Determinants of contraceptive continuation in rural Bangladesh
Akhter, H.H.; Ahmed, S.
Journal of Biosocial Science 24(2): 261-268
1992
ISSN/ISBN: 0021-9320 PMID: 1583038 DOI: 10.1017/s0021932000019775Document Number: 314408
In 1988, the proportional hazards model method was used to analyze data on 2717 women from rural areas to examine determinants of contraceptive continuation in Bangladesh. The major determinants included total number of living children and total number of dead children (p.001) when all contraceptives used (oral contraceptives (OCs), IUD, condom, and rhythm/withdrawal) were combined. When individual contraceptives were examined however, living children was not significant for condoms. Children dead significant only for OCs but associated with all other methods. The coefficient value for son desired was higher than additional children desired for all methods combined and for OCs and IUDs (.1689 vs. .0472; .1113 vs.-.0232; and 1.5671 vs. -0.4019, respectively). This showed a higher desire for sons than daughters, but it was not significant, although it was significant in the 1985 Contraceptive Prevalence Survey. This may be because the 1988 study excluded women with no living children. Thus the desire for more children may affect contraceptive use more strongly among women of low parity. The desire for more children was significant only for condom users (p.05). Desire for more children was stronger than the desire for a son among condom and traditional/withdrawal users, suggesting that these users may be in an early phase of family formation when son preference is not yet crucial. Distance of hospital from a woman's home was significant only for condom users (p.05). Women's age and education of both spouses were weak predictors of contraception continuation.