The phrase gets used as though it named a category of remedy — herbs, needles, tinctures, the natural end of the shelf. It does not. It names a relationship to evidence.
What the word actually marks
Wikipedia's article puts it about as plainly as it can be put: alternative medicine refers to practices that aim at the healing effects of medicine but “by definition lack biological plausibility, testability, repeatability, or supporting evidence of effectiveness”. The defining feature is not what the remedy is made of. It is that nobody has shown it works.
That makes the label unstable in a way worth understanding, because things move out of it.
The ones that worked stopped being called alternative
Willow bark was a folk remedy for pain and fever for centuries. Its active compound became aspirin. French lilac was a traditional treatment for the symptoms of diabetes; the compound derived from it became metformin, now the most prescribed diabetes drug in the world. Foxglove gave us digoxin. The opium poppy gave us morphine.
None of those are called alternative medicine, and the reason is simply that they were tested and passed. When a remedy is shown to work it stops being an alternative to medicine and becomes medicine. What is left under the label is, by construction, the part that has not made that crossing.
“Natural” is not a safety claim
Digoxin comes from a flower and will stop your heart at the wrong dose. Warfarin came from spoiled sweet clover and is a serious anticoagulant. Nicotine, strychnine and ricin are all entirely natural. Nothing about a plant origin implies a gentle one, and nothing about synthesis implies harm.
What a plant origin does reliably imply is a variable dose, because a plant is not a factory. That is the practical difference, and it is the subject of the next piece in this section.
Inert is harmless; active is not
Some alternative remedies contain nothing that does anything. Those are, at worst, a waste of money — and at best, a placebo, which is a real effect on symptoms and no effect at all on disease.
The ones that do contain something active are a different proposition, because active compounds interact. St John's wort induces the liver enzymes that clear many prescription drugs, and can reduce the blood levels of antidepressants, transplant drugs, anticoagulants and hormonal contraceptives. That is not a hypothetical risk; it is a documented and clinically significant one. A remedy strong enough to help is strong enough to interfere.
Which is why the recommendation at the bottom of every page in this section is the same and is not a formality: tell a pharmacist what you are taking. They will check the whole list against each other, for free, without an appointment.