Content area
Abstract
This dissertation is an investigation of the patterns and consequences of contraceptive risk-taking among users of female methods of contraception (the sponge, cervical cap, and diaphragm). Secondary analysis is performed on data from two clinical trials of barrier methods: a study of 1437 users of the diaphragm and the sponge conducted by Family Health International in 1981-1982, and a study of 1394 users of the diaphragm and the cervical cap conducted by the University of California Medical School in 1981-1985.
The two general issues which are addressed are (1) the extent to which contraceptive risk-taking contributes to contraceptive failure, and (2) the characteristics which predict contraceptive risk-taking. Specific hypotheses about the types of women likely to take contraceptive risks are elaborated based on a model of risk-taking as an expression of ambivalence toward having an unplanned pregnancy.
Piece-wise constant hazard models are used to estimate the risk of contraceptive failure for various subgroups of women, including and excluding periods of contraceptive risk-taking. Logistic regression analysis with random effects is employed in the analysis of the correlates of risk-taking behavior.
Results indicate that contraceptive risk-taking is an important direct cause of contraceptive failure. The risk of failure is much higher during intervals of inconsistent contraceptive use than during intervals in which no risks were taken. Furthermore, most factors which have been associated with contraceptive failure in previous research are not associated with the risk of failure during consistent use. An important exception is the finding that women who have had a child are at higher risk of failure than nulliparous women if they use the sponge or the cervical cap; there is no such effect for diaphragm users.
Factors which are associated with the propensity to take contraceptive risks include having had a previous abortion, being dissatisfied with one's contraceptive method, having low educational attainment, and having taken previous contraceptive risks. Hypotheses which predicted that risk-takers would have characteristics associated with ambivalence toward pregnancy received little support in the empirical analysis.
