OBJECTIVES: Using linked administrative databases from Ontario, our study examined the “real world” cost, effectiveness and cost-effectiveness of Rituximab in diffuse-large-B-cell lymphoma. METHODS: Patients were defined as those who had a diagnosis of diffuse-large-B-cell lymphoma according to ICD-O histology classification between January 1997 and December 2007 and received either CHOP (cyclophosphamide, doxorubicin, vincristine and prednisone) or R-CHOP (CHOP plus Rituximab) as first line treatment.  We used a historical cohort design to compare the overall survival, toxicity profiles, direct costs, and cost-effectiveness of CHOP before Rituximab was approved (pre-era CHOP) with R-CHOP after Rituximab approval (post-era RCHOP).  R-CHOP and CHOP patients were hard-matched on age, and then subsequently matched on propensity scores by use of a 1:1 matching algorithm.  Propensity scores were calculated from demographic and clinical history information.  We estimated resource use and direct medical costs using the linked administrative data.  To analyze censored cost data, we employed and compared different methods, including the simple non-adjusted average, the Kaplan-Meier sample average estimator, inverse probability weighting estimator, Pfeifer and Bang’s estimator (2005) and Basu’s two-part estimator (2010).   RESULTS: A total of 1131 matched pairs of patients were evaluated.  3-year overall survival was significantly improved in the post-era RCHOP group compared to pre-era CHOP (69% [95%CI 66-71] vs 59% [95%CI 56-62]; Klein test p<0.001).  Groups did not differ in the frequency of adverse events, but 3-year direct cost was significantly higher in the post-era RCHOP group.  The incremental cost-effectiveness ratio varied depending on the method employed. CONCLUSIONS: This study illustrated how different methods can be applied to observational data to estimate costs and cost-effectiveness.    The results from this study can be compared to those from clinical trials and economic models.  This will help drug decision-makers calibrate healthcare policies while helping researchers evaluate assumptions made and methods used in economic models.