A Joint Spatially Correlated Survival Model for Timing of Childbirth and Marriage in South Africa
Samuel O. Manda
Biostatistics Unit, South African Medical Research Council, Pretoria, South Africa

In the sub-Saharan region of Africa, an understanding of the factors affecting the timing of first childbirth and marriage may help in reducing maternal mortality ratios and in increasing universal access to reproductive health; both of which are targets under the “improve maternal health” goal, one of the eight Millennium Development Goals

We used data from the South African Demographic and Health Survey 1998 to investigate socioeconomic and demographic differences in the timing of childbirth and marriage among women aged 15 to 49 years. We employed proportional hazards models, which were modified to include both spatially uncorrelated and correlated district frailty random effects for each milestone. Then model was expanded to accommodate the two sets of milestones simultaneously using, where similarities and differences in the risk factors and spatial effects were modelled using a shared spatial component model.

Keywords: Maternal and child health; South Africa; Spatial survival models

Biography: Samuel Manda is a senior specialist biostatistician in the Biostatistics Unit at the South African Medical Research Council. He has previously been at the Universities of Waikato and Auckland in New Zealand, and Leeds in the United Kingdom. His research interests and expertise are in the state-of-the art Bayesian methodology within the modeling and analysis of complex health survey data, spatial epidemiology, bioinformatics and health service research. He has published widely in international journals in these areas. He is consulted locally and internationally on methodological issues in medical and epidemiological statistics.