Gender bias in under-five mortality in low/middle-income countries. Issue 2 (28th July 2017)

• Record Type: Journal Article
• Authors: Costa, Janaína Calu; da Silva, Inacio Crochemore Mohnsam; Victora, Cesar Gomes
• Is Part Of: BMJ global health. Volume 2:Issue 2(2017)
• Publication Date: 2017-07-28
• Subjects: Under-five mortality rate -- Gender bias -- Demographic and Health Surveys; World health -- Periodicals; 362.105
• DOI: 10.1136/bmjgh-2017-000350
• Languages: English
• ISSNs: 2059-7908

Abstract

Abstract : Introduction: Due to biological reasons, boys are more likely to die than girls. The detection of gender bias requires knowing the expected relation between male and female mortality rates at different levels of overall mortality, in the absence of discrimination. Our objective was to compare two approaches aimed at assessing excess female under-five mortality rate (U5MR) in low/middle-income countries. Methods: We compared the two approaches using data from 60 Demographic and Health Surveys (2005–2014). The prescriptive approach compares observed mortality rates with historical patterns in Western societies where gender discrimination was assumed to be low or absent. The descriptive approach is derived from global estimates of all countries with available data, including those affected by gender bias. Results: The prescriptive approach showed significant excess female U5MR in 20 countries, compared with only one country according to the descriptive approach. Nevertheless, both models showed similar country rankings. The 13 countries with the highest and the 10 countries with the lowest rankings were the same according to both approaches. Differences in excess female mortality among world regions were significant, but not among country income groups. Conclusion: Both methods are useful for monitoring time trends, detecting gender-based inequalities and identifying and addressing its causes. The prescriptive approach seems to be more sensitive in the identification of gender bias, but needs to be updated using data from populations with current-day structures of causes of death.