Bacillus Calmette-Guérin (BCG) ALL-IN Meta-analysis

The COVID-19 pandemic, which began in 2020, caused widespread illness, deaths, and strain on healthcare systems worldwide. Early in the pandemic, scientists explored whether the Bacillus Calmette-Guérin (BCG) vaccine, originally developed to prevent tuberculosis, might also protect against COVID-19 and other respiratory infections. While early observations hinted at possible benefits, results from twelve large randomized controlled trials (RCTs) have since shown that BCG vaccination does not appear to meaningfully reduce the risk of COVID-19 infection, severe illness, or death.

However, it is important to continue combining all available and emerging data in a rigorous way to confirm these findings and to test whether small or time-dependent effects might exist across different populations or study designs. This project uses a “living and prospective” meta-analysis approach, meaning that new studies are added to the analysis as soon as their data become available rather than waiting for all trials to be completed and published.

Our statistical approach is called ALL-IN, short for Anytime Live and Leading INterim meta-analysis. This meta-analysis framework allows results to be updated continuously, essentially providing a live evidence summary that evolves as new information arrives. Within this framework, we use two specialized tools:
• E-value: This is a measure that shows how strongly the observed results support a real effect of the vaccine rather than being due to chance or unmeasured factors. The higher the E-value, the more confident we can be that the association is not explained away by unknown variables.
• Anytime valid confidence intervals: Traditional analyses assume data are only examined once at the end of a study. In contrast, anytime valid confidence intervals allow researchers to look at data repeatedly, day by day, without inflating the risk of false conclusions. This provides a way to monitor evolving evidence responsibly as new trials report results.

By applying these methods, the ALL-IN analysis combines ongoing and completed RCTs into a continuously updated estimate of BCG’s possible impact on COVID-19 outcomes. Although published trials so far suggest little or no effect, this living analysis can detect subtle patterns and confirm consistency across studies in near real time. Such dynamic evidence can guide researchers and policymakers on whether additional BCG trials are warranted or whether attention should shift to other preventive strategies. This study demonstrates how modern, real-time meta-analytic methods can provide continuously updated and statistically reliable insights from multiple trials at once. Even when confirming that BCG vaccination is unlikely to help against COVID-19, this work illustrates how adaptive, transparent monitoring of emerging data can accelerate public health decision-making during global health crises.