But that sample may not be representative of the demographics of the people who got infected. It is possible that a lot more people in care homes have been infected than the grocery store sample, because of the policy of sending covid patients there (and the fact that the staff must have physical contact with each of the patients every day and therefore it will spread within a care home may very quickly).

If it is the case, you don't expect the same death rate to scale up as the wider population continues to get infected (if it is still happening, the virus is rather showing signs of going away).