Admissions rates are a lie. Having nothing but small clinics and private practice offices can make it difficult to get anything beyond shadowing exposure, though I agree not impossible. It also decreases the diversity of what you have to choose from. Not all LACs are created equal, places like Amherst and Williams obviously have a huge array of ECs to choose from, but not all LACs do, which is why I suggested that an applicant should make sure. It is in my opinion not true to say that medical research is mostly confined to medical schools. Many non-medical school affiliated universities have robust biomedical and medical research scenes, and another large chunk of biomedical research can be had at non-university affiliated institutions and corporations.
Most medical students at top research medical schools have a background in science research, look it up in the MSAR. I would characterize it as "a lot" (often 1-2 years minimum) but that is dependent on subjective assessment. Regardless, I never said it had to be science or medical research, only that the ideal LAC will have a robust research environment to pick from.