How about maintenance? There's ample evidence from the UK, Switzerland, and elsewhere showing that diacetylmorphine maintenance (treating heroin addiction with heroin) is incredibly effective[0][1].
The effectiveness of treating existing users is the primary concern, but a nice bonus is that it also seems to have served as a deterrent for new heroin users due to its social image as a medical problem, rather than a criminalized behavior[2].
(Remember that, when administered in a controlled setting and in predetermined quantities (and purity), heroin use carries few of the dangers commonly associated with its use on the street, and is fairly similar to the related drugs that you already receive in the hospital[3])
[0] https://en.wikipedia.org/wiki/Heroin_maintenance
[1] http://www.nejm.org/doi/full/10.1056/NEJMoa0810635
[2] https://en.wikipedia.org/wiki/Heroin_maintenance#cite_note-9
[3] Such as morphine (the similarity of the names "morphine" and "diacetylmorphine" is not a coincidence!)
That's obviously better for users, but it's better for wider society too. People who are addicted to drugs will take drugs, so forcing them to do so hidden away in places without safe needle disposal means members of the public or cleaning staff risk needle stick injury.
That's a shame because research is difficult and also ignored if politicians find it doesn't suit their view.
Substitution therapy (eg, methadone instead of heroin) has strong supporters and detractors and it's hard to get good quality research to see whether substitution is better than abstinence.
Indeed, Russell Brand (comedian, actor, addict) has been talking about the benefits of abstinence based treatment and asking why it's not more available in England, which suggests to me that most people don't get abstinence based treatment. (Although he's talking about drugs and not alcohol).
It is weird that cognitive behavioural models are really old now and are only just starting to get a lot of use.