BACKGROUND  Randomized clinical trials remain the gold standard in evaluating different drug therapies on outcomes but are resource intensive.  Retrospective studies using observational data are inexpensive but prone to selection bias due to non-random differences between the intervention and comparator groups.  The Propensity Score (PS) method is a novel, multivariate adjustment procedure that reduces confounding and selection bias.  METHODS: This case-control study used the Health Facts® database (Cerner Corporation, Kansas City, MO), which integrates patient information from hospitals throughout the United States.  Cancer patients receiving rasburicase or allopurinol were eligible for study inclusion. Both drugs reduce uric acid (UA) elevation otherwise resulting from tumor lysis syndrome.  The PS is the probability of receiving rasburicase and is calculated from a non-parse logistic regression model that is adjusted for age, demographics, lab tests, ICU admission, Charlson Comorbidity Index, hospital characteristics, and other factors. Each rasburicase patient’s PS is matched via a 5:1 digit Greedy algorithm to that of an allopurinol patient to reduce confounding.  Alternatively, the PS is added as another predictor in a non-parse Generalized Linear Model when estimating the effect of rasburicase on outcomes. RESULTS: Of 8,257 patients, 71 rasburicase patients met our inclusion criteria and were matched to 71 allopurinol patients.  Before matching, the median [25th, 75th percentile] total costs of rasburicase patients vs. allopurinol patients were $30,681 [$17,259, $58,241] vs. $18,106 [$8,937, $39,084].  After initial matching, median costs were lower with rasburicase: only $30,681, vs. $58,438 (P = 0.007), while pre-Rx minus post-Rx mean UA differences were significantly greater with rasburicase vs allopurinol (9.3 vs. 0.9 mg/dL, P<0.001).  CONCLUSIONS: Multivariate PS adjustment reduces selection bias and confounding when quantifying treatment effects of drug therapies.  A preliminary PS analysis demonstrated that rasburicase was associated with greater UA reduction than allopurinol, as well as lower medical care costs.