OBJECTIVES: Clostridium difficileinfection (CDI) has been found to be associated with increased inpatient length of stay (LoS), however the causal direction is unclear. Many have attempted to correct for potential endogeneity bias by case matching patients at the hospital level or by using instrumental variable regressions.  We propose an alternative method using propensity score matching on a nationwide dataset to isolate the added effect of the disease using hospital level data. METHODS: Using the Hospital Episode Statistics (HES) dataset for England, four propensity scoring techniques were tested to estimate the causal effect of CDI on LoS of patients over 50 years old, who have been diagnosed with diabetes, chronic obstructive pulmonary disorder (COPD), heart failure, and/or chronic kidney disease.  Methodological variations include radius matching, nearest neighbour matching with and without replacement, and kernel matching. RESULTS: Controlling for a number of covariates, nearest neighbour matching with replacement produces the least biased and most consistent estimates at 15.22 days longer LoS after contracting CDI, with a post-matching pseudo-R2value of zero and a mean absolute standardized bias of 0.51% compared to 41.53% before matching. Radius matching produces less consistent results at 17.05 days, with a mean standardized bias of 2.93% after matching.  Kernel matching does not adequately account for bias likely due to the largely unbalanced nature of the treatment versus control group. CONCLUSIONS: Propensity score matching serves as an alternative method to traditional regression approaches to estimate the causal effect of CDI on patient LoS.  The results are consistent with those previously derived in the literature and demonstrate the heavy burden of CDI on the English National Health Service.