Identification of a population at high risk for Clostridium difficile infection (CDI) would enable CDI prevention strategies to be designed. The purpose of this study was to create a clinical risk index that would predict those at risk for CDI. A CDI risk index was therefore developed, based on a cohort of hospital patients given broad-spectrum antibiotics, and divided into a development and validation cohort. Logistic regression equations helped identify significant predictors of CDI. A scoring algorithm for CDI risk was created using identified risk factors and collapsed to create four categories of CDI risk. The area under the receiver operating characteristic (aROC) curve was used to measure goodness-of-fit. Among 54 226 patients, 392 tested positive for C. difficile. Age 50-80 years [odds ratio (OR: 0.5; P<0.0116)], age >80 years (OR: 2.5; P<0.0001), haemodialysis (OR: 1.5; P=0.0227), non-surgical admission (OR: 2.2; P<0.0001) and increasing length of stay in the intensive care unit (OR: 2.1; P<0.0001) were significantly associated with CDI. A simple risk index using presence of significant variables was significantly associated with increasing risk for CDI in both development (OR: 3.57; P<0.001; aROC: 0.733) and validation (OR: 3.31; P<0.001; aROC: 0.712) cohorts. An OR-derived risk index did not perform as well as the simple risk index. This easily implemented risk index should allow stratification of patients into risk group categories for development of CDI and help fashion preventive strategies.