Introduction

In resource-poor settings, gender inequality remains a major driver of poor health outcomes for women and girls (World Health Organization, 2011; Derrick, Reference Derrick2014). Countries characterized by poor gender equity often experience higher rates of maternal morbidity and mortality (Filippi et al., Reference Filippi, Ronsmans, Campbell, Graham, Mills, Borghi and Osrin2006; Asling-Monemi et al., Reference Asling-Monemi, Naved and Persson2009), and higher rates of infant and child morbidity and mortality, particularly among girl children (Asling-Monemi et al., Reference Asling-Monemi, Naved and Persson2009; Brinda et al., Reference Brinda, Rajkumar and Enemark2015). The negative effects of gender inequity on female health are most pronounced in resource-poor settings in East and South Asia, the Middle East and multiple regions in Africa, where cultural norms supporting a preference for male children are strongest (Hesketh & Xing, Reference Hesketh and Xing2006a; Fuse, Reference Fuse2010). Such norms confer social and economic utility on male children and social and economic expense on female children. Families are faced with the choice between raising a son who is culturally integral to the family’s livelihood, lineage, inheritance and religion, or a daughter who is an additional financial responsibility on the family’s already constrained resources (e.g. dowry, low earning potential; Das Gupta et al., Reference Das Gupta, Zhenghua, Bohua, Zhenming, Chung and Hwa-Ok2003; Fikree & Pasha, Reference Fikree and Pasha2004; Milazzo, Reference Milazzo2014). Having male children in such environments is a greater-good obligation rather than mere preference (Brunson, Reference Brunson2010), driving a bias towards sons and creating the potential for negative health effects for girl children.

The effects of gender on female health are found in many cultural contexts. In the extreme form, gender bias favouring sons may result in female-fetal abortion and female infanticide (Fuse, Reference Fuse2010), and physical and/or medical neglect of female infants and children, examples of which can be seen from studies across low- and middle-income countries. In India, girls are breast-fed for a shorter period and are given less milk overall compared with boys (Fledderjohann et al., Reference Fledderjohann, Agrawal, Vellakkal, Basu, Campbell, Doyle and Stuckler2014), and are less likely to be immunized compared with boys (Corsi et al., Reference Corsi, Bassani, Kumar, Awasthi, Jotkar, Kaur and Jha2009). In Egypt, boys receive better treatment for diarrhoeal diseases compared with girls (Yount, Reference Yount2003), and in India, Bangladesh and Pakistan, girls receive less preventive care and less nutritious food compared with boys (Fikree & Pasha, Reference Fikree and Pasha2004). In Turkey, daughters are given less access to overall resources than sons (Altindag, Reference Altindag2016). Such practices contribute to malnutrition (Raj et al., Reference Raj, McDougal and Silverman2014), poorer growth patterns (Koolwal, Reference Koolwal2007), discrimination in school attendance and care practices (Hesketh & Xing, Reference Hesketh and Xing2006b) and increased mortality (Chamarbagwala, Reference Chamarbagwala2011; Dyson, Reference Dyson2012; Jayachandran, Reference Jayachandran2014). A recent trend towards a desire for ‘balanced families’, i.e. one in which there are similar numbers of boys and girls, is increasing in some resource-poor countries with low birth rates – such as those in Eastern Europe and the Caucuses (Garenne & Hohmann, Reference Garenne and Hohmann2014). This may have particular implications for same-sex twin pairs and has led to increases in sex-selective abortion in Armenia and Azerbaijan (Garenne & Hohmann, Reference Garenne and Hohmann2014).

Overall, females should have lower rates of first-year mortality (Naeye et al., Reference Naeye, Burt, Wright, Blanc and Tatter1971; Waldron, Reference Waldron1983; Asling-Monemi et al., Reference Asling-Monemi, Naved and Persson2009). X-chromosome-linked vulnerabilities more readily inherited by male children (Waldron, Reference Waldron1983) make them more susceptible to respiratory problems and low birth weight (Snow, Reference Snow2008). Higher levels of testosterone are also shown to depress infants’ immune systems (Ahmed & Talal, Reference Ahmed and Talal1990; Bouman et al., Reference Bouman, Jan Heineman and Faas2005). In settings where son preference is prevalent, cultural norms that reduce the desire for female children risk usurping the inherent biological advantage of female children.

Singleton births have provided researchers with the most practical means by which to study gender bias and associated health disparities during the prenatal, neonatal and postnatal periods, simply by being more common than multiple births. Smits and Monden found the average twinning rate across 76 resource-poor countries to be 1 in 76 (or 13 per 1000 births), and the average triplet rate to be 1 in 6700 (Smits & Monden, Reference Smits and Monden2011). There is little evidence for how, or if, the disadvantages associated with gender bias and son preference extend to multiple births – a population inherently at risk of a number of health challenges. Compared with singletons, twins in resource-poor settings are more likely to be born preterm and have a low birth weight, both of which present increased risk for adverse perinatal outcomes (Vogel et al., Reference Vogel, Torloni, Seuc, Betrán, Widmer, Souza and Merialdi2013), including mortality (Vogel et al., Reference Vogel, Torloni, Seuc, Betrán, Widmer, Souza and Merialdi2013; Miyahara et al., Reference Miyahara, Jasseh, Mackenzie, Bottomley, Hossain, Greenwood and Roca2016). A range of studies focused on twins in resource-poor settings have shown the negative health effects experienced by twins. Twins are more likely to be smaller in the post-neonatal period in Malawi (Kalanda et al., Reference Kalanda, Verhoeff and Brabin2006) and Ghana (Van der Mei et al., Reference Van der Mei, Heijstra and Boersma2003), and twins are more likely to be undernourished in Malawi (Kalanda et al., Reference Kalanda, Verhoeff and Brabin2006) and The Gambia (Jaffar et al., Reference Jaffar, Jepson, Leach and Greenwood1998). In Malawi, Tanzania and Zambia, twins under five years of age are at higher risk for mortality than singletons (Justesen & Kunst, Reference Justesen and Kunst2000).

Due to the distinct health challenges faced by female offspring and by twins in resource-poor settings, being both a female and a twin may result in adverse health outcomes that are much more severe than those seen in female singletons or male twins. Few studies have explored this issue directly, though Miyahara et al. (Reference Miyahara, Jasseh, Mackenzie, Bottomley, Hossain, Greenwood and Roca2016) found female twins in The Gambia to have a higher mortality risk than male twins (Miyahara et al., Reference Miyahara, Jasseh, Mackenzie, Bottomley, Hossain, Greenwood and Roca2016). Studies examining twins in resource-poor settings have not taken into account how the sex composition of twins is associated with morbidity and mortality outcomes. This study analysed how two key covariates – the sex composition of twins and societal gender inequity – are associated with eight mortality and morbidity outcomes in 38 resource-poor countries. This research fills gaps in the literature on the intersection between sex composition of twins and health in resource-poor settings. Understanding the effects of son preference on twins’ health outcomes will contribute towards the goal of reducing gender inequity, as detailed in Global Goal number five (World Health Organization, 2015).