Chronic care management for Medicare patients.

is comparable with other clinical interventions considered to be effective and indicated. Dr Robey and colleagues are correct in suggesting that past abstinence is one of the strongest predictors of future abstinence. We also acknowledge that skewed data in our sample may represent dichotomous populations with respect to duration of abstinence. Further analyses of moderators within our population are exploring whether duration of previous abstinence moderates the efficacy of varenicline using a reduce-to-quit approach. It may be possible to define populations who may derive the greatest benefit from a reduce-to-quit approach with varenicline.