Agree that the question is prima facie reasonable but I just don't see how it will affect decision making.
Do a PhD if you love research and want a career that incorporates bench work. Otherwise, don't.
For residencies, 10-15 years ago, the PhD could mean quite a bit (read, a lot in IM, peds, path, and (child) neuro - which I would say were the traditional fields). Nowadays, it's just another credential (part of the credentials inflation across the board) and probably less useful than AOA.
Honestly, my impression is that the people who are doing the best nowadays in securing tip top residency spots are those who do the 1 year research tracks which are often translational/clinical research. These tracks tend to select excellent students based on prior CV + MS1/2 performance, but they also put the student in a position/lab tailored towards a particular field (e.g. rad onc interest, you go to a rad onc lab), which then gives you publications in the field and networking in the field, which can be very useful come application time. You often will get another nice credential with it such as an MS or MPH or something.