Change in the Prevalence and Determinants of Consanguineous Marriages in India Between National Family and Health Surveys of 1992-1993 and 2015-2016.

This study aimed to determine the changing prevalence of consanguineous marriage in India between two national-level surveys. The primary hypothesis was whether region of residence and religious affiliation continue to play a significant role in determining consanguineous marriage even after controlling other potentially significant confounding variables. Data from the 81,781 and 85,851 ever-married women during the National Family and Health Surveys (NFHS) survey periods 1992-1993 (NFHS-1) and 2015-2016 (NFHS-4), respectively, were used in the analysis. Multinomial and binary logistic regression analyses examined determinants of consanguineous marriage types and of paternal and maternal first-cousin marriages, respectively. In both analyses a systematic model-building procedure was adopted. Altogether, four models were estimated. In the final model (model 4) of both the analyses, all respondent background characteristics (region of residence, religious affiliation, sociodemographic, household wealth) and years of survey were included. Although the overall prevalence of consanguineous marriage in India declined significantly (16%), it was not uniform across respondent background characteristics. The northern region of India (154%) showed a significant increase in consanguineous marriage, whereas eastern (31%), central (2.3%), northeastern (40%), and southern (8%) regions showed a significant decline. Significant declines in consanguineous marriage were found for Hindus (16%) and Muslims (29%); for Muslims of eastern (48%), central (29%), western (31%), and southern (27%) regions; and for Hindus in the western region (37%). Relative risk ratios estimated using multinomial logistic regression models suggest those living in the southern region show 9.55 ( < 0.001), 5.96 ( < 0.001), and 38.16 ( < 0.001) times more likelihood in the prevalence of first-cousin, second-cousin, and uncle-niece marriages, respectively, compared to the northern region after controlling all other confounding variables. Muslims also showed 3.76 ( < 0.001) and 2.91 ( < 0.001) times more likelihood in first-cousin and second-cousin marriages, respectively, compared to Hindus. Adjusted odds ratios (AORs) estimated using binary logistic regression models suggest those living in southern and northeastern regions were 1.25 ( < 0.001) and 1.36 ( < 0.05) times more likely, respectively, to marry a maternal first cousin compared to the northern region. The AOR estimates also show that Muslims were 1.11 ( < 0.01) times more likely to marry a maternal first cousin compared to Hindus. The authors conclude that, despite significant development in the socioeconomic condition of India during the postglobalization era (beginning in 1992-1993), region of residence and religious affiliation continue to play significant role in determining consanguineous marriage.