Objectives: To evaluate the impact of a high-risk antibiotic stewardship programme on reducing antibiotic use and on hospital Clostridium difficile infection (CDI) incidence rates. A secondary objective was to present the possible utility of time-series analysis as an antibiotic risk classification tool. Methods: This was an interventional, retrospective, ecological investigation in a medium-sized hospital over 6.5 years (January 2004 to June 2010). The intervention was the restriction of high-risk antibiotics (second-generation cephalosporins, third-generation cephalosporins, fluoroquinolones and clindamycin). Amoxicillin/clavulanic acid and macrolides were classified as medium-risk antibiotics based on time-series analysis findings and their use was monitored. The intervention was evaluated by segmented regression analysis of interrupted time series. Results: The intervention was associated with a significant change in level of use of high-risk antibiotics (coefficient -17.3, P< 0.0001) and with a borderline significant trend change in their use being reduced by 0.156 defined daily doses/100 bed-days per month (P=0.0597). The reduction in the use of high-risk antibiotics was associated with a significant change in the incidence trend of CDI (P=0.0081), i.e. the CDI incidence rate decreased by 0.0047/100 bed-days per month. Analysis showed that variations in the incidence of CDI were affected by the age-adjusted comorbidity index with a lag of 1 month (coefficient 0.137051, P=0.0182). Significant decreases in slope (coefficient -0.414, P=.0309) post-intervention were also observed for the monitored medium-risk antibiotics. Conclusions: The restriction of the high-risk antibiotics contributed to both a reduction in their use and a reduction in the incidence of CDI in the study site hospital. Time-series analysis can be utilized as a risk classification tool with utility in antibiotic stewardship design and quality improvement programmes.