining long-term side effects. These issues, compounding from clinician visits through data analysis, complicate the current process of detangling long COVID and vaccine side effects.
Symptoms or side effects?
Clinicians and health care professionals not only have to deal with the medical facts but also engage with patients that bring their own ideas, concerns and expectations to each visit. During the pandemic, there was incredible amounts of dis/misinformation saturating the news, social media and the internet at large. The novelty of the virus and the vaccine, the lack of a control group and the reporting challenges made it difficult for clinicians to identify between COVID, long COVID and vaccine side effects accurately. Clinicians had to comprehensively assess each patient through a diagnosis of exclusion to determine the cause, be it long COVID, vaccine side effects or comorbidity.
The industry relies on post-marketing reporting of side effects to continue safety surveillance. Still, the extensive analysis and examination required of clinicians combined with the overwhelming pressure put on healthcare systems as hospitalisations spiked during the pandemic led to further inconsistencies in reporting. There were cases early on when pericarditis was assumed to be coincidental, but with yellow card reporting, it was then included as a side effect and could be accurately monitored.
Difficulties in data collection
Collecting and assessing individual cases is critical in the greater pharmacovigilance process. These cases form the building blocks for broader aggregate analysis and understanding of the emerging safety profile, and the industry depends on reporting of adverse event reports by both clinicians and patients. However, neither party are obligated to do so, and both encounter obstacles to providing complete and accurate data. As mentioned, the overwhelming healthcare burden and lack of details for accurate assessment posed a reporting barrier for clinicians, while not all patients know their ability to report. If they are aware, patients may not know how to provide complete and accurate data. In addition, for physicians, the lack of feedback or acknowledgement of a report submission can be a demotivator to reporting. Despite these challenges, a vast amount of data is still available, though it may differ from database to database.
The vaccine’s safety surveillance had novel challenges in data collection based on the sheer volume of data available. Where other drugs may be administered to smaller groups or subpopulations, the COVID-19 vaccines were offered to all populations, not just those involved in the trials, to curb the rapidly worsening global pandemic. This means there is virtually no control group, and there is an immense pool of patients worldwide with important data – the next step is managing and analysing that data to provide actionable insights.
Big Data solutions
To facilitate data collection in addition to frontline case reporting from patients and clinicians, organisations launched data capture aids, like the CDC’s V-Safe, a post-vaccination health checker app for mobile phones. Data captured is then stored in associated databases, such as VigiBase, FAERs, VAERs or other registries that may differentiate between COVID-19 events in vaccinated versus unvaccinated patients to enable more accurate conclusions regarding vaccine safety. Identifying signals from multiple, large-scale datasets can be a complicated, time and resource-consuming process. However, automated solutions offer key opportunities for creating more routine data collection and more granularity and flexibility in data analysis.
ICON’s proprietary SIGNET technology is helping to tackle some of the data collection and analysis challenges by allowing for timely aggregation and pattern identifications within large databases. For example, starting with the VigiBase dataset containing more than 83 million rows of data, SIGNET identified cohorts of interest and provided interactive visuals that allowed for immediate appraisal and granular examination of the data using the generated details tables. We reviewed and analysed the data from multiple angles and intersected the criteria in different ways using the embedded and custom filters explicitly developed for the use case. We can thoroughly compare data from different cohorts, including vaccine-associated events and all patients with COVID-related symptoms. We can further divide and explore this data with clear visualisations to identify trends, but SIGNET can only provide as much detail as the data itself offers. The fast pace of development in the industry at the time meant there was no universal standardisation of coding in the data. As such, there are still missing elements, such as onset dates, limiting our ability to divine insights.
The dividing line between long COVID symptoms and COVID-19 vaccine side effects is not definitive – from patients’ and clinicians’ perspectives to insights mapped in several hundred-million-point databases, we are conducting good pharmacovigilance to develop a clearer picture of these afflictions. As the burden on the healthcare system has eased, we are collecting higher-quality case data and real world data from PASS. We are getting closer to answering whether these events are symptoms or side effects and are working towards insights in the near future.
For additional information on the challenges of assessing the long-term effects of COVID-19 vaccines and the details of the pharmacovigilance efforts, watch our on-demand webinar.