This article uses linked data from the 1995 Morocco DHS calendar and the 1992 Morocco DHS service-availability module to study the effect of service environment on contraceptive discontinuation, switching, and adoption of a modern method following a birth. The 1995 Morocco DHS also collected information on the source of supply for each episode of use of a modern method recorded in the calendar, allowing study of the association between the source of supply and discontinuation and switching rates. Multilevel event-history models are used to evaluate the impact of individual-level sociodemographic characteristics and community-level indicators of family planning service provision. The findings show that the presence of a nearby public health center is associated with higher modern-method adoption after a birth and lower method-failure rates; the presence of a pharmacy is associated with lower discontinuation due to side effects or health concerns. The degree of method-choice potential has a positive impact on both the rate of switching from the pill to another modern method and on modern-method adoption after a birth.
PIP: This study examined the relationship between the service context and the separate components of contraceptive adoption and continuation in Morocco. Data were obtained on contraceptive adoption and continuation from the 1995 Demographic and Health Survey (DHS) among 4753 women 15-49 years old (3324 were interviewed in 1992) and on the service context from the 1992 DHS Service Availability Module. Four reasons were given for switching or discontinuation. Switching was due to pregnancy, adopting another modern method, adopting a traditional method, or not using any method. Event history analysis (discrete time) findings reveal that the number of methods available significantly increased postpartum adoption and only weakly related to switching from the pill. Women who used a nongovernment source were more likely to discontinue using for method related reasons and to stop using the pill at discontinuation. The availability of public health centers within 5 km increased postpartum method adoption. The availability of pharmacies in the community significantly decreased the risk of discontinuing the pill because of side effects and health concerns. One caveat is that considerable unexplained cluster level variation is likely to be due to service quality measures, which were not available. Estimates could also be biased due to endogeneity of variables.