The use of non-randomised study designs to evaluate vaccine safety can be challenging due to 1) differences between vaccinated and unvaccinated people and 2) difficulties in determining when to start follow-up for unvaccinated. Self-controlled-case series (SCCS) avoids the first challenge by comparing adverse events within individuals, with each person acting as their own control. However, this design requires outcome events not to influence vaccination timing. This condition is often violated when an adverse event can either delay or completely prevent future vaccination and is especially problematic when studying vaccines intended to be given in multiple doses. Various approaches to address this exist, but how they compare is unclear. A different approach known as target trial emulation avoids the second challenge through designing a hypothetical trial and then emulating this using existing data to answer a research question. However, it’s been rarely applied in the context of vaccine safety.
We will use two case studies to compare these study design options: a known safety problem of intussusception (a serious bowel disorder) following rotavirus vaccination, and acute neurological outcomes following COVID-19 vaccination. Using SCCS, adverse events will be compared between high-risk times shortly after vaccination and baseline risk times within individuals. In target trial emulations, we will evaluate each vaccine dose's safety by comparing outcomes in vaccinated and unvaccinated individuals matched weekly on factors influencing vaccination and outcomes. This approach will enhance our understanding of these study designs, clarify solutions for multi-dose vaccines, and provide an updated evaluation of COVID-19 vaccine safety.
We aim to compare approaches to handle violation of a key SCCS assumption when adverse events influence subsequent vaccination by either delaying or contraindicating it (event-dependent exposures). We will also compare implementation of target trial emulation framework for safety evaluation of a vaccine when it is either rapidly administered (COVID-19 vaccines) or routinely administered (rotavirus vaccine). We will leverage the established safety signal of intussusception following rotavirus vaccination as a case study of methodological interest and evaluate the safety of COVID-19 vaccine on acute neurological outcomes as a different case study while exploring relevant methodological aspects.
We will include first recorded cases of intussusception (infants for rotavirus vaccine evaluation) and acute neurological outcomes (adults of aged above 18 years for COVID-19 vaccine evaluation) in the SCCS analyses. The 1-21 and 1-42 days after vaccination will be used as risk periods for outcomes for rotavirus and COVID-19 vaccine study respectively. Conditional Poisson regression will be used to estimate relative incidence and 95% confidence interval for the safety outcomes for each dose and vaccine. We will compare the use of pre-vaccination window of varying lengths and use of modified SCCS model for event-dependent exposure as approaches to overcome violation of event-dependent exposure assumption in SCCS. We will also explore the use of varying lengths of risk periods and observational periods as sensitivity analyses.
In the target trial emulation, for each dose, exposed and unexposed individuals will be matched on pre-defined baseline confounders for both case studies. Weekly sequential trials will be emulated, and individuals will be followed to the earliest of outcome event, death or deregistration from general practice. Cox proportional hazard models will be used to estimate hazard ratio and 95% confidence intervals for each outcome.
We will utilise the linkages to HES-APC to ascertain hospitalised cases for outcomes.
Acute neurological outcomes (Guillain barre syndrome, Bell’s palsy, transverse myelitis, generalized convulsions, acute disseminated encephalomyelitis, encephalitis, and aseptic meningitis) and Intussusception.
Ian Douglas - Chief Investigator - London School of Hygiene & Tropical Medicine ( LSHTM )
Monica Mtei - Corresponding Applicant - London School of Hygiene & Tropical Medicine ( LSHTM )
Adoracion Navarro Torne - Collaborator - GSK
Anna Schultze - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Edward Parker - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Ian Douglas - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Yun "Angel" Wong - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation