COST-EFFECTIVENESS OF METHODS TO SCREEN FOR ESOPHAGEAL VARICES

Purpose:  To evaluate the cost-effectiveness PillCam endoscopy versus esophagogastroduodenoscopy in screening for esophageal varices.

Methods:  A decision analytic model was developed with cirrhotic patients offered PillCam or endoscopic screening or empirical treatment without screening to reduce the likelihood and severity of variceal bleeding.  Patients could have no varices, or small (<6mm) or large (≥6mm) varices at the onset of screening. Those being screened with the PillCam could have varices that were not detected;  endoscopic screening was associated with risks of esophageal perforation or complications related to sedation. Patients receiving empirical treatment with beta blockers could incur needless expense but insignificant risk of harm. Each branch pathway led to a Markov model with these health states: No Varices, Small Varices, Large Untreated Varices, Large Varices Treated with Beta Blockers, Varices Treated with Band Ligation, Post Initial Bleeding, Post Subsequent Bleeding, and Death, either from bleeding or from other causes.  Cost-effectiveness was conducted from a payer perspective, with costs estimated using Medicare payments and effectiveness expressed in life years. All model inputs were derived from published studies. One way sensitivity analyses were carried out as were probabilistic sensitivity analyses using Monte Carlo simulations.

Results:  Base-case analysis Expected Costs/Expected Effectiveness: PillCam $22,589 /  12.8 life-years;  EGD $23,083 /  12.7 life-years; Empirical Beta Blockers $19,281 / 13.7.  Probabilistic sensitivity analysis comparing screening strategies found virtual equivalence between the two methods.  Dominance of empirical treatment over screening was robust over a variety of sensitivity tests.

Conclusions:  Empirical prophylaxis with beta blockers dominates both screening strategies in patients with liver cirrhosis. This is the second study to arrive at this conclusion, but current guidelines recommend endoscopic screening and most discussion in the literature concerns whether PillCam screening could be of equivalent or greater benefit.  Our findings slightly favor PillCam over endoscopy, but are very sensitive to several model parameters.