Regional variation in total cost per radical prostatectomy in the healthcare cost and utilization project nationwide inpatient sample database.

摘要

PURPOSE: Surgical treatment for prostate cancer represents a large national health care expenditure. We determined whether state level variation in the cost of radical prostatectomy exists and whether we could explain this variation by adjusting for covariates associated with cost. MATERIALS AND METHODS: Using the 2004 Healthcare Cost and Utilization Project National Inpatient Sample of 7,978,041 patients we identified 9,917 who were 40 years old or older with a diagnosis of prostate cancer who underwent radical prostatectomy without cystectomy. We used linear regression to examine state level regional variation in radical prostatectomy costs, controlling for the local area wage index, patient demographics, case mix and hospital characteristics. RESULTS: The mean +/- SD unadjusted cost was Dollars 9,112 +/- Dollars 4,434 (range Dollars 2,001 to Dollars 49,922). The unadjusted mean cost ranged from Dollars 12,490 in California to Dollars 4,650 in Utah, each significantly different from the mean of Dollars 8,903 in the median state, Washington (p <0.0001). After adjusting for all potential confounders total cost was highest in Colorado and lowest in New Jersey, which were significantly different from the median, Washington (Dollars 10,750 and Dollars 5,899, respectively, vs Dollars 8,641, p <0.0001). The model explained 85.9% of the variance with regional variation accounting for the greatest incremental proportion of variance (35.1%) and case mix variables accounting for an incremental 32.3%. CONCLUSIONS: The total cost of radical prostatectomy varies significantly across states. Controlling for known total cost determinants did not completely explain these differences but altered ordinal cost relationships among states. Cost variation suggests inefficiencies in the health care market. Additional studies are needed to determine whether these variations in total cost translate into differences in quality or outcome and how they may be translated into useful policy measures.