The median delay between vaccination and the onset of narcolepsy in that case was 42 days. Not nearly the four years magic number that sharken was waiting for. I would call that soon after it is administered. Far more people are well beyond that period after the Pfizer-Biontech vaccination than were ever administered Pandemrix.

> And the same is also true for covid's long term effects

Those long term effects are visible soon after infection. Both of your examples seem to be in support of my point.

In this context it is not as relevant how long the median time is for side-effects to show, but rather how long it takes before the vaccine is no longer used.

When you said "visible" I assumed you mean visible to the OP - but you meant visible in general. So you are not wrong, but as we can see from the case of Pandremix, a median of 42 made it take two years before the usage of the vaccine was stopped.

There are a lot of reasons why it would probably be different in the case of covid (e.g. more media presence from the beginning) but in general, you cannot directly compare thus number (like the 42 days) with the 4 years of the OP.

You could argue that the same could be happening with COVID vaccines, but I think that is unlikely given the unprecedented level of worldwide scrutiny on these.

In the case of Pandremix, the vaccine was continued to be used for 2 years or so after the first cases of narcolepsy became visible.