Overweight/Obesity among children in the United States has increased dramatically over the last three decades (1, 2), rising from 5% in 1971–1974 to 13.4% in 2017–2018 among children aged 2–5 years (3). Previous research has found that the prevalence of childhood obesity by family income varies by ethnic group, and sex (4, 5). During 2017–2018, Obesity was most prevalent in Hispanic children aged 2–19 years (25.6%), followed by African Americans (24.2%), Caucasians (16.1%), and Asians (8.7%) (6), and was greater in low-income households (18.9%) than in high-income households (10.9%) (7, 8). Studies of National Health and Nutrition Examination Surveys (NHANES) data have used the concentration index to quantify socioeconomic inequalities in overweight/obesity among adolescents (5) and adults (9). Several authors have recommended the use of inequality measures to study social inequalities in health (10–12). Unlike the standard method (regression analysis), it has the advantage of using the entire population to overcome the effect of small sample sizes in some subgroups and allows comparisons over time (13, 14). Given the complexity of obesity (15), which has many contributing elements such as biological, behavioral, genetic, environmental, and developmental components (16), it is crucial to identify the presence and evolution of socioeconomic inequalities between ethnicity groups over time. Therefore, this study aimed to assess the levels and changes over time in socioeconomic inequalities of overweight/obesity among American children under the age of five by sex and ethnicity, using NHANES data from 2001–02 to 2017–18.

Analyses of pooled data from NHANES 2001–02 through 2017–18 (Table 1) revealed that children aged 2 years and older made up the majority of respondents in all ethnic groups; Males outnumbered females (50.9% and 49.1%, respectively). Pooled data covering 2001–2018 showed that Caucasians had a prevalence of 6.7%, African Americans had a prevalence of 8.6%, and other ethnicities had a prevalence of 8.0%.

Between 2001–02 and 2011–12, the prevalence of childhood overweight/obesity in the United States decreased from 7.3% to 6.3%, and had increased to 8.1% by 2017–18. However, this pattern varied widely by ethnicity and sex (Figure 1). From 2011–12 to 2017–18, the prevalence of overweight/obesity in males has steadily increased from 5.9% to 10.5%. While the prevalence of overweight/obesity among females began to rise in 2009–2010 (5.9%), stabilized between 2013–14 and 2015–16 (8.5%), and then decreased to 5.5% in 2017–18.

The greatest dispersions between the richest and poorest quintiles were observed among males of other ethnicities (Figure 2), with the poorest quintile becoming more overweight/obesity in recent years (from 10.0% to 8.5% between 2009–10 and 2017–18) and the richest quintile becoming less overweight/obesity (from 8.3% to 0.3% between 2009–10 and 2017–18). Between 2015–16 and 2017–18, the gap between the richest and poorest Caucasian females widened, with an increase in overweight/obesity among the poorest and a decrease among the richest (from 12.2% to 14.6% and 8.0% to 2.8%, respectively).

Figure 3 plots the SII and CIX results, along with their 95% confidence intervals. For both the 2015–16 and 2017–18 surveys, the SII and CIX were significantly lower than 0 among the overall Caucasian children ((SII = −11.83, IC 95% = −23.17, −0.49 and CIX = −7.368, IC 95% = −13.92, −0.82) and (SII = −11.52, IC 95% = −22.13, −0.91 and CIX = −7.24, IC 95% = −13.27, −1.21), respectively) and among males of other ethnicities [(SII = −13.93, IC 95% = −26.95, −0.92) and CIX = −8.55, IC 95% = −0.86, −16.25] and (SII = −21.19, IC 95% = −40.65, −1.74) and CIX = −13.11, IC 95% = −1.42, −24.80), respectively). Indicating a higher concentration of overweight/obesity in a quintile of the poorest households. With the exception of African American females in the 2013–14 survey, for whom SII and CIX were significant (SII = 12.60, 95% CI = 0.24, 24.97 and CIX = 7.86, 95% CI = 15.59, 0.12), SII and CIX were non-significantly larger than 0 in the overall African American children, indicating a higher concentration of overweight/obesity in a quintile of the richest households.

Annualized changes in relative wealth inequality are plotted in Figure 4. The annualized relative changes in the CIX among African Americans for both males and females are not significantly different from 0 in terms of wealth disparity. indicating a negligible difference between years. The CIX of wealth inequality for overweight/obesity showed annualized relative decreases between 2005–06 and 2007–08 among Caucasian males (CIX = −13.81, 95% CI = −4.30, −23.32), and between 2011–12 and 2013–14 among Caucasian females (CIX = −13.31, 95% CI = −1.11, −25.52), indicating an increase in wealth inequality. Between 2003–04 and 2005–2006, the CIX for overweight showed an annualized relative decrease among females of other ethnicities (CIX = −16.77, 95% CI = −5.94, −27.61), indicating an increase in wealth inequality. In contrast, a decrease in the CIX for males over the same period (CIX = −11.93, 95% CI = −1.20, −22.65) was followed by an increase between 2005–06 and 2007–2008 (CIX = 19.0, 95% CI = 30.14, 7.94), indicating a significant increase followed by a significant decrease in wealth inequality.

The concentration index (CIX) technique has been applied in the area of adolescent overweight in the 1999–2002 NHANES surveys (5). The present study focused on children under 5 years of age and used the CIX technique to examine trends in socioeconomic inequality of overweight/obesity from the 2001–02 to 2017–18 surveys. The findings of this study showed that the prevalence of childhood overweight/obesity in the United States declined between 2001–02 and 2011–12, then increased up until 2017–18. This pattern varied greatly by ethnic group and by sex. Wealth inequalities in overweight/obesity have emerged significantly in recent surveys, with the poorest children of Caucasians and other ethnicities bearing the brunt of the burden.

Our findings are consistent with previous research that found a significant decrease in the prevalence of overweight/obesity among children aged 2–5 years between 2003–04 and 2011–12 (1, 31) and reported that overweight/obesity varies by ethnic groups, with African Americans, American Indians, and Mexican Americans being more affected than non-Hispanic Whites (4, 32). There are hereditary risks for childhood obesity that have been described. When one or both parents are overweight/obese, the risk of having an overweight/obese child increases by 2–3 times or 15 times, respectively (33). The authors identified a number of factors that have contributed to the obesity epidemic among African Americans, including incorrect BMI classification, which may have led to 30%–60% of African American children being classified as overweight/obese (34). In addition, racial segregation (35, 36) and racism (37, 38) may prevent African Americans from escaping obesogenic factors associated with their environments (39).

Our findings are consistent with previous research suggesting that obesity is not always concentrated in low-income households (2, 5, 7, 40). Children from low-income Caucasian and Hispanic households were most likely to be overweight or obese. However, this relationship did not apply to African American children (41, 42). Although more African American children in the richest households were overweight or obese than in the poorest, the difference was only marginally significant (with the exception of 2013–14).

Our findings showed that among Caucasians and other ethnicities, wealth inequalities in childhood overweight/obesity emerged significantly in 2015–16 and 2017–18, which is consistent with previous research indicating that the strength of the relationship between income and the prevalence of overweight/obesity has increased over time (2); whereas other studies have found this relationship to be absent (5) or inconsistent across racial/ethnic groups (40).

The current study also found that the prevalence of overweight/obesity among male children of other ethnicities revealed persistent wealth inequalities between 2001–02 and 2017–18. These inequalities were due in part to a continual rise in the prevalence of overweight in the poorest group since 2009–10, while a continual decline had been seen in the richest group over the same time period. A stronger association between family income and childhood overweight was observed among Mexican-American children aged 2–5 years; however, the correlation did not differ significantly between males and females, independent of racial/ethnic groups (40). Another study found an inverse association among Mexican-American females only (43).

In the 2001–02 and 2003–04 surveys, overweight/obesity was concentrated among females of other ethnicities in the richest group, but the direction completely changed in 2005–06, with the poorest being the most affected. These findings are partially consistent with a previous study, which found that the direction of the relationship between income and overweight/obesity shifted for all children after 2004 (40).

Our findings are consistent with previous findings that overweight/obesity was more concentrated among African American males and females in the richest group. In contrast, our findings from 2001 to 2018 pooled data revealed that this relationship was significant only in African American females. It has been reported that obesity is increasingly linked to poverty, food insecurity (44) and the risk of being passed on to future generations. Authors have reported that childhood obesity is strongly related to the socioeconomic status of their parents, and that countries with the greatest wealth inequality have higher rates of childhood obesity (45). Mothers are less likely to breastfeed when they are obese and of low socioeconomic status. Thus, non-breastfed infants are more likely to have unhealthy eating habits, become obese and have delayed cognitive development by age 3 (45). On the other hand, evidence linking food insecurity and obesity is limited to adult women in developing countries (46).

The environment in which children and adolescents from low-income households grow may also contribute to their increased risk of being overweight or obese (7, 47, 48). Children from low-income or racial/ethnic minority families most frequently reside in underdeveloped areas with few stores (49), which restricts their access to nutritious foods like fruits and vegetables (50, 51) and encourages the consumption of inexpensive, high-calorie, high-fat foods (52). Moreover, these areas frequently lack parks for outdoor recreation (53). Debates are still ongoing about the relative culpability of energy intake vs. energy expenditure (physical activity) in weight gain. But most studies point to overconsumption of energy-dense foods as the primary culprit, and a decrease in physical activity is unlikely to be responsible for a dramatic increase in obesity in lower socioeconomic groups (54, 55).

Given the complexities of the factors influencing childhood obesity development (53, 56–58). Multidisciplinary interventions addressing individual, family, economic, environmental, social, and cultural barriers are required to prevent the progression of obesity (59).

To our knowledge, we are among the first to investigate wealth inequalities in overweight/obesity in United States children under the age of five using inequality measures proposed by the authors (5, 60). The advantage of using the SII and CIX is particularly relevant in the current study because it takes into account the entire population to overcome the effect of small sample sizes in some subgroups (5, 61). In addition, analysis on pooled data from all 9 surveys was conducted to increase the power of the analysis.

Limitations of this study include the cross-sectional nature of the NHANES data, which does not give insight on causality or causes; income may not correlate with individuals' actual socioeconomic status (62). This would necessitate the use of a longitudinal design to establish temporal precedence, as well as the inclusion of additional environmental, behavioral, and biological measurements (63). The CDC advises utilizing WHO growth standards for children under the age of two, and CDC growth benchmarks for children two and older (64). Nonetheless, we followed the WHO growth guidelines (65). Although these standards are probably not optimum for all parts of the pediatric population, they were applied to all children worldwide, regardless of ethnicity, socioeconomic status, or food type.

This study adds to previous research using the CIX technique in the domain of wealth inequality in the overweight/obese population in the United States. Overall, our findings give an update and reinforce the notion that overweight/obesity in children under the age of five has increased and that related wealth inequalities are a public health problem in the United States. In the United States, rising childhood overweight/obesity is associated with rising wealth inequality. We propose that policies and programs aimed at preventing childhood overweight/obesity should include minorities and low socioeconomic groups, but that population-based treatments should target al.l ethnic groups.