OBJECTIVES: According to international surveys, half of the subjects with indications for statin therapy are treated and half the treated subjects is still adherent after six month from starting therapy. Poor adherence to statin therapy is the main factor of cardiovascular prevention failure and healthcare costs increase. Unfortunately, tools to support stakeholders in monitoring current medical practice are unavailable. The aim of our work was to perform a population-based retrospective analysis to evaluate the characteristics of patients treated with statins and their adherence to treatment through the linkage of administrative databases of the LHS of Ferrara (approximately 350,000 beneficiaries). METHODS: All subjects aged >18 years receiving at least a prescription for statins between January 1st, 2004 and June 31st, 2005 were enrolled. In each subject we recorded age, sex, concurrent chronic therapy, previous hospital admissions and, starting from the first prescription, a treatment profile in the following 6 months. Adherent subjects were defined as having a PDD-standardized (mean daily dose/PDD) >0.8. The pharmacological patterns were compared among three periods lasting 6 months each. RESULTS: Treated subjects decreased from 20,445 and 20.221 of the first two periods to 17,756 of the third period mainly for reduction of newly treated subjects (from 5,108 and 4342 to 3,688). Newly treated in the third period were more frequently male, older and showed a higher prevalence of concurrent drug treatments and of previous cardiovascular hospitalizations. Adherent subjects increased from 29,7% and 31.9% in first two periods to 45.4% in third period (OR 1648, CI 1.579-1.721, p<0.001). CONCLUSION: Poor adherence was associated with younger age, lower prevalence of concurrent drug treatments and of previous cardiovascular hospitalizations. Adequate systems are required to monitor and assess actual practice, to highlight and size critical areas, to account stakeholders for practice improvement through adherence to standard rather than for cost containment.