Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Nov 1.
Published in final edited form as: J Abnorm Psychol. 2021 Nov;130(8):875–885. doi: 10.1037/abn0000719

Context matters: Neighborhood disadvantage is associated with increased disordered eating and earlier activation of genetic influences in girls

Megan E Mikhail 1, Sarah L Carroll 1, D Angus Clark 2, Shannon O’Connor 3, S Alexandra Burt 1, Kelly L Klump 1
PMCID: PMC8634784  NIHMSID: NIHMS1738258  PMID: 34843291

Abstract

Emerging evidence suggests socioeconomic disadvantage may increase risk for eating disorders (EDs). However, there are very few studies on the association between disadvantage and EDs, and all have focused on individual-level risk factors (e.g., family income). Neighborhood disadvantage (i.e., elevated poverty and reduced resources in one’s neighborhood) is associated with increased risk for anxiety/depression and poor physical health. To date, no studies have examined phenotypic associations between neighborhood disadvantage and disordered eating, or how any form of disadvantage may interact with genetic individual differences in risk for EDs. We examined phenotypic and etiologic associations between neighborhood disadvantage and disordered eating in 2,922 girls ages 8–17 from same sex twin pairs recruited through the Michigan State University Twin Registry. Parents rated the twins on nine items assessing core disordered eating symptoms (e.g., weight preoccupation, binge eating), and neighborhood disadvantage was calculated from 17 indicators of contextual disadvantage (e.g., median home value, neighborhood unemployment). Puberty was measured using the Pubertal Development Scale to examine whether associations were consistent across development. At a phenotypic level, greater neighborhood disadvantage was associated with significantly greater disordered eating symptoms in girls at all stages of puberty (β = .07). Moreover, genotype x environment models showed that girls living in more disadvantaged neighborhoods exhibited stronger and earlier (i.e., during pre/early puberty) activation of genetic influences on disordered eating. Results highlight the critical importance of considering contextual disadvantage in research on etiology and risk for disordered eating, and the need for increased screening and treatment for EDs in disadvantaged youth.

Keywords: disordered eating, eating disorders, disadvantage, etiology, genotype x environment interaction

General Scientific Summary

This was the first study to examine the association between neighborhood disadvantage and disordered eating in girls. We found that neighborhood disadvantage was associated with significantly greater disordered eating symptoms across all stages of puberty in girls, and that girls living in disadvantaged neighborhoods exhibited stronger and earlier (i.e., during pre/early puberty) activation of genetic influences on disordered eating. Results indicate that neighborhood disadvantage is a critical contextual factor with implications for etiology and risk for disordered eating in girls.

Historically, eating disorders (EDs, including anorexia nervosa, bulimia nervosa, binge-eating disorder, and other specified feeding and eating disorders) have been viewed as predominantly affecting people from middle-to-upper class socioeconomic backgrounds (Gard & Freeman, 1996). However, recent research suggests that socioeconomic disadvantage (e.g., food insecurity) may increase risk for key disordered eating symptoms (e.g., weight/shape concerns, binge eating) that predict later development of EDs (Becker et al., 2017, 2019; Coffino et al., 2020; Lydecker et al., 2019). Importantly, all research to date has focused on individual-level or proximal disadvantage (i.e., individual/family socioeconomic status (SES)) rather than the influence of the wider socioeconomic context on disordered eating. Neighborhood disadvantage, defined as elevated poverty and reduced community resources (e.g., safe housing, grocery stores) in the area surrounding a person’s home (Attar et al., 1994), is a significant risk factor for anxiety/depression and externalizing disorders (Burt et al., 2015; Choi et al., 2021). Although neighborhood disadvantage is correlated with familial income and poverty (r’s ~ .3 to .5; Hackman et al., 2012; Roubinov et al., 2018), it is a distinct process that has been shown to have unique and significant effects on psychopathology (Burt, 2014).

Many possible pathways could link neighborhood disadvantage to disordered eating. Disadvantaged contexts are characterized by substantially increased stress, a transdiagnostic risk factor for multiple forms of psychopathology (Brenner et al., 2013; Mennis et al., 2016). Other features of disadvantaged neighborhoods may specifically increase risk for disordered eating, including maladaptive food environments such as “food deserts” (limited access to fresh foods and grocery stores) and “food swamps” (an overabundance of highly palatable foods) that are common in disadvantaged neighborhoods and associated with increased body mass index (BMI) (Cooksey-Stowers et al., 2017; Dubowitz et al., 2012).

In addition to examining whether neighborhood disadvantage is linked to disordered eating, it is also important to identify why and/or how neighborhood disadvantage may impact disordered eating. Neighborhood disadvantage could influence disordered eating through environmental main effects, where contextual influences directly impact ED outcomes. For example, living in a neighborhood with high fast-food availability could increase risk for dysregulated eating through the direct effects of environmental factors that encourage palatable food consumption.

By contrast, neighborhood disadvantage could influence disordered eating indirectly through genotype x environment interactions. Genotype x environment interactions occur when the impact of genetic risk differs based on a person’s life experiences and context (Burt, 2011). In other words, a person may have underlying genetic risk for disordered eating (heritability = ~50% in late adolescence/adulthood; Mikhail et al., 2019), but this genetic risk may be more likely to lead to disordered eating in certain environmental contexts.

There are two primary types of genotype x environment interactions. Diathesis-stress interactions occur when genetic influences on a trait are strengthened in the presence of environmental stress. Should etiologic moderation take the form of a diathesis-stress genotype x environment interaction, people who are genetically predisposed to disordered eating would experience activation of latent genetic risk when exposed to neighborhood disadvantage. If diathesis-stress genotype x environment interactions are operative, we would therefore expect to see stronger genetic influences on disordered eating among individuals living in more disadvantaged neighborhoods. Diathesis-stress genotype x environment interactions involving neighborhood disadvantage have previously been observed for other internalizing phenotypes (e.g., depression; Strachan et al., 2017).

Alternatively, environmental factors may become more influential in more disadvantaged environments, overshadowing individual differences in genetic risk. These types of “bioecological” genotype x environment interactions can occur when environmental stressors that are outside the average/expected range of experience disrupt normative development (Bronfenbrenner & Ceci, 1994; Burt, 2014) (e.g., genetic differences in height may be obscured by limited access to food during a famine). In the presence of a bioecological genotype x environment interaction, we would expect to see stronger environmental influences on disordered eating among people living in more disadvantaged neighborhoods.

Only a few studies have examined genotype x environment interactions with respect to EDs despite calls for research that integrates genetic and environmental risk (Wade & Bulik, 2018). Twin studies are an ideal approach to examining genotype x environment interactions for complex traits that are influenced by multiple genes, such as EDs, because they can examine the totality of additive genetic influences on a phenotype of interest. Twin research has revealed stronger genetic influences on disordered eating in the presence of proximal stressors, such as weight-related teasing (Fairweather-Schmidt & Wade, 2017) and parental divorce (Suisman et al., 2011), suggesting diathesis-stress genotype x environment interactions. A small body of literature taking a candidate gene approach also suggests stronger associations between genetic variants and disordered eating among people who have experienced stressful life events (Akkermann et al., 2012; Micali et al., 2017; Rozenblat et al., 2017), but such studies should be interpreted with caution given the significant limitations associated with examining a single gene (Border & Keller, 2017). Importantly, to date, no research has examined genotype x environment interactions for any form of disadvantage (at individual, family, or neighborhood levels) and disordered eating, despite the importance of this work for etiologic models and prevention/treatment of EDs among people living in disadvantaged contexts.

In the present study, we examined phenotypic associations and genotype x environment interactions between neighborhood disadvantage and disordered eating symptoms in a large, population-based sample of female twins (N = 2,922) that spanned child and adolescent development (i.e., ages 8–17). We examined a broad measure of core disordered eating symptoms to capture the full spectrum of pathology, including distressing symptoms that may not meet diagnostic thresholds but contribute to impairment (Hart et al., 2020) and future risk for severe EDs in youth (e.g., binge eating, weight/shape concerns) (Le Grange & Loeb, 2007). Past studies show dramatic developmental differences in disordered eating in girls across puberty, with substantial increases in disordered eating symptoms (Nagl et al., 2016) and genetic influences (from 0% to ~50% heritability; Klump et al., 2003, 2007, 2012; Mikhail et al., 2019; O’Connor et al., 2020) from pre- to post-puberty. We modeled these developmental changes to examine whether phenotypic and etiologic associations between neighborhood disadvantage and disordered eating differed across this critical developmental stage.

Methods

Participants

Analyses included 2,922 female twins from same-six twin pairs (ages 8–17; M =12.49, SD = 3.00) (see Table 1 for demographic information) from the population-based Michigan Twins Project (MTP) (Burt & Klump, 2013, 2019; Klump & Burt, 2006). MTP twins are demographically representative of Michigan (Burt & Klump, 2019) and vary widely in family SES (see Table 1). Similar to Michigan population rates, ~14% of MTP youth live in families whose income is at or below the federal poverty level (Burt & Klump, 2013). Study procedures were approved by the Michigan State University Institutional Review Board.

Table 1.

Descriptive statistics for participant demographics and symptoms (N = 2,922)

Participant Characteristics Mean (SD) or % of Sample (N) Range
Age 12.49 (3.00) 8.08–17.98
Zygosity (N listed as number of pairs)
 Monozygotic same sex female pairs 46.3% (676) —
 Dizygotic same sex female pairs 53.3% (779) —
 Female same sex pairs of unknown zygosity 0.4% (6) —
Race/ethnicity
 White (non-Latinx) 82.8% (2,420) —
 Black/African American (non-Latinx) 7.8% (228) —
 Latinx/Hispanic 1.6% (48) —
 Asian American 1.0% (28) —
 Native American/American Indian 0.4% (12) —
 More than one race 4.1% (120) —
 Other/Unknown 2.3% (66) —
BMI percentile 55.89 (29.77) 0.5–99.5
Raw BMI 19.93 (4.53) 12.88–44.64
Pubertal status (PDS score) 2.55 (.98) 1–4
Combined parental income (in thousands of dollars) $85.80 (52.97) $0–$300+
Mother’s education level
 Less than high school 3.9% (112) —
 High school graduate 18.4% (524) —
 Less than 4 years of college 31.7% (904) —
 College graduate (4–6 years of college) 32.4% (924) —
 Post-graduate education 13.5% (386) —
Father’s education level
 Less than high school 5.5% (142) —
 High school graduate 27.1% (706) —
 Less than 4 years of college 27.0% (704) —
 College graduate (4–6 years of college) 28.6% (744) —
 Post-graduate education 11.8% (308) —
Area Deprivation Index (ADI) percentile rank relative to all census tracts in the United States 38.84 (26.59) 1–100
Symptom Measures Mean (SD) or % of Sample (N) Sample Range Possible Range Cronbach’s alpha
MTP-ED total score 1.76 (2.80) 0–16 0–18 .84
Reported having AN, BN, or BED 1.3% (35) — — —
Reported being treated for AN, BN, or BED 0.5% (14) — — —
Internalizing symptoms 1.85 (1.90) 0–10 0–10 .63
Open in a new tab

Note: BMI = body mass index; MTP-ED = Michigan Twins Project Eating Disorder Survey; AN = anorexia nervosa; BN = bulimia nervosa; BED = binge eating disorder; internalizing symptoms = score on the Emotional Symptoms subscale of the Strengths and Difficulties Questionnaire. N’s may not add up to the total N for all variables due to missing values.

Measures

Zygosity Determination.

Zygosity was determined using a well-validated physical similarity questionnaire (Lykken et al., 1990) completed by the twins’ parents that shows over 95% accuracy based on validation studies using serological analysis and genetic markers (Lykken et al., 1990; Peeters et al., 1998).

Disordered Eating.

Disordered eating was assessed using the Michigan Twins Project Eating Disorder Survey (MTP-ED), a nine-item, parent-report questionnaire typically completed by the twins’ mother. The MTP-ED contains questions regarding body dissatisfaction (i.e., distress regarding body shape), weight preoccupation (i.e., fear of gaining weight), and disordered eating behaviors (i.e., dieting, binge eating, purging; see Table S1 for items) modeled in part on items from the Eating Disorder Examination-Questionnaire (EDE-Q; Fairburn & Beglin, 1994). Each item is rated on a 3-point scale from 0 (not true) to 2 (certainly true).

The MTP-ED showed good internal consistency in both the full sample (α = .83) and subsamples across puberty (pre/early puberty girls: α = .79; mid/late puberty girls: α = .85) in this study. Internal consistency was similar in child (ages 8–12; α = .82) and adolescent (ages 13–17; α = .85) participants. As expected, MTP-ED scores were significantly, positively correlated with age, pubertal status, BMI percentile, and internalizing symptoms (see Table S3). To further evaluate criterion validity, we examined whether MTP-ED scores were higher in participants whose parents indicated they had been diagnosed with a lifetime ED (AN, BN, or BED) on a checklist of physical and mental health conditions on the MTP intake questionnaire. The MTP-ED discriminated between girls with and without a lifetime parent-reported ED (M(SD) with no ED: 1.75 (2.78); M(SD) with lifetime ED: 3.08 (4.33), p = .016).

To establish convergent validity with existing measures, we examined twin self- and parent-report versions of the MTP-ED that were included in the Twin Study of Mood, Behavior, and Hormones during Puberty (TSMBH; Klump et al., 2018) (N = 559, 196 of whom (6.7%) overlap with the current sample) from the MSUTR. MTP-ED parent and twin self-report were significantly correlated (r = .37, p < .001), with higher child-parent agreement than previously observed for other established symptom measures for relevant phenotypes (e.g., parent-child report rs = .18–.30 for measures of binge eating and depressive symptoms; Vo et al., 2019). Moreover, parent-reported MTP-ED scores for twins in the TSMBH were significantly correlated with parent-reported total scores on the EDE-Q (r = .79, p < .001) and Minnesota Eating Behavior Survey1 (MEBS; von Ranson et al., 2005) (r = .72, p < .001). Self-reported MTP-ED scores from twins in the TSMBH were similarly correlated with their self-reported total scores on the EDE-Q and MEBS (r = .77 for both measures, p < .001).

Neighborhood Disadvantage.

Neighborhood disadvantage was operationalized using a well-validated (Kind & Buckingham, 2013; Singh, 2003), census-tract level Area Deprivation Index (ADI) incorporating 17 indicators of neighborhood disadvantage (e.g., unemployment rate, median home value; see Table S2 for all indicators). The ADI score for each family was coded using publicly available data from the American Community Survey for the census-tract that contained the family’s address (https://www.neighborhoodatlas.medicine.wisc.edu/). Raw ADI scores were converted into percentiles relative to other families in the sample, with higher scores indicating greater neighborhood disadvantage.

Neighborhood disadvantage as measured by the ADI is correlated with poorer physical (Kind et al., 2014; Powell et al., 2020) and mental (Burt et al., 2020) health, as well as higher BMI (Sheets et al., 2020) and lower physical activity (Miller et al., 2020). The ADI also has excellent internal consistency (α = .95 in past research; Singh, 2003). Notably, family SES was only moderately correlated with the ADI in the current study (r = −.48; 23% variance shared), suggesting that individual and neighborhood disadvantage are related but unique constructs.

Pubertal Development.

Pubertal maturation was assessed using the five-item, parent-report Pubertal Development Scale (PDS; Petersen et al., 1988) that assess the physical changes of puberty (e.g., growth spurts, breast development). Parent-rated PDS correlates strongly with professionally-rated Tanner staging and shows good psychometric properties (α = .91–.96; Koopman-Verhoeff et al., 2020). As in past research (Klump et al., 2003, 2012) PDS items were averaged to create an overall score.

BMI Percentile.

Age-specific BMI percentiles were calculated from parent-reported height and weight using CDC growth charts (https://www.cdc.gov/healthyweight/xls/bmi-group-calculator-us-062018-508.xlsm).

Statistical Analyses

Data Preparation.

MTP-ED and PDS scores were prorated if one item was missing and marked as missing if more than one item was missing. While participant-reported BMI shows good concordance with objective measures (Gorber et al., 2007), we took a conservative approach in setting extreme BMI values <0.5th percentile or >99.5th percentile to missing. MTP-ED scores were log transformed to account for positive skew and standardized prior to analysis. Race/ethnicity was included as a covariate in phenotypic analyses, but was not significantly associated with disordered eating (see Results). Thus, we did not include race/ethnicity as a covariate in genotype x environment analyses. Unfortunately, we did not have enough participants of color to examine whether race/ethnicity moderated genotype x environment effects. However, prior research has shown that genetic/environmental influences on disordered eating are similar across race (Munn-Chernoff et al., 2015), and we did not have a priori reason to expect differences in genotype x environment interactions associated with neighborhood disadvantage across race. Youth from disadvantaged neighborhoods tend to have higher BMIs (Alvarado, 2016), and higher BMIs are associated with disordered eating (Neumark-Sztainer et al., 2007). All analyses were therefore conducted with and without BMI percentile to directly assess its impact.

Phenotypic Analyses.

Multilevel models (MLMs) with a random intercept to account for nesting of twins within families were used to examine phenotypic associations between neighborhood disadvantage and disordered eating. We could not include random slopes for neighborhood disadvantage because levels of neighborhood disadvantage did not differ within family. Including random slopes for covariates that could differ within family (e.g., pubertal status) did not change any results, and so we present the more parsimonious random intercept models below. Models included puberty x neighborhood disadvantage interactions to examine whether associations were similar across development.

Genotype x Environment Analyses.

We conducted extended univariate, double moderator twin models (van der Sluis et al., 2012) with disadvantage and pubertal status as the moderators (see Figure S1). This model examines differences in additive genetic (A; i.e., genetic influences that sum across genes), shared environmental (C; i.e., environmental factors that increase similarity between co-twins), and non-shared environmental (E; i.e., environmental factors that differentiate co-twins, including measurement error) influences on disordered eating across levels of neighborhood disadvantage and puberty. Because moderators are included in the means model, values of A, C, and E reflect the etiology of disordered eating after removing the variance shared with the moderators. Analyses included 12 parameters of interest: 3 path coefficients (a, c, e in Figure S1) that capture genetic/environmental influences at the lowest level of the moderators (i.e., lowest level of neighborhood disadvantage for girls in pre-puberty), and 9 moderation coefficients that capture linear increases/decreases in the initial ACE path coefficients as a function of pubertal development (βXP, βYP, βzP in Figure S1), neighborhood disadvantage (βXD, βYD, βzD in Figure S1), and their interaction (βXPD, βYPD, βzPD in Figure S1). The extended genotype x environment model described by van der Sluis (2012) allowed us to include twins who were discordant on pubertal status while correcting for potential biases in significance testing resulting from the correlation between pubertal status and disordered eating. All twins were concordant on ADI percentile, as this is a family-level variable; thus, the correlation between ADI percentile and disordered eating was not a concern for significance testing within the twin models.

We first fit the full model with all path estimates and moderators freely estimated, then fit submodels based on the full model results to identify a best-fitting model. This approach allowed us to identify relevant submodels without conducting an excessive number of tests. Best-fitting models were those that had a non-significant difference in minus twice the log-likelihood (−2lnL) between the full and nested model, and minimized Akaike’s Information Criterion (AIC), Bayesian Information Criterion (BIC), and sample-size adjusted BIC (SABIC). If AIC, BIC, and SABIC identified different models as best-fitting, the model that optimized two out of three indices and had a non-significant change in −2lnL was selected as the best-fitting.

ADI percentiles and PDS scores were floored at 0, then scaled from 0–1 for interpretability. Following prior recommendations (Purcell, 2002), tables and figures report unstandardized path coefficient and moderator estimates that reflect absolute differences in genetic and environmental influences across the moderators.

Results

Phenotypic Analyses

A wide range of neighborhood disadvantage and disordered eating symptoms was represented (see Table 1). Indeed, 338 participants (11.6%) lived in disadvantaged neighborhoods above the national 75th percentile on the ADI, and 17.1% scored above the MPT-ED mean for girls with a parent-reported ED diagnosis. Neighborhood disadvantage was significantly correlated with MTP-ED scores at the bivariate level (r = .09, p < .001) (see Table S3), and associations remained significant after controlling for puberty and race/ethnicity in the MLM (β = .07, p = .003) (see Table 2). Girls at more advanced pubertal stages exhibited greater disordered eating symptoms (β = .32, p <.001), but the lack of a significant puberty x ADI interaction (p = .932) indicated that neighborhood disadvantage was associated with greater ED symptoms for girls across all stages of development.

Table 2.

MLM examining associations between neighborhood disadvantage and disordered eating symptoms across puberty

BMI Percentile Not Included as a Covariate
Variable β SE p 95% CI
Intercept .01 .03 .598 −.04, .06
Neighborhood disadvantage .07 .03 .003 .03, .12
Pubertal status .32 .02 <.001 .28, .36
Neighborhood disadvantage x puberty −.002 .02 .932 −.05, .04
Race/ethnicity
 Black/African American (non-Latinx) −.08 .09 .369 −.27, .10
 Latinx/Hispanic .11 .18 .562 −.25, .46
 Asian American .17 .25 .490 −.31, .65
 Native American/American Indian −.28 .42 .501 −1.10, .54
 More than one race −.07 .12 .563 −.30, .16
 Other/unknown −.04 .17 .807 −.37, .29
BMI Percentile Included as a Covariate
Variable β SE p 95% CI
Intercept .03 .03 .232 −.02, .08
Neighborhood disadvantage .03 .03 .285 −.02, .08
Pubertal status .24 .02 <.001 .20, .29
Neighborhood disadvantage x puberty −.003 .02 .881 −.05, .04
Race/ethnicity
 Black/African American (non-Latinx) −.11 .10 .290 −.30, .09
 Latinx/Hispanic .06 .19 .738 −.31, .43
 Asian American .07 .25 .786 −.42, .55
 Native American/American Indian −.24 .46 .592 −1.14, .65
 More than one race −.21 .12 .088 −.45, .03
 Other/unknown −.13 .16 .415 −.45, .19
BMI percentile .32 .02 <.001 .28, .36
Open in a new tab

Note: MLM = multilevel model; neighborhood disadvantage = Area Deprivation Index (ADI) percentile; pubertal status = continuous Pubertal Development Scale score; BMI = body mass index. The outcome for all models is standardized, log-transformed Michigan Twins Project Eating Disorder Survey (MTP-ED) total score. Reference group for race/ethnicity is white/Caucasian. Effects significant at p < .05 are bolded.

When BMI percentile was added to the model, the association between neighborhood disadvantage and disordered eating was in the same direction, but was attenuated (β = .03, p = .285; see Table 2). These results suggest that girls living in more disadvantaged neighborhoods may have more disordered eating symptoms, in part, because they tend to be at higher body weights, which may result in greater weight stigma and body dissatisfaction.

Genotype x Environment Analyses

Importantly, more modest phenotypic associations between neighborhood disadvantage and disordered eating do not preclude the possibility of a genotype x environment interaction. In fact, more modest associations may be expected if the relationship between neighborhood disadvantage and disordered eating differs across genetic and/or environmental risk. Interestingly, and as shown in Supplemental Material (see Tables S4–S5 and Figure S2), genotype x environment results were nearly identical with and without BMI percentile regressed out of the MTP-ED total score. Thus, we focused our results/interpretations on analyses that controlled for BMI percentile.

Parameter estimates from the full genotype x environment model are presented in Figure 1. Panels (presented from left to right) show differences in additive genetic influences (ADI-A), shared environmental influences (ADI-C), and non-shared environmental influences (ADI-E) on disordered eating across neighborhood disadvantage. ADI percentile is represented on the x-axis, such that greater values indicate greater neighborhood disadvantage. While continuous PDS scores were used in analyses, it not possible to graphically depict two continuous moderators. Therefore, differences in A, C, and E across neighborhood disadvantage are depicted at the extremes and mid-point of pubertal development: pre-puberty (PDS = 1), mid-puberty (PDS = 2.5), and post-puberty (PDS = 4). Parameter estimates in Figure 1 seemed to show substantial differences in genetic effects across both neighborhood disadvantage and puberty. At lower levels of disadvantage, genetic influences on disordered eating were much greater in post-pubertal than pre-pubertal girls, a finding that replicates past work (Klump et al., 2003, 2007, 2012). By contrast, at the highest levels of the ADI, genetic influences on disordered eating were substantial in pre-pubertal as well as post-pubertal girls. These differences across neighborhood disadvantage and puberty were striking and were in contrast to effects for nonshared environmental (which remained stable across neighborhood disadvantage and puberty) and shared environmental (which seemed to differ across puberty, but not neighborhood disadvantage) factors (see Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Additive genetic (A), shared environmental (C), and non-shared environmental (E) influences on disordered eating across puberty and neighborhood disadvantage. ADI = Area Deprivation Index percentile (higher values indicate greater neighborhood disadvantage); BMI = body mass index; Pre-Puberty = Pubertal Development Scale (PDS) score of 1; Mid-Puberty = PDS score of 2.5; Post-Puberty = PDS score of 4.

In terms of model-fitting, we first fit a no moderation model to test for significant moderation effects. This model fit poorly, with a significant change in −2lnL (see Table 3). We then fit a model that constrained all E moderators to zero, and a model that constrained all E moderators and the ADI C moderator to zero. Both of these constraints improved model fit (see Table 3) and thus, they were included in all subsequent models. Finally, we tested a series of models that constrained the puberty x ADI moderators for A and/or C to zero to test whether associations between neighborhood disadvantage and these components of variance differed significantly across puberty.

Table 3.

Model fit comparisons for genotype x environment models with BMI percentile regressed out

Model −2lnL χ2 Δ (df) p AIC BIC SABIC
Full model
Nested submodels 5811.810 — — 5857.810 5976.846 5903.786
 No moderation 5866.872 55.062 (9) <.001 5894.871 5967.328 5922.857
 Constrain all E mods 5813.364 1.554 (3) .670 5853.363 5956.873 5893.342
 Constrain all E mods, ADI C mod 5813.440 1.630 (4) .803 5851.440 5949.774 5889.421
 Constrain all E mods, ADI C mod, puberty X ADI C mod 5813.576 1.766 (5) .880 5849.575 5942.734 5885.556
 Constrain all E mods, ADI C mod, puberty X ADI A mod 5816.540 4.730 (5) .450 5852.540 5945.699 5888.522
 Constrain all E mods, ADI C mod, puberty X ADI AC mods 5818.284 6.474 (6) .372 5852.284 5940.267 5886.266
Open in a new tab

Note: ADI = Area Deprivation Index percentile; BMI = body mass index; puberty = Pubertal Development Scale (PDS) score; mod(s) = moderator(s); −2lnL = minus twice the log-likelihood; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; SABIC = sample size adjusted Bayesian Information Criterion; full model = model with paths and all moderators; A = additive genetic variance; C = shared environmental variance; E = nonshared environmental variance.

As shown in Table 3, the best-fitting model retained the puberty x ADI A moderator, but constrained the puberty x ADI C moderator to zero. This model was best fitting on two out of three fit indices (AIC, SABIC) and showed a non-significant change in −2lnL. Parameter estimates from the best-fitting model were nearly identical to the full model (see Table 4 and Figure 1), and showed significantly greater genetic influences on MTP-ED scores for pre/early pubertal girls in disadvantaged neighborhoods. Strikingly, the heritability of disordered eating was more than twice as high in pre-pubertal girls from the most disadvantaged neighborhoods (h2 = .70) as compared to pre-pubertal girls from the most advantaged neighborhoods (h2 = .34). These findings suggest that neighborhood disadvantage is associated with much earlier increases in genetic influences on disordered eating in girls than are typically observed in more advantaged environments.

Table 4.

Unstandardized path and moderator estimates for the full and best-fitting genotype x environment models

Standardized, Log-Transformed MTP-ED Total Score, With BMI Percentile Regressed Out
Model a c e βxP βyP βzP βxD βyD βzD βxPD βyPD βzPD
Full model .410 (.124, .696) −.386 (−.805, .034) .482 (.347, .616) .547 (−.029, 1.122) .847 (.256, 1.439) −.015 (−.226, .196) .498 (.002, .994) .195 (−.665, 1.055) .064 (−.243, .372) −.720 (−1.711, .271) .031 (−1.079, 1.141) .023 (−.454, .499)
Best-fitting .429 (.268, .589) −.301 (−.520, −.082) .510 (.478, .541) .431 (.112, .750) .901 (.600, 1.203) — .480 (.212, .748) — — −.525 (−1.004, −.045) — —
Open in a new tab

Note: ADI = Area Deprivation Index percentile (higher values indicate greater neighborhood disadvantage); a = additive genetic influences at the lowest levels of the moderators; c = shared environmental influences at the lowest levels of the moderators; e = non-shared environmental influences at the lowest levels of the moderators; βxP, βyP, βzP = coefficients for moderation of genetic/environmental variance by pubertal status; βxD, βyD, βzD = coefficients for moderation of genetic/environmental variance by neighborhood disadvantage; βxPD, βyPD, βzPD = coefficients representing changes in the moderating effects of neighborhood disadvantage across pubertal status (i.e., the disadvantage x puberty interaction). 95% confidence intervals of parameter estimates are included in parentheses. Effects significant at p < .05 are bolded.

Discussion

This is the first study to examine associations between neighborhood disadvantage and disordered eating, and results highlight a potentially critical role for neighborhood disadvantage in the etiology of disordered eating in girls. Girls living in more disadvantaged neighborhoods exhibited significantly greater disordered eating symptoms across all stages of pubertal development, suggesting that neighborhood disadvantage is associated with increased disordered eating even among pre-pubertal girls who are typically at much lower risk (Baker et al., 2012; Klump et al., 2012; Nagl et al., 2016). Genotype x environment interaction models revealed marked differences in the etiology of disordered eating that could underlie these phenotypic associations. Specifically, we replicated pubertal increases in genetic influences on disordered eating in girls (Baker et al., 2012; Harden et al., 2014; Klump et al., 2003, 2007, 2012), but only for girls living in advantaged contexts. When examining girls living in the most disadvantaged neighborhoods, we observed that genetic influences were as great in pre/early-puberty as in late/post-puberty. This could indicate that genetic influences on disordered eating were potentiated much earlier for girls living in disadvantaged contexts, potentially leading to greater and/or earlier disordered eating for girls with genetic risk.

Moving forward, it will be important to identify the “active ingredients” in neighborhood disadvantage that may contribute to etiologic moderation and phenotypic associations with disordered eating. Interestingly, the phenotypic association between neighborhood disadvantage and disordered eating was attenuated after accounting for BMI, while genotype x environment results were unaffected by BMI. This could suggest different mechanisms operating at phenotypic and etiologic levels. At the phenotypic level, food deserts/food swamps in disadvantaged neighborhoods may lead to greater BMI (Cooksey-Stowers et al., 2017; Dubowitz et al., 2012), which could in turn lead to weight stigma, body dissatisfaction, and disordered eating. Indeed, prior research has shown that disadvantage is associated with greater weight stigma among adults (Becker et al., 2017). At an etiologic level, BMI did not substantially alter genotype x environment findings, suggesting that other factors are involved in shifts in genetic influences. These factors may include early life stress that is very common in disadvantaged contexts, including economic stress, educational challenges, limited access to green spaces, and sensory-related (e.g., increased noise pollution) stressors. This early stress may exacerbate genetically-based differences in the physiological stress response (Gillespie et al., 2009) or neural systems involved in emotion processing and stress reactivity (Gard et al., 2020), ultimately leading to earlier expression of individual differences in genetic risk for disordered eating. Within this proposed framework, increased BMI may be a proximal risk factor for phenotypic increases in disordered eating in girls from disadvantaged neighborhoods, while enduring and chronic stress may be a more distal factor that activates genetic influences on overall risk for disordered eating symptoms. Though these hypotheses are speculative, they represent one set of possible mechanisms to be tested in future (ideally longitudinal) research.

Disadvantage takes multiple forms, some relatively distal (e.g., school or neighborhood disadvantage) and others more proximal (e.g., family-level financial disadvantage, food insecurity, lack of health insurance). While this study focused on a more distal form of disadvantage, prior research on food insecurity suggests at least some proximal forms of disadvantage are also associated with disordered eating (Becker et al., 2017, 2019; Coffino et al., 2020; Lydecker et al., 2019). It is currently unknown whether proximal and distal forms of disadvantage show differential associations with disordered eating, or if they interact with genetic/environmental risk in similar ways. Though measures of disadvantage were limited in our dataset, we had one proximal measure of disadvantage (family socioeconomic status (SES), a composite of mother’s education level, father’s education level, and combined parental income) that we could examine in post-hoc phenotypic and twin analyses to begin to address these questions. Note that family SES in our data is coded in the opposite direction as neighborhood disadvantage – that is, lower family SES indicates greater proximal levels of disadvantage.

As shown in Table S6, when examined on its own, lower family SES was associated with significantly greater disordered eating symptoms. When family SES and neighborhood disadvantage were included in the same model, there was no significant interaction between neighborhood disadvantage and family SES, and the association between lower family SES and greater disordered eating symptoms remained significant while greater neighborhood disadvantage remained associated with more disordered eating at a trend level. These results suggest that individuals tend to have the greatest levels of disordered eating when they are high on multiple indices of disadvantage. Our findings from the twin genotype x environment models were broadly consistent with results for neighborhood disadvantage; genetic influences on disordered eating were greater for youth from lower SES (more disadvantaged) families, suggesting that both distal and more proximal forms of disadvantage may be associated with activation of genetic influences (see Tables S7–S8 and Figure S3). However, we also observed some moderation of shared environmental influences by family SES (i.e., lower shared environmental influences among low SES families; see Tables S7–S8 and Figure S3) that we did not observe with neighborhood disadvantage. Taken together, results suggest that key associations between disadvantage and disordered eating may be consistent regardless of how disadvantage is measured (e.g., greater phenotypic symptoms of disordered eating and generally greater genetic influences in more disadvantaged contexts), while others may depend on the specific form of disadvantage examined and may reflect partially distinct etiologic pathways. Additional research that incorporates multiple indices of disadvantage is needed to more fully elucidate the implications of experiencing disadvantage in several domains.

While our study had many strengths (e.g., large population-based sample, a well-validated measure of disadvantage), some key limitations should be noted. We only collected parent reports of disordered eating and BMI. Our results replicated numerous previous findings with self-report measures of disordered eating, including associations with key variables (BMI, pubertal development, internalizing symptoms; Hay et al., 2015; Smink et al., 2014; Thomas et al., 2020), phenotypic increases in disordered eating across puberty (Baker et al., 2012; Klump et al., 2012; Nagl et al., 2016), and increases in genetic influences across puberty in girls from more advantaged contexts (Baker et al., 2012; Harden et al., 2014; Klump et al., 2003, 2007, 2012). Using parent report also decreased concerns about comprehension of disordered eating items among the younger children in our sample, and our measure exhibited good psychometric properties across the full age range. Nevertheless, replication with self-reported symptoms is needed to confirm and replicate results. This is particularly true given that parents may be unaware of the full extent of disordered eating symptoms experienced by youth due to the secrecy often attending such symptoms (Bartholdy et al., 2017). Interestingly, however, parents did not always report lower levels of disordered eating than youth in our validation sample. Specifically, youth reported somewhat greater disordered eating symptoms than their parents did in 37.3% of cases, but parents reported higher levels of disordered eating than youth did in 22.6% of cases (in the remaining 40.1% of cases, parent and child report matched exactly). Thus, it is possible that parent report may, in some cases, capture symptoms that youth themselves fail to report due to stigma or embarrassment.

Prior simulation studies of genotype x environment twin models involving neighborhood disadvantage have found that samples of 1,000 families are well powered to detect small-to-medium sized effects of genetic and non-shared environmental moderation in population-based samples (i.e., not enriched for disadvantage) (Burt et al., 2020). Thus, our sample of approximately 1,500 families was sufficiently powered to detect moderation of A and E. However, results from the simulation study cited above suggest lower power to detect shared environmental moderation in our sample, and we were likely only powered to detect shared environmental moderation of a large effect size. Additional limitations on power may have arisen through zygosity errors due to lack of genotyping (though we note that this is likely a relatively small source of error given that the method we used to determine zygosity is 95% accurate) and the limited response scale (i.e., 0–2) of our disordered eating measure. In genotype x environment models, limited power would have primarily affected the results of model fitting (i.e., the best fitting model may have not retained C moderation when such moderation in fact exists), rather than the results of the full model. This is because the full model reflects the raw data, and allows all parameters to be estimated freely rather than constraining any to zero based on non-significance. Because the full model showed minimal moderation of shared environmental influences by neighborhood disadvantage, the overall pattern of findings and our conclusions were likely not unduly influenced by lower power to detect effects. Nevertheless, replication in larger samples and samples enriched for disadvantage is needed and could reveal whether etiologic moderation (particularly for shared environmental influences) is more pronounced when greater numbers of participants from highly disadvantaged neighborhoods are included.

Our analyses focused exclusively on girls. While disordered eating predominantly affects girls (Hay et al., 2015; Nagl et al., 2016; Smink et al., 2014), boys also experience EDs, and it will be important for future studies to examine whether associations between neighborhood disadvantage and disordered eating differ in boys. Finally, our data were cross-sectional, which prevented us from identifying neighborhood disadvantage as a prospective risk factor. Prospective studies are needed to elucidate the short-term and long-term associations between neighborhood disadvantage and risk for EDs. Additionally, it is difficult to definitively rule out the possibility that factors that covary with neighborhood disadvantage may drive the associations observed in this study without an experimental design.

Despite these limitations, our findings have important implications for treatment and research. It is imperative to screen for EDs among disadvantaged youth, especially among girls with a family history of EDs. Routine screening for EDs is rare in general (Johnston et al., 2007), and perhaps particularly among disadvantaged populations. Relatedly, disadvantaged youth have limited access to ED treatment in the USA (Newacheck et al., 2003; Sonneville & Lipson, 2018). Accessible, affordable treatment options are needed, including for youth traditionally viewed as being at lower risk for EDs (e.g., pre-pubertal girls). Finally, our results demonstrate that the etiology of disordered eating may differ significantly in disadvantaged environments. Etiologic models of disordered eating are incomplete without consideration of the context in which symptoms occur.

Supplementary Material

Supplemental Material

Acknowledgments

A subset of the results reported in this article were previously presented at the 2020 International Conference on Eating Disorders in Sydney, Australia. Study procedures were approved by the Michigan State University Institutional Review Board (protocol LEGACY06-829M). This research was supported by a grant from the National Institute of Mental Health (R01 MH111715-03S1). The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIMH.

Footnotes

1

The Minnesota Eating Behavior Survey (MEBS; previously known as the Minnesota Eating Disorder Inventory [M-EDI]) was adapted and reproduced by special permission of Psychological Assessment Resources, 16204 North Florida Avenue, Lutz, Florida 33549, from the Eating Disorder Inventory (collectively, EDI and EDI-2) by Garner, Olmstead, Polivy, Copyright 1983 by Psychological Assessment Resources. Further reproduction of the MEBS is prohibited without prior permission from Psychological Assessment Resources.

References

  1. Akkermann K, Kaasik K, Kiive E, Nordquist N, Oreland L, & Harro J (2012). The impact of adverse life events and the serotonin transporter gene promoter polymorphism on the development of eating disorder symptoms. Journal of Psychiatric Research, 46(1), 38–43. [DOI] [PubMed] [Google Scholar]
  2. Alvarado SE (2016). Neighborhood disadvantage and obesity across childhood and adolescence: Evidence from the NLSY children and young adults cohort (1986–2010). Social Science Research, 57, 80–98. [DOI] [PubMed] [Google Scholar]
  3. Attar BK, Guerra NG, & Tolan PH (1994). Neighborhood disadvantage, stressful life events and adjustments in urban elementary-school children. Journal of Clinical Child Psychology, 23(4), 391–400. [Google Scholar]
  4. Baker JH, Girdler SS, & Bulik CM (2012). The role of reproductive hormones in the development and maintenance of eating disorders. Expert Review of Obstetrics & Gynecology, 7(6), 573–583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bartholdy S, Allen K, Hodsoll J, O’Daly OG, Campbell IC, Banaschewski T,…Schmidt U (2017). Identifying disordered eating behaviours in adolescents: How do parent and adolescent reports differ by sex and age? European Child & Adolescent Psychiatry, 26(6), 691–701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Becker CB, Middlemass KM, Gomez F, & Martinez-Abrego A (2019). Eating disorder pathology among individuals living with food insecurity: A replication study. Clinical Psychological Science, 7(5), 1144–1158. [Google Scholar]
  7. Becker CB, Middlemass K, Taylor B, Johnson C, & Gomez F (2017). Food insecurity and eating disorder pathology. International Journal of Eating Disorders, 50(9), 1031–1040. [DOI] [PubMed] [Google Scholar]
  8. Border R, & Keller MC (2017). Commentary: Fundamental problems with candidate gene-by-environment interaction studies–reflections on Moore and Thoemmes (2016). Journal of Child Psychology and Psychiatry, 58(3), 328–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brenner AB, Zimmerman MA, Bauermeister JA, & Caldwell CH (2013). Neighborhood context and perceptions of stress over time: An ecological model of neighborhood stressors and intrapersonal and interpersonal resources. American Journal of Community Psychology, 51(3–4), 544–556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bronfenbrenner U, & Ceci SJ (1994). Nature-nurture reconceptualized in developmental perspective: A bioecological model. Psychological Review, 101(4), 568–586. [DOI] [PubMed] [Google Scholar]
  11. Burt A (2011). Some key issues in the study of gene–environment interplay: Activation, deactivation, and the role of development. Research in Human Development, 8(3–4), 192–210. [Google Scholar]
  12. Burt SA (2014). Research review: The shared environment as a key source of variability in child and adolescent psychopathology. Journal of Child Psychology and Psychiatry, 55(4), 304–312. [DOI] [PubMed] [Google Scholar]
  13. Burt SA, Clark DA, Pearson AL, Klump KL, & Neiderhiser JM (2020). Do neighborhood social processes moderate the etiology of youth conduct problems? Psychological Medicine, 50(9), 1519–1529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Burt SA, & Klump KL (2013). The Michigan state university twin registry (MSUTR): An update. Twin Research and Human Genetics, 16(1), 344–350. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Burt SA, & Klump KL (2019). The Michigan State University Twin Registry (MSUTR): 15 years of twin and family research. Twin Research and Human Genetics, 22(6), 741–745 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Burt SA, Klump KL, Kashy DA, Gorman-Smith D, & Neiderhiser JM (2015). Neighborhood as a predictor of non-aggressive, but not aggressive, antisocial behaviors in adulthood. Psychological Medicine, 45(13), 2897–2907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Burt SA, Pearson AL, Carroll S, Klump KL, & Neiderhiser JM (2020). Child antisocial behavior is more environmental in origin in disadvantaged neighborhoods: Evidence across residents’ perceptions and geographic scales in two samples. Journal of Abnormal Child Psychology, 48(2), 265–276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Choi JK, Teshome T, & Smith J (2021). Neighborhood disadvantage, childhood adversity, bullying victimization, and adolescent depression: A multiple mediational analysis. Journal of Affective Disorders, 279, 554–562. [DOI] [PubMed] [Google Scholar]
  19. Coffino JA, Grilo CM, & Udo T (2020). Childhood food neglect and adverse experiences associated with DSM-5 eating disorders in US national sample. Journal of Psychiatric Research, 127, 75–79. [DOI] [PubMed] [Google Scholar]
  20. Cooksey-Stowers K, Schwartz MB, & Brownell KD (2017). Food swamps predict obesity rates better than food deserts in the United States. International Journal of Environmental Research and Public Health, 14(11), 1366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dubowitz T, Ghosh-Dastidar M, Eibner C, Slaughter ME, Fernandes M, Whitsel EA,…Escarce JJ (2012). The Women’s Health Initiative: The food environment, neighborhood socioeconomic status, BMI, and blood pressure. Obesity, 20(4), 862–871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Fairburn CG, & Beglin SJ (1994). Assessment of eating disorders: Interview or self-report questionnaire? International Journal of Eating Disorders, 16(4), 363–370. [PubMed] [Google Scholar]
  23. Fairweather-Schmidt AK, & Wade TD (2017). Weight-related peer-teasing moderates genetic and environmental risk and disordered eating: Twin study. The British Journal of Psychiatry, 210(5), 350–355. [DOI] [PubMed] [Google Scholar]
  24. Gard MC, & Freeman CP (1996). The dismantling of a myth: A review of eating disorders and socioeconomic status. International Journal of Eating Disorders, 20(1), 1–12. [DOI] [PubMed] [Google Scholar]
  25. Gard AM, Maxwell AM, Shaw DS, Mitchell C, Brooks-Gunn J, McLanahan SS,…Hyde LW (2021). Beyond family-level adversities: Exploring the developmental timing of neighborhood disadvantage effects on the brain. Developmental Science, 24(1), e12985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Gillespie CF, Phifer J, Bradley B, & Ressler KJ (2009). Risk and resilience: Genetic and environmental influences on development of the stress response. Depression and Anxiety, 26(11), 984–992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Goodman R (1997). The Strengths and Difficulties Questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38(5), 581–586. [DOI] [PubMed] [Google Scholar]
  28. Gorber SC, Tremblay M, Moher D, & Gorber B (2007). A comparison of direct vs. self-report measures for assessing height, weight and body mass index: A systematic review. Obesity Reviews, 8(4), 307–326. [DOI] [PubMed] [Google Scholar]
  29. Hackman DA, Betancourt LM, Brodsky NL, Hurt H, & Farah MJ (2012). Neighborhood disadvantage and adolescent stress reactivity. Frontiers in Human Neuroscience, 6, 277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Harden KP, Kretsch N, Moore SR, & Mendle J (2014). Descriptive review: Hormonal influences on risk for eating disorder symptoms during puberty and adolescence. International Journal of Eating Disorders, 47(7), 718–726. [DOI] [PubMed] [Google Scholar]
  31. Hart LM, Gordon AR, Sarda V, Calzo JP, Sonneville KR, Samnaliev M, & Austin SB (2020). The association of disordered eating with health-related quality of life in US young adults and effect modification by gender. Quality of Life Research, 1–13. [DOI] [PubMed] [Google Scholar]
  32. Hay P, Girosi F, & Mond J (2015). Prevalence and sociodemographic correlates of DSM-5 eating disorders in the Australian population. Journal of Eating Disorders, 3(1), 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Johnston O, Fornai G, Cabrini S, & Kendrick T (2007). Feasibility and acceptability of screening for eating disorders in primary care. Family Practice, 24(5), 511–517. [DOI] [PubMed] [Google Scholar]
  34. Kind AJ, & Buckingham WR (2018). Making neighborhood-disadvantage metrics accessible—the neighborhood atlas. The New England Journal of Medicine, 378(26), 2456–2458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kind AJ, Jencks S, Brock J, Yu M, Bartels C, Ehlenbach W,…Smith M (2014). Neighborhood socioeconomic disadvantage and 30-day rehospitalization: A retrospective cohort study. Annals of Internal Medicine, 161(11), 765–774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Klump KL, & Burt SA (2006). The Michigan State University Twin Registry (MSUTR): Genetic, environmental and neurobiological influences on behavior across development. Twin Research and Human Genetics, 9(6), 971–977. [DOI] [PubMed] [Google Scholar]
  37. Klump KL, Culbert KM, Slane JD, Burt SA, Sisk CL, & Nigg JT (2012). The effects of puberty on genetic risk for disordered eating: Evidence for a sex difference. Psychological Medicine, 42(3), 627–637. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Klump KL, Fowler N, Mayhall L, Sisk CL, Culbert KM, & Burt SA (2018). Estrogen moderates genetic influences on binge eating during puberty: Disruption of normative processes? Journal of Abnormal Psychology, 127(5), 458–470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Klump KL, McGue M, & Iacono WG (2003). Differential heritability of eating attitudes and behaviors in prepubertal versus pubertal twins. International Journal of Eating Disorders, 33(3), 287–292. [DOI] [PubMed] [Google Scholar]
  40. Klump KL, Perkins PS, Burt SA, McGue M, & Iacono WG (2007). Puberty moderates genetic influences on disordered eating. Psychological Medicine, 37(5), 627–634. [DOI] [PubMed] [Google Scholar]
  41. Koopman-Verhoeff ME, Gredvig-Ardito C, Barker DH, Saletin JM, & Carskadon MA (2020). Classifying pubertal development using child and parent report: Comparing the pubertal development scales to Tanner staging. Journal of Adolescent Health, 66(5), 597–602. [DOI] [PubMed] [Google Scholar]
  42. Le Grange D, & Loeb KL (2007). Early identification and treatment of eating disorders: Prodrome to syndrome. Early Intervention in Psychiatry, 1(1), 27–39. [DOI] [PubMed] [Google Scholar]
  43. Lydecker JA, & Grilo CM (2019). Food insecurity and bulimia nervosa in the United States. International Journal of Eating Disorders, 52(6), 735–739. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Lykken DT, Bouchard TJ, McGue M, & Tellegen A (1990). The Minnesota twin family registry: Some initial findings. Acta Geneticae Medicae et Gemellologiae: Twin Research, 39(1), 35–70. [DOI] [PubMed] [Google Scholar]
  45. Mennis J, Mason M, Light J, Rusby J, Westling E, Way T,…Flay B (2016). Does substance use moderate the association of neighborhood disadvantage with perceived stress and safety in the activity spaces of urban youth? Drug and Alcohol Dependence, 165, 288–292. [DOI] [PubMed] [Google Scholar]
  46. Micali N, Crous-Bou M, Treasure J, & Lawson EA (2017). Association between oxytocin receptor genotype, maternal care, and eating disorder behaviours in a community sample of women. European Eating Disorders Review, 25(1), 19–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Mikhail ME, Culbert KM, Sisk CL, & Klump KL (2019). Gonadal hormone contributions to individual differences in eating disorder risk. Current Opinion in Psychiatry, 32(6), 484–490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Miller A, Pohlig RT, & Reisman DS (2020). Social and physical environmental factors in daily stepping activity in those with chronic stroke. Topics in Stroke Rehabilitation, 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Munn-Chernoff MA, Grant JD, Agrawal A, Koren R, Glowinski AL, Bucholz KK,…Duncan AE (2015). Are there common familial influences for major depressive disorder and an overeating–binge eating dimension in both European American and African American Female twins? International Journal of Eating Disorders, 48(4), 375–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Nagl M, Jacobi C, Paul M, Beesdo-Baum K, Höfler M, Lieb R, & Wittchen HU (2016). Prevalence, incidence, and natural course of anorexia and bulimia nervosa among adolescents and young adults. European Child & Adolescent Psychiatry, 25(8), 903–918. [DOI] [PubMed] [Google Scholar]
  51. Neumark-Sztainer DR, Wall MM, Haines JI, Story MT, Sherwood NE, & van den Berg PA (2007). Shared risk and protective factors for overweight and disordered eating in adolescents. American Journal of Preventive Medicine, 33(5), 359–369. [DOI] [PubMed] [Google Scholar]
  52. Newacheck PW, Hung YY, Jane Park M, Brindis CD, & Irwin CE Jr (2003). Disparities in adolescent health and health care: Does socioeconomic status matter? Health Services Research, 38(5), 1235–1252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. O’Connor SM, Culbert KM, Mayhall LA, Burt SA, & Klump KL (2020). Differences in genetic and environmental influences on body weight and shape concerns across pubertal development in females. Journal of Psychiatric Research, 121, 39–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Peeters H, Van Gestel S, Vlietinck R, Derom C, & Derom R (1998). Validation of a telephone zygosity questionnaire in twins of known zygosity. Behavior Genetics, 28(3), 159–163. [DOI] [PubMed] [Google Scholar]
  55. Petersen AC, Crockett L, Richards M, & Boxer A (1988). A self-report measure of pubertal status: Reliability, validity, and initial norms. Journal of Youth and Adolescence, 17(2), 117–133. [DOI] [PubMed] [Google Scholar]
  56. Powell WR, Buckingham WR, Larson JL, Vilen L, Yu M, Salamat MS,…Kind AJ (2020). Association of neighborhood-level disadvantage with Alzheimer disease neuropathology. JAMA Network Open, 3(6), e207559–e207559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Purcell S (2002). Variance components models for gene–environment interaction in twin analysis. Twin Research and Human Genetics, 5(6), 554–571. [DOI] [PubMed] [Google Scholar]
  58. Roubinov DS, Hagan MJ, Boyce WT, Adler NE, & Bush NR (2018). Family socioeconomic status, cortisol, and physical health in early childhood: the role of advantageous neighborhood characteristics. Psychosomatic Medicine, 80(5), 492–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Rozenblat V, Ong D, Fuller-Tyszkiewicz M, Akkermann K, Collier D, Engels RC,…Krug I (2017). A systematic review and secondary data analysis of the interactions between the serotonin transporter 5-HTTLPR polymorphism and environmental and psychological factors in eating disorders. Journal of Psychiatric Research, 84, 62–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Sheets LR, Kelley LEH, Scheitler-Ring K, Petroski GF, Barnett Y, Barnett C,…Parker JC (2020). An index of geospatial disadvantage predicts both obesity and unmeasured body weight. Preventive Medicine Reports, 18, 101067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Singh GK (2003). Area deprivation and widening inequalities in US mortality, 1969–1998. American Journal of Public Health, 93(7), 1137–1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Smink FR, van Hoeken D, Oldehinkel AJ, & Hoek HW (2014). Prevalence and severity of DSM-5 eating disorders in a community cohort of adolescents. International Journal of Eating Disorders, 47(6), 610–619. [DOI] [PubMed] [Google Scholar]
  63. Sonneville KR, & Lipson SK (2018). Disparities in eating disorder diagnosis and treatment according to weight status, race/ethnicity, socioeconomic background, and sex among college students. International Journal of Eating Disorders, 51(6), 518–526. [DOI] [PubMed] [Google Scholar]
  64. Strachan E, Duncan G, Horn E, & Turkheimer E (2017). Neighborhood deprivation and depression in adult twins: Genetics and gene x environment interaction. Psychological Medicine, 47(4), 627–638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Suisman JL, Burt SA, McGue M, Iacono WG, & Klump KL (2011). Parental divorce and disordered eating: An investigation of a gene-environment interaction. International Journal of Eating Disorders, 44(2), 169–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Thomas KS, Williams MO, & Vanderwert RE (2021). Disordered eating and internalizing symptoms in preadolescence. Brain and Behavior, 11(1), e01904. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. van der Sluis S, Posthuma D, & Dolan CV (2012). A note on false positives and power in GxE modelling of twin data. Behavior Genetics, 42(1), 170–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Vo PT, Racine SE, Burt SA, & Klump KL (2019). Convergence in maternal and child reports of impulsivity, depressive symptoms, and trait anxiety, and their predictive utility for binge-eating behaviors. International Journal of Eating Disorders, 52(9), 1058–1064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. von Ranson KM, Klump KL, Iacono WG, & McGue M (2005). The Minnesota Eating Behavior Survey: A brief measure of disordered eating attitudes and behaviors. Eating Behaviors, 6(4), 373–392. [DOI] [PubMed] [Google Scholar]
  70. Wade TD, & Bulik CM (2018). Genetic influences on eating disorders. In Agras WS & Robinson A (Eds.), Oxford library of psychology. The Oxford handbook of eating disorders (p. 80–105). Oxford University Press. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Material
NIHMS1738258-supplement-Supplemental_Material.pdf (468.4KB, pdf)

RESOURCES