Background:
National Fund of Resources (FNR), public non-state agency of Uruguay, finances highly-specialized medical procedures and high-cost drugs. Given the impact of tobacco use as a serious Public Health problem and as a risk management strategy, in 2004 FNR implemented a Tobacco Dependence Treatment Program. It provides professional training, financing pharmacological treatment with nicotine substitutes and bupropion, provides a computerized clinical record system and performs periodic analysis of process and outcome indicators. The present work analyzes the samples studied in two consecutive evaluations.

Methods:
The objective is to detect variables, at the onset of the program, that predict the probability of success, measured in continuous abstinence for 6 months. A logistic regression was performed with: sex, age at the beginning of treatment, age at onset of tobacco use, daily consumption, maximum consumption, medication use, educational level and number of controls in the first 30 days of treatment.

Results:
824 patients were included in the analysis, the variables retained by the model were the number of controls as a factor that favors abstinence (OR = 1.33, 1.17-1.51), and mean daily consumption at admission (OR = 0.97, 0.95-0.99) and the maximum consumption (OR = 0.99, 0.97-1.0) as the factors that do not favor. Significant effect was obtained for abstinence rate, from 4 contact with the team (OR = 1.94, 1.10-3.41).

Conclusions:
The magnitude of consumption at admission is a well-known predictor, which is related to dependence degree, however the fact that in the model was maintained independently of maximum consumption, suggests that consumption at admission reflects if there are control measures implemented before onset program, indirect indicator of motivation. On the other hand, the results suggest that it is justified to concentrate time and resources necessary to increase adherence in the first month, to achieve at least 4 patients' contacts with the program.