Birth spacing and child survival in rural Senegal
Ronsmans, C.
International Journal of Epidemiology 25(5): 989-997
1996
ISSN/ISBN: 0300-5771 PMID: 8921485 DOI: 10.1093/ije/25.5.989Document Number: 377990
This study 1) describes the distribution of birth intervals, 2) explores the extent to which child mortality reduces the birth interval, and 3) examines the association between a short birth interval and neonatal and postneonatal mortality and mortality over the age of 1 year in the next birth. The study was conducted among a rural population living east of Dakar, Senegal. Annual data were obtained from the demographic surveillance system during 1983-87. Prolonged breast feeding was widespread. Four separate samples were selected from the 4852 single births that occurred during 1983-87 among the mothers in the 22 study villages. Standard life tables were used to construct mortality levels for the entire sample. Subsamples were used to establish the nonparametric Kaplan-Meier survival probability of a next birth, the average length of birth intervals, the effect of a short preceding birth interval on subsequent neonatal and infant mortality, and the effect of a short birth interval on subsequent mortality among children aged over 1 year. The probability of dying in the first year was 103.5/1000 live births, and 134.3/1000 among children aged 1-5 years. The median birth interval was 33 months. After a neonatal death, 50% of women had another birth within 15 months. The proportion with a short birth interval under 14 months declined if a child survived infancy. Most neonatal and postneonatal deaths occurred before the estimated date of the next conception. A large proportion of deaths among infants aged 1-2 years occurred after the date of next conception. Preceding birth intervals of 1-2 years did not affect neonatal and postneonatal mortality. Child mortality in the second year is attributed to abrupt weaning. Children had a 2.58 times higher risk of dying postneonatally during the rainy season. Mortality declined over the study period. Birth order, maternal age, sex, or literacy did not largely affect child mortality.