Several studies have examined the benefits and costs of drug treatment; however, they have typically focused on the benefits and costs of a single treatment episode. Although beneficial for certain analyses, the results are limited because they implicitly treat drug abuse as an acute problem that can be treated in one episode. We developed a Monte Carlo simulation model that incorporates the chronic nature of drug abuse. Our model represents the progression of individuals from the general population aged 18-60 with respect to their heroin use, treatment for heroin use, criminal behavior, employment, and health care use. We also present three model scenarios representing an increase in the probability of going to treatment, an increase in the treatment length of stay, and a scenario in which drug treatment is not available to evaluate how changes in treatment parameters affect model results. We find that the benefit-cost ratio of treatment from our lifetime model (37.72) exceeds the benefit-cost ratio from a static model (4.86). The model provides a rich characterization of the dynamics of heroin use and captures the notion of heroin use as a chronic recurring condition. Similar models can be developed for other chronic diseases, such as diabetes, mental illness, or cardiovascular disease.