Regional variation in early marriage in Kenya

Résumé

Background: This study seeks to the establish the association between region and early marriage and whether or not the association has changed over time. Method: The study used a national representative sample of 21315 ever married women aged 20-49 years drawn from the 2014 Kenya Demographic and Health Survey. Descriptive statistics and logistic regression analysis were used to analyse the data. Results: The study found early marriage prevalence of 32.4% [95% CI: 31.81 – 33.06] and considerable regional variation. At the national level, early marriage was significantly associated with region [Rift Valley: aOR =1.22, 95%; 1.07 – 1.39, p value = 0.000; Western: aOR=1.31, 95%; 1.12 – 1.52, p value = 0.000; Nyanza: aOR =2.08, 95%; 1.80 – 2.40, p value = 0.000; North Eastern: aOR=0.63, 95%; 0.45 – 0.82, p value = 0.05; Eastern: aOR= 0.84, 95%; 0.72 – 0.97, p value = 0.05], education [primary education: aOR= 0.63, 95% CI ; 0.56 -0.70; secondary education: aOR= 0.201, 95% CI; 0.18- 0.24, higher education: aOR= 0.05, 95% CI; 0.04 - 0.06], household wealth status [poor: aOR= 1.31, 95% CI; 1.20 – 1.44; middle level: aOR=1.24, CI; 1.13 – 1.36], religion [Muslim: aOR= 1.36, 95% CI; 1.14 – 1.61, other non-Christian faith: aOR= 1.28, 95% CI; 1.03 -  1.59], premarital sex [ yes: aOR= 1.91, 95% CI; 1.78 – 2.05] , pregnancy [yes: aOR= 1.44, 95% CI; 1.24 – 1.68] and age cohort [30-39: aOR = 0.71, 95% CI; 0.66 – 0.77, 40-49: aOR= 0.80, 95% ; 0.74 – 0.87]. The study also established region specific drivers of early marriage. Lack of education, premarital sex and premarital pregnancy were key drivers in all the regions. Poverty and religion were drivers in the majority of the regions.