The best evidence for alcohol screening and brief intervention in primary care supports efficacy, at best, not effectiveness: You say tomāto, I say tomăto? That’s not all it’s about

Several meta-analyses find that patients with nondependent, unhealthy, alcohol use (identified by screening) who are randomized to receive brief counseling subsequently report drinking less than those randomized to control groups [1-4]. Most studies do not have biological confirmation, and those that do generally do not find effects on such outcomes, nor on any hard outcomes (though some do [4]). The lack of evidence for effects on these outcomes raises the possibility that alcohol screening and brief intervention efficacy remains unknown, since the modest effects on drinking could easily be due to social desirability bias [5] (i.e., patients are told they are in a study of counseling to reduce their drinking, they are randomly assigned to brief counseling with that goal, and then they are asked whether they reduced their drinking, so they may therefore report less drinking to please the interviewer; this would be expected to occur more often in an intervention than in a control group). This possibility should be resolved by further study (i.e., randomized trials with biological outcomes). But if we take the positive trial findings as being valid— and many scientists and practitioners have [6]—the question is whether they represent efficacy or effectiveness. This question is important because if alcohol screening and brief intervention (ASBI) has efficacy, the next steps would be to determine how to implement it and retain effectiveness in real-world practice. But if the evidence