Advances in risk-benefit evaluation using probabilistic simulation methods: an application to the prophylaxis of deep vein thrombosis.

Abstract

OBJECTIVE: To demonstrate the use of probabilistic simulation modeling to estimate the joint density of therapeutic risks and benefits. Published data are used to introduce the risk-benefit acceptability curve as a novel method of illustrating risk-benefit analysis. STUDY DESIGN AND SETTING: Using published data, we performed a second-order Monte Carlo simulation to estimate the joint density of major bleeding and deep vein thrombosis (DVT) secondary to enoxaparin or unfractionated heparin. Within a Bayesian framework, beta-distributions for the probabilities of experiencing a DVT and major bleed were derived from the clinical trial, and incremental probabilities were calculated. RESULTS: The incremental risk-benefit pairs from 3,000 simulations are presented on a risk-benefit plane. To accommodate different risk preferences, the results are also illustrated using a risk-benefit acceptability curve, which incorporates different risk-benefit acceptability thresholds (mu), or the number of major bleeds one is willing to accept in order to avert one DVT. Finally, a net-benefit curve is used to illustrate the risk-benefit ratio and the derivation of 95% confidence intervals around the ratio. CONCLUSION: Modern simulation methods permit the estimation of the joint density of risks and benefits with their associated uncertainty, and within a Bayesian framework, facilitate the estimation of the probability that a therapy is net-beneficial over different preference thresholds for risk-benefit trade-offs.