There's a secondary issue here. Sometimes, we look at outcome data: what is the decrease in getting sick or dying after taking a vaccine? This is the number you

There's a secondary issue here. Sometimes, we look at outcome data: what is the decrease in getting sick or dying after taking a vaccine? This is the number you actually care about. Other times, we look at a number that correlates: what is the increase in _antibody levels_? One example from this article: > A recent (but not yet peer-reviewed) study of more than half a million U.S. veterans showed that the Johnson & Johnson shot’s protection against infection (whether or not that leads to disease or hospitalization) had plummeted from 88 percent to 3 percent by mid-August, while the other vaccines’ effectiveness had declined much less. Research published in the past few months also signals that Moderna’s shot beats out Pfizer’s in terms of both antibody count and hospitalizations prevented, while a National Institutes of Health study released earlier this month found that Moderna’s booster shot lifted participants’ antibody levels a bit more than Pfizer’s, and that both mRNA boosters were miles ahead of J&J’s. My understanding is that we don't know how much antibody levels need to increase to provide a level of safety. Or, for that matter, if at some point, further antibody levels don't help at all! If there's a light rain, a full-body drysuit will keep you dry. But so will an umbrella! The drysuit is overkill! We certainly want antibodies against the virus, but I don't believe we know how much is overkill. One reason this might matter is if we could have given smaller shots to people and ramped up injections much sooner. Or even in the present, we could send more injections to other countries for lower cost, if a certain volume of vaccine could be given to more people.