Child mortality is a crucial indicator of the Sustainable Development Goals (SDGs), reflecting the objective of good health and well-being (Abreha et al., 2024; Xie et al., 2024). Previously, the Millenium Development Goals (MDGs) had focused on reducing child mortality. These goals were later superseded by the Sustainable Development Goals (SDG), which stressed the importance of child mortality in terms of both local and global concerns targeting country-specific needs (Bryce et al., 2013; Campbell, 2017; Lomazzi et al., 2014; Nepal, 2018). In recent decades, recognising the significance of a child’s first five years, crucial strides have been made towards reducing preventable child deaths. The collective contributions of governments, organisations, communities, healthcare practitioners and families globally have yielded a remarkable 62% decline in under-five mortality from 1990 to 2022 (UNICEF et al., 2024). While laudable progress has been achieved, the sobering reality of nearly five million under-five deaths in 2022 underscores the persistent threat (UNICEF et al., 2024). The rate of under-five mortality improvement also fluctuates across time frames and geographical regions (UNICEF et al., 2024). For instance, the under-five mortality reduction during the initial phase of the SDG era (2015–2022) was slower than in the preceding MDG era (2000–2015). Sub-Saharan Africa and Southern Asia jointly accounted for almost 80% of under-five deaths in 2022, with Southern Asia contributing to 26% of the global total. Within the South Asian context, Nepal reported 27 under-five deaths per 1000 live births in 2022 (UNICEF et al., 2024). The recent Nepal Demographic and Health Survey (NPDHS) unveiled a notable 72% decrease in under-five mortality from 1996 to 2022 (Ministry of Health and Population and ICF, 2023). Nonetheless, Nepal remains distant from attaining the SDG benchmark of 20 deaths per 1000 live births (Khatiwada et al., 2024). Furthermore, these fatalities are profoundly tragic, given that the majority could have been averted. Although how to prevent child death remains a burning question in low- and middle-income countries, the progress made on this issue has been quite remarkable relative to its starting point (Bryce et al., 2013; Liu et al., 2012). It is evident that efforts and steps that have been taken to address the issue of under-five mortality have gradually resulted in a global decrease in the child mortality rate (Lozano et al., 2011; Wang et al., 2014). Child mortality is considered one of the best indicators of a country’s economic development, and several studies have shown how child mortality and economic development are interconnected (O’Hare et al., 2013; Preston, 1975). Hence, there is a pressing need for further research in the field.

Despite substantial evidence of the familial role in reducing under-five mortality, this aspect is often overlooked in Nepal. With household dynamics changing in recent years, it has become increasingly important to understand how a household can impact a child’s well-being (Kayastha and Shreshtha, 2003). The present study examined the heterogeneity in the under-five mortality risk in Nepal, focusing on the influence of household generations and household structure from 1970 to 2022. Interestingly, our research findings revealed that the nuclear family structure remained stable in Nepal over the study period. This observation contrasts with the results of other region-based studies done in Nepal, which showed an increase in nuclear families (Ghimire and Axinn, 2006; Rai, 2023). The findings uncovered significant disparities in under-five mortality risk based on household dynamics, highlighting the critical impact of household characteristics on child health outcomes. Additionally, factors contributing to the mortality risk heterogeneity across different generations of households and household structures were explored. The results of this study are summarised as follows:

The study’s findings revealed that the risk of under-five mortality was higher in one- or two-generation households. There is ample evidence showing that extended households can provide substantial caregiving and support for children, positively affecting their well-being (Kumar and Ram, 2013; Menashe-Oren et al., 2023; Zewudie et al., 2020). Extended families can offer care not only while a child is growing (through immunisation, daily care, etc.), but also during pregnancy and delivery (Kumar and Ram, 2013). Such social support is often lacking in nuclear households. Additionally, households with more members can contribute financially, helping families with income deficits to provide better resources and proper nutrition for children (Griffiths et al., 2002; Kumar and Ram, 2013; Treleaven, 2023). However, these scenarios may change over time. Living with dependent extended family members can sometimes burden the household, and parental authority over child health decisions may be hindered (Kumar and Ram, 2013). Moreover, some studies have found that living in extended households increases the risk of child mortality (Akinyemi et al., 2016). There is also evidence that women living in nuclear households have greater autonomy than women living in extended households, where resource allocation seems to be constraining (Matsumura and Gubhaju, 2001).

The contributions of mother’s education and mother’s age at childbirth to under-five mortality in Nepal were found to be significant. Our results showed that with decreasing levels of mother’s education and age at birth, the incidence of under-five mortality increased. These findings are in line with several other studies (Ghimire et al., 2019; Pandey et al., 2013; Pandey et al., 2017; Sarkodie, 2021; Singh and Tripathi, 2013; Sohail and Neupane, 2019). Having more education can increase mothers’ awareness of better pregnancy practices, knowledge and access to available healthcare facilities. Our findings were similar to the results to several other studies showing that children born to mothers and fathers with no education or primary-level education are at greater risk of under-five mortality (Akter et al., 2015; Ghimire et al., 2019; Khadka et al., 2015). Similarly, we found that father’s education significantly influenced under-five mortality, consistent with other studies (Dendup et al., 2018; Goli et al., 2017; Khadka et al., 2015; Sohail and Neupane, 2019). The decomposition analysis showed that mother’s age at birth narrowed the gap in under-five mortality. In recent decades, there have been several socio-economic transformations leading to improvements in women’s education and empowerment, which have, in turn, significantly affected women’s age at marriage. Parents are not only the primary caregivers of children, but are also the main providers of economic, social and emotional support to them. Hence, the age, education and income of parents are important factors in children’s growth and development.

The study found that shorter birth intervals were associated with higher mortality risk. This finding is consistent with other studies showing a significant association between birth interval and infant mortality (Dadi, 2015; Houle et al., 2013; Kozuki and Walker, 2013; Rutstein, 2005). There is ample evidence indicating that a short birth interval might hamper a mother’s physical and psychological recovery due to a lack of proper nutrition and sleep, and due to several underlying conditions, which can arise when a woman gives birth again before fully recovering from the previous birth. Moreover, if two children are spaced closely, issues regarding the distribution food, healthcare and time in the family may arise, leading to adverse child health outcomes. Additionally, socio-economic background and physical health are crucial factors in the risk of under-five mortality. The multivariate decomposition analyses revealed that the preceding birth interval narrowed the gap in under-five mortality in both cases. It is reasonable to assume that long birth intervals significantly reduce the chances of child mortality. Similarly, multiple childbirths might be associated with short birth intervals, leading to increased child mortality. The results clearly showed that the third-, fourth- and last-born children were more prone to under-five mortality, which is in line with previous findings (Mishra et al., 2018; Modin, 2002; Paul et al., 2021b; Srivastava et al., 2021). However, some studies have found that first-born children are more likely to experience under-five mortality, as younger mothers may lack parenting experience and knowledge (Siervo et al., 2010).

Our study found that the birth cohort of the child was significantly associated with under-five mortality in Nepal, with older birth cohorts facing a greater risk of under-five mortality. This finding is in line with several other studies showing improvements in child mortality in Nepal over the decades (Pradhan et al., 2012; Sreeramareddy et al., 2013; Thapa, 2008). A potential explanation for this observation is that in recent decades, efforts to expand female education, increase awareness, ensure access to healthcare and skilled obstetric services, improve healthcare sectors and increase political commitments to ensuring newborn survival in Nepal have reduced the risk of child mortality (Bishai et al., 2016; Pradhan et al., 2012).

The household wealth quintile was found to be significant in determining under-five mortality in Nepal. The results of the decomposition analysis showed that the household wealth quintile played a role in widening the under-five mortality gap between the groups varying by generations and household structure. These results are consistent with other studies showing that child mortality is much higher among the poorest quintile households (Boettiger et al., 2021; Chao et al., 2018; Fagbamigbe et al., 2022). Wealth is considered one of the most vital determinants of good health (Semyonov et al., 2013). The household wealth quintile has several aspects. Generally, parents from poor households have less financial security, education and awareness than their rich counterparts. Hence, children born into wealthier households have better access to food, sanitation, healthcare facilities, etc., than their poor counterparts, which decreases their chances of experiencing child mortality. In short, poor households find it difficult to afford basic needs, resulting in health shocks among children (Boyle et al., 2006; Lartey et al., 2016).

The present study showed that mortality among children under age five was lower in Nepal’s hill and Terai regions than in the mountain region. This finding is also in line with several other studies showing that child mortality in Nepal is significantly higher in the mountain areas (Ghimire et al., 2019; Goli et al., 2017; Pandey et al., 2013; Sreeramareddy et al., 2013). The decomposition analysis showed the role of the ecological region in widening the gap in under-five mortality by generations in households and household structure. One potential explanation for this gap is that mountain areas have higher concentrations of poor populations, increasing the chances of that the residents lack access to proper transportation and healthcare centres, which could, in turn, lead to elevated infant mortality (National Planning Commission, 2016). The people living in mountain regions are less willing to pay for advanced medical care because most of them are disadvantaged ethnic populations (Nawal and Goli, 2013). However, recent improvements in communication and transportation infrastructure in these regions have helped to narrow the gap in under-five mortality. This study also found that under-five deaths were higher in rural areas than in urban areas. This finding is also in line with some previous studies on child mortality (Goli et al., 2017; Khadka et al., 2015; Mishra et al., 2018; Sreeramareddy et al., 2013). Compared to rural areas, urban areas tend to have better healthcare facilities, better communication and transportation infrastructure, greater awareness, and, most importantly, more healthcare centres, which may, in turn, lead to lower child mortality.

A major share of Nepal’s population is comprised of migrants from Tibet or of Indo-Aryan heritage (Basra et al., 2024), and almost 124 languages and dialects are spoken in Nepal (Regmi, 2024), making it a socio-culturally diverse and multi-lingual country. Our study revealed that under-five mortality was higher among populations that spoke Maithili, Bhojpuri or other languages than among the Nepali-speaking population. This finding is in line with a recent report on maternal services showing that Maithili- and Bhojpuri-speaking women have a lower rate of institutional delivery (Khatri et al., 2024). One plausible explanation for this gap is that people speaking any language other than the native Nepali language face language barriers in accessing healthcare facilities. Moreover, cultural upbringing plays a vital role in care-seeking attitudes.

This study has a few limitations. First, the household structure may change over the birth history of mothers, and the DHS datasets provide information on household type at the time of the survey, and not during each child’s birth. It is possible that the household composition may have changed after the death of a child or before the time of the survey. To address this issue, we performed sensitivity analysis by comparing model estimates for the recent sample of children born within five years of the survey date with model estimates for the full sample. The findings revealed that the model estimates were similar across the full and recent samples, thereby documenting the robustness of our findings. However, this is a standard limitation in demographic studies, and future research using datasets with longitudinal household data could address this issue. This issue has been mentioned in a similar study that examined the association of household structure with under-five mortality, suggesting that longitudinal data might be used to consider the de facto household structure at the time of death rather than at the time of the survey, which would be a more accurate way to study this association (Menashe-Oren et al., 2023). Second, the study could not account for factors such as the absence of household members due to migration, as this information was unavailable in the dataset. Finally, child mortality is shaped by complex socio-economic interactions, and an intersectional approach could further enrich our understanding of these dynamics in future research.

The present study aimed to show the gap in under-five mortality by generations in the household and household structure. The study found that the wealth quintile of the household, mother’s age at birth, mother’s educational level, the preceding birth interval and the ecological region were significantly associated with Nepal’s under-five mortality, and considered their role in widening the under-five mortality gap between the groups. Several studies have shown how crucial a mother’s education is to a child’s overall development. One study found that almost 26% of married women of reproductive age in Nepal have an unmet family planning need, making this a pressing concern for the country (Ghimire et al., 2019). Laws must be enacted that consider both men’s and women’s minimum age at marriage and women’s education. Women’s education is an essential issue to focus on because it will benefit both mothers and their children by increasing awareness, knowledge and practices regarding pregnancy. Similarly, a household’s wealth plays an important role in child development as it affects access to education, healthcare and other rudimentary requirements. Moreover, healthcare facilities must be improved in rural areas of Nepal at the grassroots level. Finally, there is a need for more exploration of why under-five mortality varies by ecological region.