Questions about the Ladapo Post-Vaccination Mortality Study
Ladapo estimated the risk of cardiac death post-COVID vaccination in residents of Florida, USA. The study was based on the Self-Controlled Case Series (SCCS) design. SCCS allows each case to serve as its own control. The exposure period was defined as the 28 days “following vaccination”. The baseline period was defined as the remaining time up to 25 weeks following vaccination. The study period was December 15, 2020, to June 1, 2022.
The primary finding of the study was:
“In the 28 days following vaccination, a statistically significant increase in cardiac-related deaths was detected for the entire study population (RI = 1.07, 95% CI = 1.03 - 1.12).”
Four publications are cited for the SCCS method. One of the referenced studies estimated the risk of death post-COVID vaccine using the SCCS design. Another looked at the risk of death post-bupropion using the SCCS design. The two other references provided description of the SCCS method.
The most problematic statement from Ladapo is:
“Follow-up began on the day of their last COVID-19 vaccination. Participants were not censored upon death, rather, they were followed for the entire 25-week follow-up period.”
The meaning of that statement is not clear. However, handling of cases where the observation period is missing, also known as censoring, is discussed directly in a cited study, Whitaker et al:
“Provided that the deaths of interest are rare, the case series method can then be applied.”
Nevertheless, the fact remains that a case simply cannot serve as its own control after death.
The questions/concerns about this study are:
No authorship is provided except for "Florida Department of Health". Authorship adds legitimacy to a publication. Why no authorship?
Strictly speaking, the mortality outcome is not possible with the Self-Controlled Case Series. To what extent does this limit the reliability of the results reported from this study? (UPDATE: Bias from "non-recurrent events" is probably small - https://pubmed.ncbi.nlm.nih.gov/30284323/ )
The authors state: "Confounding by age may be present in the 60 years or older age group, which may explain the slightly elevated risk for cardiac-related deaths following vaccination." The question is "Why?" Confounding by age occurs in the Self-Controlled Case Series due to the difference in age of the subject between the risk and baseline periods. In this study, the risk period is the one-month period following vaccination and the baseline period is the 5-month period following that. On average, the subjects are 3 months older during the baseline period than during the risk period. As a result of the slight age difference, there is expected to be a slightly greater incidence of cardiac death during the baseline period. In other words, there is a bias in this study towards lowering the estimation of the vaccination risk or even towards showing protection due to vaccination. Why aren't the study estimates of the risk of cardiac death due to vaccination considered to be conservative estimates?
The study looked at the mortality risk in the "28-day risk period following COVID-19 vaccination". That time period was further defined as "Follow-up began on the day of their last COVID-19 vaccination." For a vaccination such as mRNA consisting of two doses this definition of the risk period can significantly distort the estimate of the mortality risk from vaccination. While the risk period appears to be 28 days, in reality, it is the time from the initial dose until 28 days following the second dose. In the case of the mRNA and assuming a 28-day period between vaccine doses, that would lead to an estimate of doubling of the incidence of cardiac death as discussed elsewhere. The question is: How exactly was the mortality risk determined for the mRNA cases? (UPDATE: The authors reference the " SCCS package in R" for analysis. A search of "SCCS package in R" on Google shows a relatively recent software version that may handle this issue.)