This application proposes a study of the process of adoption of Norplant and its patterns of use in two sites in the U.S. The research is important because Norplant could potentially have a great impact on the health and well-being of American women. This is because Norplant, being highly effective and requiring no action from women after its insertion, is likely to become widely used and thus it has the potential for significantly decreasing the incidence of unintended pregnancies among some groups. We propose a longitudinal study of a cohort of women beginning a new segment of contraceptive use or nonuse, enrolled over one year at sites in Atlanta GA and Charlotte NC. We will investigate factors affecting women's decision to choose Norplant or other methods or nonuse at that time. We will then study the patterns of choices they subsequently make about Norplant versus other methods as they progress in their contraceptive careers over the course of three years, and how changes in their lives, experience with various methods, and other factors affect the probabilities and timing of contraceptive change, discontinuation and unintended pregnancies throughout this time. After an initial 6-month period to develop instruments and procedures, we will enroll 800 Norplant acceptors and 400 each of women who choose instead to use primarily the pill, condoms, and no method, and also 100 each of women choosing sterilization and remaining methods, using a stratified random sample design. We will conduct a personal 45-minute interview with these women at the time of enrollment, to determine reasons for choosing their method and other data necessary for investigating determinants of choice and of continuation and use-effectiveness. To determine women's changes to subsequent methods and associated continuation and use-failure rates, as well as the factors accounting for these, we will longitudinally follow the baseline users of Norplant, pill, condom and nonuse. Users of other methods will not be followed because of budget constraints. We will conduct 15-minute reinterviews with the follow-up sample at 12-month intervals after enrollment over three years. We will perform final analysis and paper writing during the final fifth year. Methods of analysis to be used include: well-established descriptive and multivariate techniques such as multiple regression, multinomial logistic regression, and hazards models; and newly developed techniques required to analyze choice of sequential multiple methods and discontinuation and use-effectiveness with these methods (mixed point processes models with random and fixed covariate effects and semi-Markov models with covariates).
