Independent and Interactive Associations of Subjective and Objective Socioeconomic Status With Body Composition and Parent-Reported Hyperphagia Among Children

Abstract

Background:

Subjective socioeconomic status (SSES) and objective socioeconomic status (OSES) have been independently associated with body composition and eating behavior in children. While low OSES may constrain access to healthier foods, low SSES has been associated with increased preference for and motivation to consume higher energy foods and portions independent of OSES. Despite these distinct ways that OSES and SSES may affect children's eating behavior and adiposity, their joint contributions remain unclear. We investigated the independent and interactive associations of SSES and OSES with children's BMI, fat mass index (FMI), and caregiver-reported hyperphagia.

Methods:

Data were derived from the Children's Growth and Behavior Study, an ongoing observational study. Multiple linear regressions used child's SSES and OSES of the family as independent factors and modeled the statistical interaction of SSES and OSES with BMI (n = 128), FMI (n = 122), and hyperphagia and its subscales (n = 76) as dependent variables.

Results:

SSES was independently and negatively associated with hyperphagia severity and OSES was independently and negatively associated with both FMI and hyperphagia severity. There was a statistical interaction effect of SSES and OSES on hyperphagia severity—lower SSES was associated with greater hyperphagia severity only at lower levels of OSES.

Conclusions:

These findings demonstrate a relationship between low OSES and child adiposity and that the relationship between child SSES and hyperphagia severity may be most relevant for children from households with lower family OSES. Future research on socioeconomic disparities in children's body composition and eating behaviors should examine the interaction of SSES and OSES.

Clinical Trial Registration:

NCT02390765.

Introduction

Socioeconomic disparities in childhood obesity are increasing in the United States (US), with higher prevalence among low-income families.^1,2 Between 2017 and 2020, nearly 20% of children aged 2–19 years were living with obesity.² Among US children from low-income families (family income 130% or less of the federal poverty level), 25.80% were living with obesity, while only 11.50% of children from high-income families (family income 350% or more of the federal poverty level) were living with obesity.²

Prior research has indicated that socioeconomic disparities in childhood obesity and unhealthy eating behaviors may emerge through differences in actual socioeconomic resources and variations in children's perceptions of their socioeconomic standing.^3–11 Yet, the moderating effect of actual resources on the relationship between status perceptions and children's weight and diet remains unclear.

Socioeconomic status (SES) is measured objectively and subjectively. Objective SES (OSES) refers to the actual socioeconomic resources of individuals or families, which are quantified by education, occupation, and/or income.¹² Conversely, subjective SES (SSES) refers to how one perceives their SES compared with others in a reference group (e.g., society).¹³

SSES represents perceived adequacy of one's socioeconomic circumstances and opportunities across the past, present, and future, which is not reflected in measures of OSES.¹⁴ Independent of OSES, low SSES is predictive of obesity,^11,15 poor self-rated health, and higher rates of mental health issues, including depression, anger, anxiety, and low self-esteem.⁹

OSES Associations with Obesity and Eating Behaviors

Several studies have demonstrated that in developed countries, lower neighborhood and family OSES is associated with higher BMI, unhealthy eating behaviors, and poor health in adolescents.^5–8 Among Belgian adolescents, parent occupation status was inversely associated with obesity, such that lower job status was associated with higher BMI.⁶

Goodman⁸ observed especially robust associations with OSES, indicating that poor parental education and low income are significantly associated with higher BMI, higher rates of depression, and poor self-perceived health in adolescents. In a study conducted by Pueyo et al.,³ there was a moderate social gradient of self-perceived health among adolescents, in which lower OSES was associated with poorer self-perceived health.

Moreover, OSES may affect motivational and behavioral patterns that increase risk for higher BMI, such as hyperphagia. Hyperphagia is defined as excessive eating and being overly preoccupied or motivated by food. In a prospective analysis of SES and appetite in early childhood, lower OSES was associated with higher food responsiveness, higher desire to drink any beverage, and higher emotional overeating, suggesting that OSES may contribute to components of hyperphagia.¹⁶

SSES Associations with Obesity and Eating Behaviors

SSES is linked to BMI and dietary habits. Lower SSES is associated with adolescents' poor self-rated health, mental health, physical activity, consumption of fruits and vegetables, and BMI.^9,11 Longitudinally, lower SSES is associated with an increased likelihood of overweight and obesity 5 years later.¹⁰ These trends correspond to findings that among female adults, lower SSES is associated with higher postlunch energy intake, which is indicative of poor energy compensation after a large meal and could result in longitudinal increase in fat mass.¹⁷

Additionally, prior studies examining the influence of experimentally manipulated SSES and relative deprivation on eating behaviors in adults suggest that low SSES may promote motivational and behavioral responses of increased energy intake, including larger portion selection, consumption of greater energy ad libitum, preference for high-energy snacks, and perceptual sensitivity to beverage energy content.^18–21

These findings suggest that low SSES may contribute to hyperphagic motivations and behaviors, although there has been limited research testing the relationship between SSES and hyperphagia among children.

SES Moderation on BMI and Diets

While low SSES and OSES may be independently associated with greater energy intake and BMI,^4,22 few studies have examined their interplay with these outcomes, beyond statistically adjusting for the other's influence. Prior studies examining the statistical interaction effect of SSES and OSES on health and well-being have suggested their additive risk or protection.

The combination of low SSES and OSES was associated with greater mental health problems among adolescents²³ and poorer self-rated health among older adults,²⁴ while higher SSES and OSES were associated with enhanced perceptions of overall health and well-being among young adults²⁵ and lower allostatic load among postpartum women.²⁶ Despite these findings, there has been limited research examining interaction effects of low SSES and OSES on behaviors that risk excess energy intake and adiposity in children.

OSES indicators such as education, income, and neighborhood quality explain less than half of the association between SSES and BMI. This suggests that there remain substantive unique effects of SSES on BMI that operate independently of OSES.²² OSES may reflect material resources that affect health, such as access to healthy foods, healthy lifestyles, and health care, while SSES may reflect psychological dispositions that impact health, such as self-esteem, life satisfaction, and resilience.^22,27

Additionally, SSES is only moderately correlated with education, occupation, and income.^13,28 Thus, we conceptualized the interplay of SSES and OSES with BMI and dietary behaviors through moderation.

SSES and OSES may interact to influence health and well-being, with OSES serving as a context for the absence or presence of resources (e.g., health care, physical/leisure activities, healthier diets) that exacerbate maladaptive responses to low SSES. Indicators of lower OSES are associated with poorer child diet quality,^29,30 and this relationship may partly operate through home food environments consisting of less healthy foods and eating practices (e.g., watching television during meals).³¹

Given the potential influence of low SSES on increased preferences for higher energy foods, larger portion sizes, and food intake,^{10,18,21,32,33} children reporting low SSES and OSES may be situated in environments that provide stronger motivation for palatable high-energy food intake. Together, low SSES and OSES may exacerbate hyperphagic tendencies and weight gain.

Identifying whether the influence of SSES on our outcomes is more relevant for those with higher or lower OSES is necessary for developing strategies to provide more holistic and successful interventions for children with obesity and who engage in maladaptive eating behaviors.

The Present Study

An exploratory analysis was performed to evaluate the independent associations of SSES and OSES with BMI, fat mass index (FMI), and caregiver-reported hyperphagia among children and whether OSES moderates the relationship between SSES and these outcomes. We hypothesized that lower SSES and OSES would be independently associated with increased BMI, FMI, and hyperphagia.

We also hypothesized there would be a statistical interaction between SSES and OSES, such that participants with low SSES and OSES would exhibit increased BMI, FMI, and hyperphagia.

Methods

Participants

Data were derived from the NICHD Intramural Children's Growth and Behavior Study, an ongoing study initiated in 2015, which examines how genes and the environment influence eating behavior and health longitudinally. Participants and their caregivers were recruited through direct mail, flyers, and previous study enrollment/eligibility.

Cognitively capable children (8–17 years old at baseline) in good general health, with BMI ≥5th percentile for age and sex,³⁴ and living in the Washington D.C. metro area were eligible for study enrollment. Child assent and caregiver consent were collected at baseline. Following baseline assessments, data were collected annually for up to 6 years.

The current study's cross-sectional analyses utilized data from baseline or year 3 follow-up (Y3) assessments, determined by the interval at which the participant's SSES was measured. For measures that were completed by caregivers, scores were averaged if both caregivers responded. The study procedures were approved by the National Institutes of Health Institutional Review Board.

Participants were excluded from analyses if they were older than 18 years or had missing data for exposures, covariates, or outcomes, resulting in three different sample sizes [n = 128 (BMI), n = 122 (FMI), and n = 76 (hyperphagia)].

Measures

Subjective socioeconomic status

The MacArthur Scale of Subjective Social Status–Youth Version¹¹ measured SSES by presenting children with a 10-rung ladder that represents American society and asking them to choose the rung that best represents their family's status based on income, education, and occupation. Higher scores indicated higher SSES. This measure was added to the study in November 2018.

Objective socioeconomic status

The Hollingshead Two-Factor Index of SES¹² measured OSES using caregivers' report of their educational and occupational background. Education and occupation scores ranged from 1 (graduate school/executive positions) to 7 (<7 years of education/service workers). The education score was weighted by three, the occupation score was weighted by five, and scores were added together to create an OSES score. The measure was reverse scored, so higher scores represented higher OSES.

BMI

Children's BMI (kg/m²) was calculated using weight and height measurements (in triplicate using a stadiometer and calibrated scale) and then converted into z-scores (BMIz) for age and sex according to CDC guidelines.³⁴

Fat mass index

Dual-energy X-ray absorptiometry (DXA) was used to determine children's total body fat mass, which was then converted to kilograms to calculate FMI (kg/m²).

Hyperphagia

The 13-item Hyperphagia Questionnaire³⁵ measured excessive preoccupation with and drive toward eating by asking caregivers to rate their child's symptoms on a 5-point scale, in which higher scores indicate more severe or frequent symptoms.

In addition to the total score (α = 0.87), three subscales were derived from the questionnaire: behavior (α = 0.52; e.g., cleverness in obtaining food), drive (α = 0.83; e.g., preoccupation with obtaining food), and severity (α = 0.72; e.g., extent that food-seeking behavior interferes with everyday functioning).^* Individual item ratings were converted into standardized z-scores before the subscale scores were computed since individual items used different rating scales.

Hyperphagia scores were highly positively skewed, with most participants reporting minimal symptoms of hyperphagia. This may have occurred because hyperphagia is one of the most severe conditions that characterize the tendency to overeat³⁶ and the questionnaire was originally developed to assess relatively severe hyperphagic behaviors in clinical populations (e.g., Prader–Willi Syndrome),³⁵ including items related to stealing or foraging for food in the trash, which may have lower frequencies in nonclinical populations.

Covariates

Covariates included the visit in which SSES was measured (baseline or Y3) since the time point when SSES was measured varied across participants. Child age (continuous) and child sex (male or female)^† were also included as covariates since these characteristics may be associated with eating behaviors of children.^37,38

Analyses

Analyses were conducted using SPSS 28.³⁹ Effects of cross-sectional statistical interactions between SSES and OSES on outcomes were tested using the Process macro model #1,⁴⁰ with low, mean, and high levels of SSES and OSES conditioned to values of −1 standard deviation (SD), mean, and +1 SD, respectively. In this model, SSES was the predictor and OSES was the moderator.

We examined the main effects of family SSES, family OSES, and the statistical interaction between the two on BMIz, FMI, and hyperphagia. Given deviations from normality (positive skew), coefficients for main effects and statistical interactions in the models were further estimated using wild bootstrapping with 5000 resamples. Wild bootstrapping was used because it is robust even when assumptions of normality and homoscedasticity are violated.⁴¹

Results

No outliers were removed from analyses. The final sample (n = 128), based on our largest analytic sample, was 47.70% female, 66.40% White, and 89.80% not Latino or Hispanic and had a mean age of 13.09 (SD = 2.52; Table 1) years. Despite the large age range (8–17 years), there was no correlation between age and any of our exposures or outcomes, and age was used as a covariate in all analyses (Table 2). SSES and OSES were positively and significantly correlated (b = 0.33, p < 0.01), yet multiple regression models, including the two variables as predictors, exhibited low multicollinearity (variance inflation factor of <2).

Table 1.

Descriptive Statistics of the Sample

Continuous variables	n	Minimum	Maximum	Mean (SD)
SSES	128	4.00	10.00	6.86 (1.23)
OSES	128	24.50	84.00	63.09 (10.46)
BMIz	128	−1.78	2.70	0.53 (0.97)
Fat mass (kg/m2)	122	1.94	17.49	5.83 (2.79)
Hyperphagia behavior	76	−2.19	14.46	0.05 (3.03)
Hyperphagia drive	76	−2.64	10.76	0.18 (3.34)
Hyperphagia severity	76	−0.91	8.05	0.09 (1.90)
Hyperphagia total	76	−5.73	33.27	0.33 (7.35)
Age (years)	128	8.00	17.00	13.09 (2.52)

Categorical variables	n	Response	Frequency	Proportion
Sex	128	Male	67	52.30%
		Female	61	47.70%
Race	128	Asian	9	7.00%
		Black	17	13.3%
		Multiple races	16	12.5%
		White	85	66.4%
Ethnicity	128	Latino or Hispanic	10	7.8
		Not Latino or Hispanic	115	89.8%
SSES interval	128	Baseline	71	55.50%
		Y3	57	44.50%

All variables were measured or reported by children, except for race/ethnicity, hyperphagia, and OSES, which were reported by parents.

BMIz, BMI as age- and sex-adjusted z-scores; FMI, fat mass index; OSES, objective socioeconomic status; SD, standard deviation; SSES, subjective socioeconomic status.

Table 2.

Spearman's Rho Correlation Coefficients for Continuous Variables With Sample Size in Parentheses

	SSES	OSES	BMIz	FMI (kg/m2)	Hyperphagia behavior	Hyperphagia drive	Hyperphagia severity	Hyperphagia total	Age (years)
SSES	1.00 (128)	—	—	—	—	—	—	—	—
OSES	0.33** (128)	1.00 (128)	—	—	—	—	—	—	—
BMIz	−0.03 (128)	−0.04 (128)	1.00 (128)	—	—	—	—	—	—
FMI (kg/m2)	0.02 (122)	−0.07 (122)	0.83** (122)	1.00 (122)	—	—	—	—	—
Hyperphagia behavior	0.03 (76)	−0.10 (76)	0.27* (76)	0.24 (73)	1.00 (76)	—	—	—	—
Hyperphagia drive	0.09 (76)	−0.19 (76)	0.22 (76)	0.18 (73)	0.65** (76)	1.00 (76)	—	—	—
Hyperphagia severity	−0.18 (76)	−0.06 (76)	0.31** (76)	0.28* (73)	0.57** (76)	0.53** (76)	1.00 (76)	—	—
Hyperphagia total	0.01 (76)	−0.17 (76)	0.29* (76)	0.24* (73)	0.89** (76)	0.88** (76)	0.68** (76)	1.00 (76)	—
Age (years)	0.03 (128)	−0.09 (128)	0.09 (128)	0.14 (122)	−0.05 (76)	−0.16 (76)	−0.22 (76)	−0.13 (76)	1.00 (128)

^* Significant at the 0.05 level (two-tailed).

^** Significant at the 0.01 level (two-tailed).

To determine whether there was a selection bias for those who were included based on SSES data availability, we performed independent sample t-tests to compare differences in OSES, BMI, FMI, hyperphagia, and age and a chi-squared test to compare differences in sex between those who were administered the SSES measure and those who were not.

We found significant differences between those who were and were not administered the SSES measure for OSES [t[316] = −2.39; p = 0.02 (without data: M = 60.16, SD = 11.79; with data: M = 63.21, SD = 10.96)], baseline hyperphagia behavior [t[292.18] = 2.34; p = 0.02 (without data: M = 6.95, SD = 2.24; with data: M = 6.42, SD = 1.67)], and baseline hyperphagia total score [t[296] = 2.35; p = 0.01 (without data: M = 16.11, SD = 4.89; with data: M = 14.86, SD = 4.16)].

BMI as Age- and Sex-Adjusted z-Scores

The BMIz model was not supported (R² = 0.07, F(6, 121) = 1.62, p = 0.15), and there were no main effects or statistical interactions of SSES and OSES (Table 3).

Table 3.

Model Summary, Main Effects, and Interaction of Subjective and Objective Socioeconomic Status (Confidence Intervals Based on 5000 Wild Bootstrap Samples)

Outcome	Analysis	R 2	F(df)	b	Bootstrapped 95% CI	p
BMIz	Model summary	0.07	1.62 (6, 121)	—	—	0.15
	SSES effect	—	—	−0.51	−1.14 to −0.09	0.20
	OSES effect	—	—	−0.06	−0.12 to −0.02	0.17
	SES interaction	—	—	0.01	−0.01 to 0.03	0.23
FMI (kg/m2)	Model summary	0.18	4.15 (6, 115)	—	—	0.00
	SSES effect	—	—	−2.08	−4.12 to 0.01	0.07
	OSES effect	—	—	−0.26	−0.47 to −0.05	0.03
	SES interaction	—	—	0.03	−0.001 to 0.07	0.06
Hyperphagia behavior	Model summary	0.08	0.99 (6, 69)	—	—	0.44
	SSES effect	—	—	−2.15	−6.14 to 1.80	0.21
	OSES effect	—	—	−0.28	−0.71 to 0.15	0.13
	SES interaction	—	—	0.04	−0.02 to 0.10	0.18
Hyperphagia drive	Model summary	0.17	2.28 (6, 69)	—	—	0.05
	SSES effect	—	—	−0.90	−4.79 to 2.88	0.61
	OSES effect	—	—	−0.24	−0.64 to 0.16	0.20
	SES interaction	—	—	0.02	−0.04 to 0.09	0.38
Hyperphagia severity	Model summary	0.23	3.38 (6, 69)	—	—	0.01
	SSES effect	—	—	−2.65	−5.17 to −0.16	0.01
	OSES effect	—	—	−0.28	−0.56 to −0.02	0.01
	SES interaction	—	—	0.04	0.003 to 0.08	0.01
Hyperphagia total	Model summary	0.15	2.06 (6,69)	—	—	0.07
	SSES effect	—	—	−5.70	−14.74 to 3.29	0.15
	OSES effect	—	—	−0.80	−1.79 to 0.16	0.06
	SES interaction	—	—	0.10	−0.05 to 0.25	0.11

CI, confidence interval.

Fat Mass Index

The main effect of SSES on FMI was not significant [b = −2.08; 95% confidence interval (CI): −4.12 to 0.01; Table 3]. There was a significant main effect of OSES on FMI (b = −0.26; 95% CI: −0.47 to −0.05), such that lower OSES was associated with higher FMI. No significant statistical interaction between SSES and OSES for FMI was identified based on bootstrapped CIs (b = 0.03; 95% CI: −0.001 to 0.07).

Hyperphagia

The hyperphagia behavior and total score models were not significant; however, the drive and severity models were significant, and only the hyperphagia severity model had significant associations for the variables of interest (Table 3). There was a significant main effect of SSES on hyperphagia severity (b = −2.65; 95% CI: −5.17 to −0.16), such that lower SSES was associated with higher hyperphagia severity.

The main effect of OSES on hyperphagia severity was also significant (b = −0.28; 95% CI: −0.56 to −0.02), such that lower OSES was associated with higher hyperphagia severity. There was a significant statistical interaction between SSES and OSES for hyperphagia severity (b = 0.04; 95% CI: 0.003–0.08). Lower SSES was associated with higher hyperphagia severity only at lower levels of OSES (b = −0.53; 95% CI: −1.04 to −0.04; Fig. 1).

Figure 1.

SSES and OSES interaction effect on hyperphagia severity, b = 0.04, 95% CI: 0.003–0.08. Lower SSES is associated with higher hyperphagia severity only at lower levels of OSES, b = −0.53, 95% CI: −1.03 to −0.06. CI, confidence interval; OSES, objective socioeconomic status; SSES, subjective socioeconomic status.

Discussion

Although we did not replicate prior findings that lower SSES and lower OSES are independently associated with higher BMIz, we did observe an independent association of OSES with FMI and hyperphagia severity and a statistical interaction, such that lower SSES was significantly associated with hyperphagia severity at lower levels of OSES. These findings suggest that low SSES and OSES may jointly contribute to increased hyperphagia severity by imposing additive risks.

Lower SSES and OSES have been consistently associated with higher BMI^5–8,10,11; however, little research has directly tested effects of their interplay on adiposity and hyperphagia among children. Prior studies examining the role of SSES and OSES in these outcomes (which have mostly focused on adults) have examined mediating relationships.^4,22,42

Although SSES may be one psychological pathway through which socioeconomic disparities in these outcomes may emerge, the association between SSES and adiposity unaccounted for by OSES suggests other ways in which these two dimensions of SES may interact.^9,11,22 Indeed, prior health studies have indicated moderating relationships between low SSES and OSES.^23–26

For example, Bøe et al.²³ found a statistical interaction effect in which higher household income and more favorable perceived economic well-being were associated with fewer symptoms of mental health problems. In this study, we demonstrate that the relationship between low SSES and OSES may extend to the severity of hyperphagic tendencies.

Given that we observed significant main effects and statistical interactions of SSES and OSES only for hyperphagia severity, socioeconomic factors may affect how severe and disruptive hyperphagia is overall, rather than predict specific hyperphagic tendencies. The finding that low SSES is associated with greater hyperphagia severity in children (especially in households with lower OSES) suggests that the effects of low SSES on increased preference for palatable high-energy foods and greater food consumption observed among adults^18,21 may also extend to children.

One limitation of this study is the small sample size. These findings would benefit from replication in other samples, especially those with greater racial/ethnic and socioeconomic representation. Given the cross-sectional nature of the study, causal relationships between SES and outcomes remain unclear. Longitudinal studies and experimental approaches are needed on this topic. The role of food insecurity also could not be examined. Finally, hyperphagia was the only behavioral outcome examined in the study. Future studies should investigate whether the interactive SES relationships extend to other food-related behaviors.

This is the first study to our knowledge that explores the moderating role of OSES in the relationship between SSES and BMI, FMI, and hyperphagia among children. Many prior studies of status in children have used measures of subjective social status rather than SSES. Although conceptually related to SSES, subjective social status is based on relative popularity, respect, and importance rather than economic resources.³³ Thus, measures of both SSES and OSES in this study are directly relevant to understanding the role of socioeconomic inequity in these outcomes.

Strengths of this study include analysis of the relationships of SSES and OSES with FMI using DXA. This is an advantage because fat mass (compared with BMI) is rarely examined as an outcome of SSES among children and FMI is a direct ratio of fat mass to height. This may be one reason why we observed a stronger relationship between SSES and OSES for FMI compared with BMIz, although neither relationship was significant. Furthermore, hyperphagia was measured by caregiver report rather than self-report. Caregiver report may limit socially desirable responding or demand characteristics since children are not directly reporting their hyperphagic symptoms.

In summary, we provide initial support for the potential interactive contribution of low SSES and OSES to greater hyperphagia severity. Some prior interventions that have sought to provide financial resources to low-OSES households did not observe improvements to diet quality.^43,44 These findings suggest that other aspects of the experiences of low OSES may be affecting diet quality beyond constraints to household budget or access to healthier diets, such as SSES.

Based on our current findings, future interventions to reduce socioeconomic disparities in eating behaviors and diet quality of children may need to consider the role of not only OSES constraints but also low SSES in eating behaviors.

Impact Statement

This study demonstrates that lower subjective socioeconomic status reported by children is associated with greater adiposity and severity of hyperphagic behaviors when they are situated within households of lower objective socioeconomic status. Studies on child eating behaviors and body composition should consider the interactions of both forms of socioeconomic disadvantage.

Authors' Contributions

J.A.Y. and M.T.K. conceived the study. J.A.Y., L.K.L., H.E.H., B.F.B., J.T.V., A.I.B., S.M.B., M.T.K., and K.Y.C. conducted primary data collection and entry. M.R.S. organized and analyzed the data. M.R.S., J.M.P.B., J.A.Y., and B.K.C. participated in drafting the article and had final approval of the submitted and published versions. All authors reviewed and edited the article before submission.

Disclaimer

The opinions and assertions expressed herein are those of the authors and are not to be construed as reflecting the views of the National Institutes of Health, US Department of Health and Human Services, Uniformed Services University of the Health Sciences, Department of Defense, or Metis Foundation.

Funding Information

This research was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (B.K.C.: ZIAHD009004-01656312; and J.A.Y.: ZIAHD00641).

Author Disclosure Statement

J.A.Y. has received grant support unrelated to this article for pharmacotherapy trials for obesity from Hikma Pharmaceuticals, Inc., Soleno Therapeutics, Inc., and Rhythm Pharmaceuticals, Inc., as well as support for basic science studies from Versanis Bio, Inc. No other potential conflicts of interest relevant to this article were reported by the other authors.