Cost analysis of magnetic resonance cholangiography in the management of inoperable hilar biliary obstruction.

Abstract

Palliation of patients with Klatskin tumors involving both hepatic ducts is usually performed with bilateral biliary stent placement. Magnetic resonance cholangiopancreatography (MRCP) offers the ability to visualize the hepatic ducts without injection of contrast, thereby reducing the patient's risk of developing postprocedure bacterial cholangitis. We used decision analysis techniques to quantitate the cost-effectiveness of MRCP before stent placement versus routine placement of bilateral biliary stents in the setting of inoperable malignant hilar obstruction. In addition to determining which strategy was most economical, we used sensitivity analysis to identify the critical factors defining relative costs. A decision analysis model was designed comparing MRCP with subsequent unilateral biliary stent placement and double biliary stent placement approaches for palliation of jaundice in a patient with inoperable malignant hilar obstruction, as viewed from the societal perspective. Baseline probabilities, obtained from the published literature, were varied through plausible ranges using sensitivity analysis. Charges were based on Medicare professional plus facility fees or diagnosis-related group rates for out- and inpatients, respectively. MRCP with subsequent directed unilateral stent placement was the least costly approach ($3806) compared with bilateral stent placement ($4275), provided the bilateral biliary stent complication rate was >3%. Bilateral stent placement needed to confer a survival advantage of at least 7 days over unilateral stent placement to become the more cost-effective approach. The use of MRCP to guide biliary stent placement in a patient with inoperable hilar obstruction reduces the overall cost of treatment. The uncertainty of any survival advantage that bilateral biliary stent placement confers over unilateral stent placement makes cost-effectiveness difficult to assess.