I think the real reason is that these sample missions want their success to not be tied to any other infrastructure that would increase the number of possible failures the mission will have.

For example for the Osiris-REx mission, it launched last 2016 and will comeback this 2023. If it had to rendezvous with the ISS for its sample return, you would have to ensure that nothing would happen to the ISS that would jeopardize the sample return. Even if you would say that there are a lot of contingencies for the ISS, 8 years is long enough for a lot of things to happen.

Imagine if something did happen to the ISS then the OSIRIS-REx mission would all be for nothing. The increase in potential failures outweigh the savings.