Plenty of evidence; but generally from poor quality trials. The trouble is that unless a trial is randomised, controlled, and double-blinded, it get's classed as poor quality. There's also minimal funding for research (no drugs, so no money from there, not on NHS, so no money from there, people that do it are interested in doing their jobs, not researchers). You can't really get decent sample sizes, again through the above limitations, and of course, in practice, it's unethical. You also have to ask yourself who a control group would be? remember, you're trying to double blind it, so the control group has to have a sham SMT - which doesn't yet exist; and you certainly can't prevent the provider from knowing whether or not they've done something. Good quality research in this field is basically impossible using good scientific method.
If you're looking at poorer quality research, then there's plenty of evidence both ways, so you can't really draw any conclusions. The only way to really get a useable result is to use a massive sample and compare Chiro/Osteo treatment to GP/hospital treatment, based on subjective measures of outcome; with good follow-up studies; the best example I'm familiar with being Meade et al in '90 & '95 (I think); which led to SMT being recommended in best practice guidelines by the RCGP, NICE, CSAG etc.
You also get the like's of Ernst in Exeter uni who seems to have an axe to grind, and who generally quotes mostly his own work, and who say's it's all a load of rubbish, on the general principal that he can't test for it in a double-blinded, RCT, and there's not the money out there to find new, acceptable research methods to find out; I mention him because he's fairly nedia-savvy, and often gets into the mainstream press; who tend to ignore that his conclusions are often challenged by peer review.
Bear in mind also, that I graduated 4 years ago, so I'm out of date in my knowledge and access to the literature. For a better answer, you'd need to ask someone still at uni, or preferably who teaches there; where they'll be 100% up to date on the latest research, and probably conducting it.