Imagine there is an injection being recommended, and it’s to lower your risk of getting a disease. The injection is new, meaning almost no time has passed to do a proper longitudinal study on its efficacy, and more importantly its safety.
While there are no long-term data available, there is a massive sample size of individuals who have taken the injection. This effectively represents a large cross-sectional dataset of injection information.
The question is, can that cross-sectional dataset, which normally would not substitute for longitudinal data, now substitute for longitudinal data because of its massive sample size?
Is there something in the statistical arsenal, specifically probability, that would reveal chronic symptoms surfacing as acute symptoms for a small but significant portion of the population (the tail of some distribution) if the sample size is large enough?