Context: Although bivariate survey data have demonstrated that the proportion of unintended pregnancies is increasing in Ecuador, the determinants of unwanted and mistimed pregnancy have yet to be identified Methods: A multinomial logistic regression analysis of the predictors of unintended pregnancy (unwanted and mistimed) was conducted using a subsample of women who were interviewed for the 1994 Demographic and Maternal-Child Health Survey for Ecuador. The study sample consisted of 4,534 women whose most recent pregnancy occurred between January 1992 and August 1994. Results: The multivariate analysis indicated that several explanatory variables significantly influenced the likelihood that a woman would classify her most recent pregnancy as unwanted or mistimed Among variables that independently raised the likelihood of unintended pregnancy were residence in the Sierra (or highlands) region, residence in a major metropolitan area, the number of previous births and use of a contraceptive method before the most recent pregnancy. In contrast, variables that significantly lowered that probability included residence in rural areas, living in a high-income household and giving birth at a relatively older age (i.e., 30-49 years). Conclusions: Services should focus on helping those groups of women who were identified in the analysis as being at increased risk of unintended pregnancy-high-parity women, women in the Sierra region and those in the metropolitan areas of Quite and Guayaquil.
