Anecdotal is real evidence, it is simply not the most reliable evidence.
A doctor tries this on sick patients, they get better, surely he should report that. The experience of the doctor and the patients is definitely something that would be recorded as "evidence", not proof, but certainly warrants further study and perhaps experimentation and clinical trials by other doctors.
It may not be the best study, it may not be "proof", but you should never make "the best" the enemy of "better".
Now if you had many, many anecdotal reports that using this treatment was harmful I could see the alarm in it. But without that you have no real evidence to be alarmed.