Defining “alternative medicine” is the chief pitfall in any discussion of the topic. It is a nebulous term.
The National Institutes of Health defines alternative medicine as a “non-mainstream approach [used] in place of conventional medicine” – effectively exiling it to a desert island and emphasizing that it can only be defined by its relation to mainstream medicine. The NIH notes that true alternative medicine is uncommon, because most people combine it with conventional techniques.
More helpful, I think, is viewing alternative medicine as “proto-medicine,” i.e. as techniques that may some day be adopted into conventional medicine if well-conducted clinical trials show a favorable ratio of benefit to harm.
The classic example of an alternative medicine moving into conventional medicine is the heart medicine digitalis. In the 1700s, Dr. William Withering in Shropshire, England heard that a local witch could do something he couldn’t: successfully treat dropsy (which is today called “edema”). Gaining the witch’s cooperation, Withering painstakingly discovered that, among her potion’s dozens of ingredients, the foxglove plant provided the therapeutic effect. He proved this by conducting formal trials of foxglove, whose scientific name is digitalis. All of this took nine years, but Withering, and the unnamed witch, achieved immortality in medical circles, and their work has benefitted untold numbers of patients. Digitalis derivatives remain in use today.
Physicians are (or should be) pragmatic. If something is provably successful, it deserves to be adopted into conventional medicine and to shed its “alternative” label. To make such a jump, a high level of evidence is required. Investing time, energy, or self in an approach that is unsupported by evidence, no matter how “natural” it seems, is an invitation to disappointment, and worse.