Contraceptive use has been identified as one of the primary factors contributing to fertility decline; however, the link between contraception use and fertility is weaker and less understood in sub-Saharan Africa. For example, despite a large increase in the contraceptive prevalence rate (CPR) in Tanzania since the early 1990s, its total fertility rate (TFR) has remained high. There has been little research on the impact of method mix in this relationship. An understanding of the dynamics of the relationship between contraceptive use, method mix and fertility would provide an important evidence to advocate and plan for future scale-up of and investment in family planning programs. Main question/hypothesis
What role has changes in contraceptive use and method mix played in determining fertility trends and differentials in Tanzania? Methodology (location, study design, data source, time frame, sample size, analysis approach) The study analyzes secondary data from the last three Demographic and Health Surveys (DHS) in Tanzania (1991/2, 1996, 2004/5) and the 199 Reproductive and Child Health Survey. The sample size of women 15-49 ranged from 4,029 to 10,329. Analysis is conducted at the national level and by urban and rural residence. Trends and differentials in TFR and CPR are assessed through calculating absolute and percent changes between surveys and testing for statistical significant changes. Bongaarts’ Proximate Determinants Model is constructed based on three indices: contraception (Cc), postpartum infecundability (Ci), and marriage/cohabitation (Cm). Due to insufficient data, abortion (Ca) is not included. In addition, relationships between variables are assessed using non-parametric Spearman’s correlations. Results/Key Findings The TFR, given CPR, was higher than expected in rural areas; however, the opposite is true in urban areas. To explain possible reasons for these patterns, 12 iterations of Bongaarts’ model were created- national, urban, and rural residence for each survey. Aside from one instance (1999 urban), there was consistency between the predicted and observed TRF, with a slight overestimation of urban TFR and underestimation of rural TFR. This is attributable, in part, to not including unmarried women or abortion. Examining trends in the indices revealed that contraception, although not the largest inhibiting factor, varied the most both overtime and between urban and rural areas, suggesting contraceptive use has played a key role in determining fertility trends and differentials. In rural areas a surprising pattern was found; the proportion of in-union women increased over time. Traditionally the opposite is found due to increases in age of first marriage. Therefore much of the fertility inhibiting gain in rural contraceptive use went to cancelling out resulting increases in Cm. Across the four surveys, urban CPR was consistently double rural CPR. In both urban and rural areas the largest average annual CPR increase happened between the first two surveys (1991/2 to 1996); 18.4% and 16.7% respectively. Between 1999 and 2004/5, the average annual increase slowed to less that 2% in both areas. Changes in method mix mirror those of CPR; a large increase in the use of long-acting methods during the between the first two surveys, followed by relatively small changes to method mix thereafter. This resulted in little change in average use effectiveness post-1996. An attempt to account for the relationship between method mix and fertility produced inconclusive results. This is likely due to how duration and consistency of use are accounted for in measures of average effectiveness. For example, in 2004/5 approximately one-third of in-union women using contraception had used their method for less than six months, while the TFR is measured over a three-year period. Bongaarts’ model was applied to hypothetical future scenarios. Results suggest meeting the existing unmet need could result in TFRs as low as 2.5 in urban areas and 4.0 in rural areas, and meeting the Government’s target of 60% CPR would bring fertility near replacement levels. 5. Knowledge contribution The study has shown that national level measures of CPR and TRF mask important, policy-relevant differential patterns. Specifically, sub-national analysis reveals large and growing differences in these two measures between urban and rural areas, suggesting the importance of focusing programmes and resources to underserved rural communities. The study has also highlight the need for additional research to better understand how consistency and duration of use area factored into measures of method effectiveness. Unmet need remains high in Tanzania (22% in 2004/5), and despite large increases in contraceptive use in the early to mid-1990s, the rate of increase has slowed considerably. The study results provide additional support for efforts to revitalize FP efforts in Tanzania; such as the Government’s National Family Planning Costed Implementation Program (NFPCIP). This is not only vital to respecting individual women’s fertility preferences, but also at a national level will result in slowed population growth and improvements in the development prospects of the country.