Abstract
Although the rate of contraceptive use in France is high, more than one-third of pregnancies are unintended. We built a dynamic microsimulation model that applies data from the French COCON study on method switching, discontinuation, and failure rates to a hypothetical population of 20,000 women, followed for five years. We use the model to estimate the adjustment factor needed to make the survey data fit the demographic profile of France by adjusting for underreporting of contraceptive nonuse and abortion. We then test three behavior-change scenarios that could reduce unintended pregnancies: decreasing method failure, increasing time using effective methods, and increasing switching from less effective to more effective methods. Our model suggests that decreasing method failure is the most effective means of reducing unintended pregnancies, but we found that all of the scenarios reduced unintended pregnancies by at least 25 percent. Dynamic microsimulations may have great potential in reproductive health research and prove useful for policymakers.