If everyone wears seatbelts then 100% of the vehicle crash victims in hospitals will be seatbelt users. But seatbelts are still highly effective at reducing injuries and deaths.
That's not what the stat is saying. You seem to be mis-interpreting it in the same way some of the other people were.
Take 1000 people who are vaccinated. Measure how many get infected in a time span T. Calculate the ratio.
Take 1000 people who are not vaccinated. Measure how many get infected in a time span T. Calculate the ratio.
In the UK the infected:non-infected ratio is higher for the first group than the second. That should be impossible as it implies effectiveness is now negative. Possible root cause - the (relatively small) group of people who refuse to take the vaccine are refusing because they know they already got it, and thus have natural immunity, but the vaccine doesn't build immunity that lasts, so as time goes on the vaccinated group ends up getting infected anyway and having to fall back on building their own natural immunity. That's just speculation but otherwise it's hard to explain what is going on here.
Take 1000 people who are vaccinated. Measure how many get infected in a time span T. Calculate the ratio.
Take 1000 people who are not vaccinated. Measure how many get infected in a time span T. Calculate the ratio.
In the UK the infected:non-infected ratio is higher for the first group than the second. That should be impossible as it implies effectiveness is now negative. Possible root cause - the (relatively small) group of people who refuse to take the vaccine are refusing because they know they already got it, and thus have natural immunity, but the vaccine doesn't build immunity that lasts, so as time goes on the vaccinated group ends up getting infected anyway and having to fall back on building their own natural immunity. That's just speculation but otherwise it's hard to explain what is going on here.