I'm not sure that the medical profession does itself any favours in saying that everything they do has been tested and trailed, because many things have not been. However all new drugs and therapies do have to go through trials before they are approved in practice, and this has been the case for a long time now. However many treatments were derived a long time ago, and some have little evidence behind them (especially in mental health - there is still a huge amount we don't know about the mind). Many of these are being weeded out, partly because when money is tight there is a stronger push towards only funding things that can be shown to work.
Of course clinicians make mistakes (I used to work in clinical risk, so know this very well), but then to err is to be human, and it is likely that if there was the same sort of volume of treatments given by the alternative sector there would be a similar error rate. That's not to say that time and effort should not be spent on trying to get it right, and understanding and rectifying what went wrong. Nor that it is not a tragedy when things do go wrong.
With the chemotherapy, it should be remembered that chemotherapy is the administration of poison. I think more thought should be put into deciding whether or not to embark on course of chemo, as often the benefits are slim and the side effects horrendous. But I don't think that we are very good at accepting that sometimes even though there might be something that sometimes offers hope the best approach is not to go down that route.
One thing I think that people forget with alternative medicines is that herbs and things can also be toxic. Just because they are natural does not mean they are safe. Indeed many medicines are based on plants and minerals, including things like digitalis (for heart attacks, but can kill you) and some of the most deadly poisons are natural (like arsenic).
One problem with some of the more holistic therapies is that the holistic approach almost certainly is a part of the therapy, and traditional research methods are not good at taking this into account. I think all therapies would be wise to have at least associations with standards that their patients/clients could access, otherwise the very good practitioners get affected by the poor reputations of the charlatans.