Impact of study design on vaccine effectiveness estimates of 2 mRNA COVID-19 vaccine doses in patients with stage 5 chronic kidney disease

Abstract

I n observational studies, the World Health Organization recommends the test-negative design (TND) to assess corona virus disease 2019 (COVID-19) vaccine effectiveness (VE), due its accuracy and efficiency.1 The TND is a variation of the case–control design that restricts the study population to individuals who are tested for severe acute respiratory syndrome corona virus 2 (SARS-CoV-2) and have symptoms consistent with COVID-19.2 Although this design helps reduce bias resulting from differences in healthcareseeking behavior,1 a major barrier when using administrative healthcare data is the requirement for individuals who are tested for SARS-CoV-2 infection to fulfill a specific case definition (i.e., being symptomatic), which requires symptom data to be well recorded. This approach presents particular challenges for smaller specialized populations, such as individuals with stage 5 chronic kidney disease (CKD), as the requirement for symptom information may further decrease an already limited sample size.

The vaccine literature suggests that effectiveness estimates in the general population may vary by study design.1,3 However, to our knowledge, no studies using the same population and time period have evaluated the impact of study design on COVID-19 VE estimates in the CKD population, and no study has used the TND. In this study, we examined the VE of 2 doses of mRNA COVID-19 vaccines (compared to unvaccinated individuals) against SARS-CoV-2 infection and severe outcomes (hospitalization and death) using 3 study designs (i.e., the TND and pseudo-TND [pTND], which was restricted to individuals tested for SARS-CoV-2 infection, with the former, which was restricted to individuals with recorded symptoms; and a cohort design) in patients with stage 5 CKD.