I know, I know, if it isn't a pill it can't possibly work. The list of references at the end of that link (which is the ONLY thing I referred them to) is mostly papers from peer reviewed lit, though.
http://www.ncbi.nlm.nih.gov/pubmed/...nel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum
RCT with a high rate of cessation.
http://www.ncbi.nlm.nih.gov/pubmed/...nel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum
Another RCT with a high rate of smoking cessation.
http://www.ncbi.nlm.nih.gov/pubmed/...nel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum
RCT showing no difference
Regardless, a quick romp through the literature reveals that smoking cessation rates at one year following hypnosis routinely come up in the mid-20% range or higher, which is about the same as varenicline or intensive behavioral modification groups.
MORE effective than the others? Maybe not. But on balance, probably better than nothing.
The cochrane database is nice in some ways, not so nice in others.
GENERALLY when they say something is effective it's true.
A lot of times, though, they quibble over minutiae and require gigantic numbers of independently run homogeneous trials to pronounce something effective. Problem being that's a near impossibility without some serious drug money behind it. Not talking conspiracy theory here, just the truth. AstraZeneca will front tons of money to get hundreds of patients to try the brand new SDNRI/antipsychotic du jour versus prozac. Who exactly will fund an RCT of hundreds of patients getting an exercise vs. TAU
And if you really understand the nitty gritty of statistics, you should realize that you can usually get a pretty decent idea of whether something is effective or not WITHOUT the statistical (rather than scientific) rigor that the cochrane database demands.
Don't get me wrong, I think they're doing a good thing, but if we sit around and wait for enough studies to be done on non-pharmacologic interventions in just about anything for cochrane to make a positive recommendation of the intervention, we'll still be sitting around and waiting when the universe implodes.
And I'll give them some serious credit for how they say things. "we have not shown that x is effective" note the syntax not "we have shown that x is NOT effective" which is a horse of an entirely different color.
That's what happens when you demand the high level of statistical rigor and trial homogeneity that they do and you only get 9 heterogeneous RCTs to look at. Then again, 9 heterogeneous RCTs is really a butt load of information.
Given the dismal rates of efficacy of cold turkey and/or patch without counseling, and the fact that the patient might have a psychological/personal/financial barrier to undergoing counseling, and you may not want to create a psychotic varenicline zombie, it might not be a bad idea to say 'eh what the heck, it's worth a try' to hypnosis.
BTW, one way or another, I haven't made up my mind about hypnosis. Sounds kinda hokey. But even Cochrane agrees that there is some positive evidence for it in management of pain, both psychologically and in terms of reduced use of analgesics, so why not for something else that's very physical and psychologically uncomfortable?