After reviewing titles and abstracts, 458 articles were eligible for potential inclusion. Upon full-text screening, articles which were not related to child mortality; included children over 5 years old; were not in SSA; or did not describe urban or rural settings, were excluded. The final search output included 21 articles. The process of data collection and selection is illustrated Fig. 2.

Fig. 2

PRISMA flow diagram outlining the process of data identification, screening, and selection according to inclusion and exclusion criteria

This review encompassed 37 SSA countries: Angola, Cameroon, Benin, Congo Democratic Republic, Burkina Faso, Burundi, Eritrea, Chad, Comoros, Nigeria, Congo, Cote d’Ivoire, Eswatini, Ethiopia, Gabon, Gambia, Liberia, Ghana, Togo, Uganda, Guinea, Kenya, Lesotho, Madagascar, Morocco, Mozambique, Namibia, Malawi, Mali, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Sao Tome and Principe, South Africa, Tanzania, Zambia, Zimbabwe. Characteristics of these papers are summarized in Table 2.

Table 2 Characteristics of papers included in this reviewOperationalization of Urbanicity and Rurality

This review stratified findings based on how urbanicity was operationalized in the included studies, as summarized in appendix C in supplementary material. Three categories emerged: (1) binary urban–rural classification (appendix Ci in supplementary material), (2) stratified urban areas (formal, informal/slum, rural) (appendix Cii in supplementary material)), and (3) continuous urbanicity using spatial metrics or satellite imagery (appendix Ciii in supplementary material).

Urbanicity as a Binary Dichotomy

Studies using a dichotomous urban–rural framework reveal substantial inconsistencies, with no clear urban advantage or penalty. In Tanzania, neonates in urban areas faced higher mortality risks than rural counterparts (OR = 1.94; p = 0.006) despite greater service coverage, suggesting a growing urban penalty [8]. Conversely, across 35 SSA countries, pro-rural inequalities in U5MR were observed in 16 countries, pro-urban in 2, and no significant difference in 17 others [5]. In Malawi and Tanzania, urban–rural differences fluctuated, with declining urban advantage or reversal to urban penalty across morbidity indicators [7, 18]. Across central Africa, nearly all rural–urban inequity in U5MR was explained by differences in covariate distribution rather than effects [19]. These contradictions highlight the limitations of the urban–rural dichotomy, masking underlying intra-urban disparities and dynamic nature of health determinants.

Urbanicity as Stratified (Formal, Informal/Slum, Rural)

Stratified analyses reveal that children in informal urban settings consistently fare worse than their formal urban counterparts. While comparing urban informal settings to rural areas yielded mixed results, intra-urban inequities consistently exceeded urban–rural differences. In Kenya, urban non-slum (OR: 1.12, 95% CI 1.08–1.16) and slum areas (OR: 1.49, 95% CI 1.41–1.57) had higher child mortality than rural areas [20]. Fotso and colleagues demonstrated faster declines in U5MR in Kenya (rural = 22%, urban = 14%, slum = 39%) [21]. Kimani-Murage and colleagues explain the reversing urban penalty in Kenya through a narrowing urban–rural gap, with a more significant decline in rural areas [12]. Despite this, U5MR in urban slums remain highest (104/1000) relative to rural (72.5/1000) and non-slum urban (73/1000) areas [12]. Gunther and Harttgen found that in Ethiopia considering urban populations as homogenous masks large intra-urban inequalities [13]. Child mortality rates in urban slums were almost three times higher (125.9/1000) compared to formal urban areas (51.1/1000) in Ethiopia [13], with intra-urban differences exceeding urban–rural differences; U5MR was 65% higher in urban slums than formal urban areas, while rural U5MR is only 16% higher than formal urban areas. However, rural child mortality was higher than in urban areas overall, even when adjusting for urban slums. This indicates that while slums pose a higher health risk, urban access to basic infrastructure is still generally better than rural areas.

Urbanicity as a Continuous Concept

Studies operationalizing urbanicity as continuous leveraged composite indices, spatial metrics or satellite-derived spatial classifications​ to capture gradients in development and infrastructure. These approaches reveal that the relationship between urbanicity and child health is neither linear not uniformly protective. Across 26 SSA countries, urban children were 10% less likely to die than rural (OR = 0.9; CI 0.87–0.94), but this advantage varied by settlement type: compared to rural villages, mortality was lower in rural towns (OR = 0.94), urban clusters with deprivation (OR = 0.91), and urban centers with deprivation (OR = 0.84). However, children in deprived urban areas still face high health burdens, with only a weak mortality advantage (OR = 0.94; CI 0.88–1.01), while non-deprived urban clusters (OR = 0.85) and centers (OR = 0.83) show better outcomes [22]. In Tanzania, core urban areas exhibited higher NMRs (39.8/1000 live births) compared to peri-urban (24.8/1000 live births) and rural (21.9/1000) zones [4]. This highlights that urban advantage is often diluted or reversed in settings with infrastructural deficits or socioeconomic deprivation, suggesting that urbanicity alone does not ensure better child health. Instead, spatial gradients in service access and living conditions may more accurately explain health disparities across the urban continuum.

Determinants Associated with Child Mortality in Urban and Rural Settings

The construction of rurality and urbanicity were inconsistent across all studies, with no clear cross-country nor urban–rural uniformity in factors underlying child mortality. Overall, the determinants were similar but varied in strength across papers in rural and urban areas. Based on the findings of included studies, four categories emerged: environmental, healthcare, sociodemographic, and disease and morbidity related. Appendix summarizes the findings across each of these four dimensions and their effect on the urban–rural gap.

Environmental Factors

Access to basic amenities like electricity, clean water, and sanitation is a significant marker of urbanicity. In West SSA, urban areas have greater access to these amenities (70%) than rural areas (20%), with these amenities and U5MR declining as city size decreased (p < 0.01) [23]. Having significantly reduced the probability of infant death (β: 0.13, p < 0.01) [19]. Over 70% of urban households in Kenya, both formal and informal, had improved water and sanitation, compared to only half of rural households [20]. However, Fotso and colleagues noted that rapid urban population growth is negatively correlated with access to safe drinking water (r = − 0.42; p = 0.07), increasing child mortality [21]. Across 22 SSA countries, those with most improved safe drinking water in urban areas had largest declines in U5MR, emphasizing the protective effect of amenities.

Housing quality also plays a critical role in child mortality related to urbanicity. There are higher risks of infant death in both urban non-slum (OR: 1.12, 95% CI: 1.08–1.16) and slum areas (OR: 1.49, 95% CI: 1.41–1.57) compared to rural areas. Improved housing was generally protective, especially in rural areas, where durable housing correlated with an 18% reduction in U5MR (OR: 0.82, 95% CI: 0.76–0.88) [20]. Conversely, in urban settings, durable housing increased child mortality risk by 2% (OR: 1.24, 95% CI: 1.04–1.49), perhaps reflecting other adverse factors in better-built high-risk areas. Similarly, Van de Poel and colleagues identified higher U5MR in premises with unfinished floors in urban areas (β: 0.13, p < 0.01), compared to no significant mortality effect in rural areas, where most dwellings have unfinished flooring [19]. This possibly represents slum dwellings and poor public health conditions.

Healthcare-Related Factors

Some authors explored access to and use of healthcare as well as health behaviors to construct urbanicity and rurality. Healthcare-related factors explored across this review include child immunization coverage, travel time and distance to the nearest health facility, ANC visits, birth by caesarean section, use of SBA, and facility delivery.

Vaccination coverage is important in explaining child mortality disparities, with countries exhibiting high levels of child vaccination coverage displaying more rapid declines in U5MR [21]. Across 23 SSA countries, urban children show higher full immunization coverage (FIC) (52.8%) than rural (40.7%, p < 0.001), often linked to socioeconomic disparities [24]. Richest wealth status was the most significant contributor (35.7%) to the urban–rural gap (p < 0.001), with 8% of the gap explained by distance to health facilities (p < 0.001) [24]. Obanewa and Newell echoed these findings in Nigeria, showing higher FIC in urban areas (69% higher in urban formal settings and 45% higher in slums) [25]. Place of delivery, maternal ANC attendance, and maternal education were significant across all settings, with maternal ANC attendance demonstrating the largest effect in rural areas (Rural: OR 8.37, 95% CI 5.34–13.12; Urban: OR 6.82, 95% CI 2.29–20.34; Slum: OR 8.07, 95% CI 2.15–30.25). In urban areas, maternal education had the largest effect (Rural: OR 4.99, 95% CI 2.48–10.06; Urban: OR 9.18, 95% CI 3.05–27.64; Slum: OR 5.03, 95% CI 1.52–16.65) [25]. Across 22 SSA countries, urbanization and urban population growth correlated with reduced child FIC (r = − 0.57, p = 0.01) [21]. Considering intra-urban inequities in FIC revealed that the general decline in FIC was more pronounced in urban slums (− 13%) than in urban areas as a whole (− 9%) in Zambia [21]. The same is exhibited in Kenya, whereby urban slum children have the lowest FIC rates in urban areas (82%) respective to rural areas (42%) [18]. Across both urban and rural areas, socioeconomic status was found to be positively correlated with facility delivery, with 90% and 33% in the highest and lowest social quintiles respectively delivering in a healthcare facility. This has also been exhibited in Tanzania (Rural: 54.4%, Urban: 88.3%) [26] and Ghana (adjusted OR: 1.59, 95% CI: 1. 07–2.37, p = 0.02) [27]. The determinants of using SBA were similar but of different strengths in rural and urban areas. Residing less than 4 km from a health facility was the greatest independent contributor to the variance in SBA in the urban areas (Rural: OR 6.05 95% CI 2.54–14.39, p < 0.001; Urban: OR 9.62, 95% CI 3.78–24.49, p < 0.001), whereas in rural areas, frequency of ANC attendance was the greatest independent contributor (Rural: OR 34.86, 95% CI 12.74–95.38, p < 0.001; Urban: OR 3.93, 95% CI 2.21–6.86, p < 0.001). Maternal education was more significant in rural areas, of 37.6% increased SBA use compared to 5% in urban areas (p = 0.001) [27].

Attendance of less than the WHO-recommended four ANC visits has been associated with higher NMRs across SSA, and thus is used as a proxy for exploring child mortality (28; 29). There is a consistently higher coverage of ANC in urban respective to rural areas. Adewuyi and colleagues found that in Nigeria, more rural women (61.1%, 95% CI: 58.5–63.5) underutilized ANC compared to their urban counterparts (22.4%, 95% CI: 20.0–25.1) [28]. Determinants were similar across urban and rural areas, but the strength of their effects differed. In rural areas, wealth index had a stronger effect (adjusted OR 2.17, 95% CI 1.68–2.81) than in urban areas (adjusted OR 2.05, 95% CI 1.51–2.79). Whereas in urban areas, lack of paternal education had the strongest effect (Rural: adjusted OR 2.03, 95% CI 1.72–2.43; Urban: adjusted OR 2.16, 95% CI 1.68–2.75).

Caesarean delivery has been correlated with increased NMRs across SSA [29] and is used as a proxy for child mortality in this review. The prevalence of caesarean delivery was higher in urban areas (10.37%, 95% CI: 8.99–11.75) than in rural areas (3.78%, 95% CI:3.17–4.39) across 28 SSA countries [30]. Approximately 81% of the rural–urban disparities in caesarean deliveries were attributable to the differences in child and maternal characteristics. Wealth index was the largest contributor to explaining the rural–urban disparity in caesarean deliveries. The likelihood of a caesarean section increased with wealth index in both urban (OR: 2.83, 95% CI: 2.11–3.80) and rural areas (OR = 2.58, 95% CI: 2.17–3.07). However, the odds were more significant in urban areas. Compared to women who had no ANC, those who had four or more ANC visits were more likely to deliver through caesarean delivery, with higher odds in rural areas (OR: 4.49, 95% CI: 3.42–5.89) compared to urban (OR: 2.71, 95% CI: 1.80–4.11).

Travel time to the nearest hospital is key in constructing urbanicity. In Tanzania, while core urban areas had the shortest average travel time to hospitals (4 min core urban; 41 min semi-urban; 89 min rural), NMRs were paradoxically higher in core urban areas (39.8/1000 live births, 95% CI: 26.3–59.9), compared to semi-urban (24.8/1000 live births, 95% CI: 19.6–31.4) and rural (21.9/1000 live births, 95% CI:16.8–28.5) areas, which displayed similar NMRs (p = 0.03) [4]. This aligns with Yelverton and colleagues who observed urban children lived closer to health facilities (1.9 km) than their rural (5 km) counterparts (p < 0.001) [31]. Lungu and colleagues found that in Malawi, even in the context of geographical proximity to healthcare services, those in urban slums may not seek healthcare, let alone of a timely nature; 61% of caregivers sought healthcare, while 53% sought it late [7].

Sociodemographic Factors

Sociodemographic factors are difficult to disentangle from other child mortality indicators outlines as they are often confounders or effect modifiers, rendering their independent influence on urbanicity and child mortality complex to assess. However, some papers included in this review examine their effects independently, specifically socioeconomic status (SES) and maternal education, revealing nuanced insights. SES emerges as the most significant determinant in whether a country experiences an urban or rural penalty for child mortality [5, 21]. Despite maintaining more favorable sociodemographic profiles, NMRs paradoxically remained highest in urban areas (OR: 1.94, p = 0.006) [8, 31]. In Zambia, urban poor children were 46% more likely to die than their poorest rural counterparts, suggesting that other unfavorable circumstances in urban settings offset socioeconomic advantage, highlighting the challenges in isolating its impact on child mortality along the urban–rural continuum [21]. Maternal education was significant in reducing child mortality across urban and rural areas; however, the effects of this were most pronounced in rural areas (Rural: OR 0.17, Urban: OR 0.06) [19]. Gruebner and colleagues identified a 22% reduction in the risk of infant death for educated mothers in urban areas (OR: 0.84, 95% CI: 0.73–0.96) compared to rural (OR:0.78) in Kenya, highlighting the larger effect in rural areas [20].

Child Disease and Morbidity Indicators

Child morbidity indicators and disease incidence demonstrating urban–rural disparities that are used to construct urbanicity include infections (diarrhoeal diseases and ARI) and childhood malnutrition and stunting. When considering urban and rural areas as dichotomous, there are conflicting findings regarding which has highest rates of infection and likelihood to seek and receive treatment. Ekholuenetale and colleagues found a higher prevalence of ARI among urban residents across 23 SSA countries, coupled with an increased likelihood to seek and receive treatment [32]. In Malawi, Lungu and colleagues had consistent findings of a declining urban advantage with respect to diarrhoeal and ARI rates yet established an urban penalty for access to and use of treatment services (Urban: 59.1%, Rural: 67.7%) [7]. Kimani-Murage and colleagues also found that those in rural areas were more likely to seek care for childhood ARIs respective to their urban counterparts (OR:1.4) [12]. However, when intra-urban inequities are accounted for there are more consistent findings. Mberu and colleagues found that slum children in Nairobi fared worse than children in rural areas of Kenya across child morbidity and health service indicators, with urban slums exhibiting a U5MR of 3.6 times higher than the rest of Nairobi as a whole [33]. Slum children are more likely to have illnesses such as diarrhea and ARI (Rural: 9.1%, Urban: 7.3% Slum: 24.6%) and simultaneously less likely to receive treatment (Rural: 58.1%, Urban: 56.7%, Slum: 42.7%) compared to their rural and urban counterparts. Shon echoes these findings whereby across 26 SSA countries, children in deprived urban settlements are 24% (OR = 1.24; CI 1.14–1.34) more likely to experience diarrhea compared to children in rural villages (OR = 0.96; CI 0.92–1) [22].

There are cross-country disparities in child stunting and malnutrition across urban and rural settings. In Tanzania, more children are stunted in rural areas (45%) compared to urban areas (35%), and underweight (Rural: 17%, Urban: 11%), with the urban–rural disparity widening due to slower rural decline (18; 26). Children from households in the lowest wealth quintiles accounted for a larger proportion of stunted children across both urban and rural areas. No interaction effect existed between residence and other determinants, and the urban–rural disparity was mainly caused by the discrepancy of the individual and household-level factors between rural and urban households [26].