Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Nov 1.
Published in final edited form as: Prev Med. 2022 Sep 3;164:107234. doi: 10.1016/j.ypmed.2022.107234

Associations of abuse and neglect with young adult health in a population-based longitudinal cohort: Findings from Project EAT

Susan M Mason a,*, Rebecca L Emery b, Jessica Friedman a, Stephanie Hanson c, Sydney Johnson a, Dianne Neumark-Sztainer a
PMCID: PMC9951821  NIHMSID: NIHMS1866941  PMID: 36063877

Abstract

Childhood maltreatment (abuse and neglect) is associated with a range of negative outcomes, but a gap remains in understanding of how specific maltreatment types, particularly neglect and non-familial sexual abuse, relate to health and behavior. This study examined the association of neglect and sexual abuse (both familial and non-familial), as well as familial physical and emotional abuse, with: depressive mood and eating disorders; tobacco and marijuana use; and BMI ≥ 25 kg/m2 and BMI ≥ 30 kg/m2 in young adults. Data came from Project EAT (Eating and Activity in Teens and Young Adults), a population-based longitudinal study of weight-related health from adolescence into young adulthood. Maltreatment before age 18 was retrospectively reported at ages 26–33. Risk differences (RDs) and 95% confidence intervals (CIs) were estimated for those with a given maltreatment type to those without, and also for the cumulative number of maltreatment types experienced. One in 3 participants reported abuse or neglect. All maltreatment types were associated with at least one adverse health outcome, with physical abuse being least consistently related to the outcomes. Emotional abuse showed the strongest association with depressive mood. All maltreatment types were associated with eating disorder diagnosis, tobacco use, and marijuana use (except physical abuse for eating disorder). There was little evidence of a maltreatment association with BMI ≥ 25 kg/m2; emotional abuse and neglect were associated with BMI ≥ 30 kg/m2. Prevention of maltreatment needs to be a top public health priority.

Keywords: Child maltreatment, Child abuse, Child neglect, Body weight, Depression, Eating disorders, Tobacco, Marijuana

1. Introduction

At least one in 3 U.S. adults report experiencing neglect and/or physical, sexual, or emotional abuse abuse in childhood (collectively, “childhood maltreatment”) (Merrick et al., 2018). Childhood maltreatment is an important determinant of later-life chronic diseases and premature mortality (Felitti et al., 1998). The mechanisms by which early abuse leads to later-life illness may be partially explained by the impacts of abuse on mental health challenges (e.g., depression and disordered eating), risk behaviors (e.g., tobacco and other drug use), and early markers of chronic disease risk (e.g., higher weight status) (Anda et al., 1999; Dong et al., 2004; Norman et al., 2012; Merrick et al., 2019). Although associations of childhood maltreatment generally, and certain specific types such as familial physical, emotional, and sexual abuse, with later health are well-established, less is known about how neglect and non-familial sexual abuse are related to mental and physical health outcomes. These associations remain especially understudied in young adulthood, a critical inflection point when trajectories of health in those with maltreatment appear to begin to diverge from those without maltreatment (Danese and Tan, 2014; Noll et al., 2007).

The most robust evidence for long-term outcomes of childhood maltreatment comes from the literature on mental health outcomes, including depression and eating disorders. Sexual, physical, and emotional abuse all appear to be associated with depression (Campbell et al., 2016; Font and Maguire-Jack, 2016), with meta-analyses finding that, of these, childhood emotional abuse is most strongly associated with depression (Norman et al., 2012; Humphreys et al., 2020; Infurna et al., 2016; Nelson et al., 2017). Notably, studies have generally not distinguished familial from non-familial sexual abuse, each of which may have distinct associations with depression. With regard to eating disorders, a meta-analysis (Caslini et al., 2016) found that childhood sexual, emotional, and physical abuse are all associated with eating disorder risk. Evidence on neglect as a risk factor for eating disorders remains scant, although non-comparison studies indicate that individuals with an eating disorder report high rates of childhood neglect (Pignatelli et al., 2017). As with other areas of research, studies of maltreatment and eating disorders have generally not disaggregated familial from non-familial sexual abuse (Caslini et al., 2016).

There is long-standing evidence of associations of childhood maltreatment with substance use including use of tobacco and cannabis. For tobacco use and/or dependence, sexual abuse appears to be a risk factor in many studies (Campbell et al., 2016; Font and Maguire-Jack, 2016; Kristman-Valente et al., 2013; Kisely et al., 2020; Mills et al., 2014); these analyses have not disaggregated familial from non-familial sexual abuse. Findings for physical abuse have been less consistent, with some studies showing it to be linked to tobacco use in adolescents generally (Mills et al., 2014; Yoon et al., 2020a), with others finding associations only when physical abuse co-occurs with emotional abuse (Lewis et al., 2019). Cannabis dependence also appears to be related to childhood physical and emotional abuse (Abajobir et al., 2017; Afifi et al., 2012) and sexual abuse (Afifi et al., 2012). Fewer studies have examined neglect as a risk factor for substance use, although a few recent studies have found neglect to predict both cigarette use (Yoon et al., 2020a), marijuana use and dependence (Abajobir et al., 2017; Yoon et al., 2020b), and substance use generally (Dubowitz et al., 2019; Kobulsky et al., 2018).

Finally, childhood maltreatment appears to be a risk factor for higher adult body mass index (BMI) (Danese and Tan, 2014). The original ACE study found physical, sexual, and emotional abuse all to be associated with having a higher BMI (i.e., ≥ 30 kg/m2), but physical abuse showed a stronger association than the other abuse types (Williamson et al., 2002), a finding echoed in subsequent studies (Li et al., 2019; Thomas et al., 2008). However, other studies (Font and Maguire-Jack, 2016) have found sexual abuse to have similar associations with high weight status as physical abuse. These studies have examined combined familial and non-familial sexual abuse, and thus the differential association of these types of abuse with weight status is not known. The associations of neglect with weight have received less attention, although a very early study (Lissau and Sørensen, 1994) found physical neglect at age 10 to predict adult higher weight status, and a meta-analysis of 6 studies echoed this early finding (Danese and Tan, 2014).

An important gap in the existing literature is a lack of separate analyses of familial and non-familial childhood sexual abuse. Theories of trauma and child development, such as betrayal trauma theory (Freyd, 2008), suggest that sexual abuse perpetration by a family member may have more severe consequences than sexual abuse by a non-family member. Although over 65% of childhood sexual abuse is perpetrated by non-family members (Finkelhor and Shattuck, 2012), existing studies of maltreatment and later health have generally not distinguished sexual abuse by perpetrator types. The scant literature on this topic has found mixed results for differential impacts of familial and non-familial sexual abuse, with one study of sexual abuse survivors finding a higher risk of psychopathology when sexual abuse was committed by family than when it was not (Aydin et al., 2015) while another found non-familial sexual abuse was more strongly associated than familial sexual abuse with nicotine dependence among smokers (Cammack et al., 2019).

The goals of the current study are to describe the prevalence and overlap of different maltreatment types and identify their associations with depressive mood, eating disorder diagnoses, tobacco and marijuana use, and higher weight status (BMI ≥ 25 kg/m2 and BMI ≥ 30 kg/m2). We use data from Project EAT (Eating and Activity in Teens and Young Adults), a population-based longitudinal cohort of young adults followed for over 15 years since adolescence to examine differences in these outcomes across subtypes of abuse and neglect. For comparison with prior literature using the ACE Scale or similar risk score measures, we also examined outcomes as a function of cumulative levels of abuse and neglect exposure.

2. Methods

2.1. Participants and procedures

Project EAT, a population-based, longitudinal study of weight-related health among adolescents, began as a cross-sectional survey in 31 public middle and high schools in the Minneapolis-St. Paul metropolitan area of Minnesota in 1998–1999. Surveys were administered by study staff to students in Health, Physical Education, and Science classrooms in either one 90-min class period or two 50-min periods. Consent for participation was obtained through either passive or active consent procedures in accordance with the wishes of the school district. A decision was subsequently made to follow up with participants who had provided sufficient contact information (N = 3672 of 4746). Mailed and online assessments were conducted in 2003–2004 (EAT-II) and 2008–2009 (EAT-III) (Larson et al., 2011; Neumark-Sztainer et al., 2011). In 2015–2016, participants who previously responded to either EAT-II or EAT-III were contacted for additional follow-up in young adulthood (EAT-IV). Follow-up surveys for EAT-IV were collected online, by mail, or by phone from 1830 of the 2270 participants that could be contacted (Larson et al., 2018; Neumark-Sztainer et al., 2018). The Institutional Review Board Human Subjects Committee at the University of Minnesota approved all protocols used in Project EAT.

The current analysis used data from participants who responded to the Project EAT-IV survey (ages 26–33 years; N = 1830) and had non-missing data on their history of abuse and neglect exposure (n = 1782).

2.2. Survey development

The EAT-IV survey was pre-tested as part of two focus groups with young adults (n = 35), and feedback guided refinements. Scale psychometric properties and test-retest reliability of items, reported below, were determined in a subgroup of 103 participants who completed the EAT-IV survey twice within a period of one to four weeks.

2.3. Measures

On Project EAT-IV, participants retrospectively reported on six types of childhood maltreatment occurring prior to age 18 years, including non-familial and familial sexual abuse, and familial physical abuse, emotional abuse, physical neglect, and emotional neglect. Items were adapted from the Minnesota Student Survey (Adelmann et al., 2000) and the Childhood Trauma Questionnaire (Bernstein et al., 1994). All abuse and neglect measures had test-retest reliabilities ≥82% except physical neglect (test-retest agreement = 69%); thus, physical neglect was not included in analyses. Childhood maltreatment variables were dichotomized at cut-points based on the Adverse Childhood Experiences (ACE) scale (Merrick et al., 2019) as described below. Details on survey questions, response options, dichotomization cut-points, and psychometrics are in Table 1.

Table 1.

Child maltreatment variables, source survey items, response options, cut-points, and psychometrics.

Child maltreatment variable Survey question Response options Cut-point Test-retest agreement
Non-familial sexual abuse “Did someone not in your family touch you in a sexual way against your wishes or force you to touch them in a sexual way?” “No,” “once,” or “more than once” No versus once or more than once 99%
Familial sexual abuse “Did someone in your family touch you in a sexual way against your wishes or force you to touch them in a sexual way?” “No,” “once,” or “more than once” No versus once or more than once 99%
Familial physical abuse “An adult in my family hit me so hard it left bruises or marks” “Never,” “rarely,” “sometimes,” “often,” and “very often” Never versus rarely, sometimes, often, or very often 90%
Familial emotional abuse “An adult in my family said hurtful or insulting things to me” “Never,” “rarely,” “sometimes,” “often,” and “very often” Never, rarely, or sometimes versus often or very often 93%
Familial emotional neglect “My family was a source of strength and support” and “someone in my family made me feel that I was important or special” “Never,” “rarely,” “sometimes,” “often,” and “very often” Sometimes, often, or very often versus rarely or never. 82%
Open in a new tab

For comparison to prior literature, a cumulative childhood maltreatment score was calculated from the count of dichotomized abuse and neglect exposures reported (total range 0–6; upper categories were collapsed to maintain cell size, such that the final variable levels were categorized as 0, 1, 2, 3+).

Depressive Mood was assessed with the 6-item Kandel-Davies scale, asking how much participants were bothered by the following over the prior 12 months (Kandel, 1982): feeling too tired to do things; having trouble going to or staying asleep; feeling unhappy, sad, or depressed; feeling hopeless about the future; feeling nervous or tense; and worrying too much about things. Response options were “not at all” (1 point), “somewhat” (2 points), or “very much” (3 points). Responses were averaged across scale items for each participant and multiplied by 10 (range 10 to 30). Cronbach’s α = 0.84. This value was dichotomized at the suggested cut-point of ≥23, indicating clinically-significant depressive mood.

Eating Disorder Diagnosis was assessed by asking participants to indicate the current and previous types of eating disorders they had been diagnosed with. Participants who indicated any of the response options (“anorexia nervosa,” “bulimia nervosa,” “binge eating disorder,” “eating disorder not otherwise specified [EDNOS],” and “other”) were coded as having had an eating disorder diagnosis. Test-retest agreement ranged from 75% for EDNOS to 100% for anorexia and bulimia nervosa.

Cigarette and Marijuana Use were assessed by asking participants how often they used cigarettes or marijuana during the past year. Response options were “never,” “a few times,” “monthly,” and “daily.” Participants were categorized as using tobacco or marijuana if they reported any use in the past year. Spearman rank-order test-retest correlations for cigarette and marijuana use were 0.87 and 0.85, respectively.

Weight Status was obtained from BMI, calculated using the standard formula (weight in kilograms divided by height in meters squared) using self-reported height and weight obtained at EAT-IV. In a substudy of 125 participants from this cohort, the validity of self-reported BMI relative to measured BMI was found to be excellent (r = 0.95 for males and r = 0.98 for women) (Quick et al., 2013). Weight status was categorized according to standard BMI cutoffs of ≥25 and ≥ 30 kg/m2.

Analyses adjusted for sociodemographic factors that may be common correlates of both maltreatment and the health outcomes of interest, including: age in years; ethnicity/race (White, Black, Hispanic, and other); self-reported sex (male, female); and childhood socioeconomic status (SES), based on the highest level of parent education at baseline, with missing or implausible values imputed using additional self-reported data such as receipt of public assistance (high school education or less vs. some college or more) (Sherwood et al., 2009). Finally, because baseline BMI could be both a confounder due to its association with the outcomes and/or a mediator if affected by childhood maltreatment, preliminary models were run with and without BMI. BMI adjustment made only a small difference in the estimates, and the final models include BMI.

2.4. Data analysis

We first examined the prevalence of each type of abuse and neglect in the total sample and by sociodemographic characteristics. We then examined the extent to which different maltreatment types are more likely to be experienced singularly or in combination with other types. Finally, we estimated risk differences for each health outcome, comparing those with a given type of maltreatment to those without that type of maltreatment. We chose to estimate difference measures because they are most appropriate for estimating public health impact and etiology (Hernan and Robins, 2020). We used the “Margins” macro in SAS 9.4 (Cary, NC) to fit separate logistic regression models for each health outcome as a function of each exposure and all covariates. The macro uses the resulting parameter estimates to predict the risk of the outcome in those with versus without exposure, at all levels of the specified covariates. These predicted risks are then used to estimate the average risk difference comparing those with versus without the specified exposure in the sample population, given the covariate distribution. To account for attrition from the original sample, all models included inverse probability weights to adjust for the baseline attributes related to attrition, including ethnicity/race, socioeconomic status, and key outcomes of the Project EAT study, such as dieting and weight status (Little, 1986). Given growing concerns about the harms of null hypothesis significance testing (Lash, 2017), we focus on effect estimation rather than testing in the presentation of our findings.

Because of prior evidence of sex differences in the impacts of abuse, we examined the interaction of sex with each form of maltreatment. We found limited evidence of modification by sex, and present main results for the combined sample. In cases where the interaction was significant, we present stratified results in a supplemental table.

3. Results

3.1. Distribution of maltreatment types by sociodemographic characteristics

Over 30% of the sample reported one or more types of abuse and neglect. Physical abuse was the most common type of maltreatment reported (Table 2). females were more likely than males to report emotional abuse and both familial sexual abuse and non-familial sexual abuse. Black and Hispanic participants and participants whose parents did not have a high school degree were far more likely than White participants and participants with greater parent education to report one or more types of childhood maltreatment.

Table 2.

Study sample demographics and baseline BMI by type of maltreatment exposure, Project EAT (n = 1782).

Variables Familial sexual abuse
 Non-familial sexual abuse
 Familial physical abuse
 Familial emotional abuse
 Emotional neglect
No
 Yes
 No
 Yes
 No
 Yes
 No
 Yes
 No
 Yes
n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD) n (row %) or mean (SD)
Total sample n (%) 1688 (94.7%) 94 (5.3%) 1644 (92.3%) 138 (7.7%) 1410 (79.1%) 372 (20.9%) 1608 (90.3%) 174 (9.7%) 1564 (87.7%) 218 (12.3%)
Age at EAT
 IV 31.1 (1.7) 30.7 (1.8) 31.0 (1.7) 31.0 (1.6) 31.1 (1.6) 30.9 (1.9) 31.1 (1.7) 30.7 (1.8) 31.1 (1.6) 30.8 (2.0)
BMI at baseline 22.4 (4.4) 22.9 (4.6) 22.4 (4.4) 22.8 (4.7) 22.4 (4.3) 22.7 (5.1) 22.3 (4.3) 23.4 (5.6) 22.4 (4.2) 22.7 (6.2)
Sex
 Male 888 (97.3%) 25 (2.7%) 875 (95.9%) 38 (4.1%) 715 (78.4%) 197 (21.6%) 843 (92.4%) 69 (7.6%) 805 (88.2%) 107 (11.8%)
 Female 796 (92.0%) 69 (8.0%) 765 (88.4%) 100 (11.6%) 690 (79.7%) 175 (20.3%) 761 (87.9%) 104 (12.1%) 755 (87.2%) 111 (12.8%)
Ethnicity/race
 White 841 (97.4%) 23 (2.6%) 812 (94%) 52 (6%) 768 (88.9%) 96 (11.1%) 793 (88.9%) 71 (8.2%) 799 (92.5%) 64 (7.5%)
 Black 290 (92.9%) 22 (7.1%) 287 (91.9%) 25 (8.1%) 235 (75.3%) 77 (24.7%) 288 (75.3%) 24 (7.8%) 271 (87%) 41 (13%)
 Hispanic 93 (90.3%) 10 (9.7%) 82 (79.7%) 21 (20.3%) 67 (65.7%) 35 (34.3%) 91 (65.7%) 12 (11.3%) 92 (89.7%) 11 (10.3%)
 Other 441 (92%) 38 (8%) 440 (91.7%) 40 (8.3%) 317 (66.2%) 162 (33.8%) 415 (66.2%) 65 (13.5%) 377 (78.7%) 102 (21.3%)
Parent education
 High school + 1039 (95.1%) 53 (4.9%) 1025 (93.8%) 68 (6.2%) 929 (85.1%) 163 (14.9%) 1004 (91.9%) 89 (8.1%) 997 (91.2%) 96 (8.8%)
 <High school 580 (94%) 37 (6%) 551 (89.2%) 67 (10.8%) 429 (69.5%) 189 (30.5%) 543 (87.9%) 75 (12.1%) 517 (83.7%) 101 (16.3%)
Open in a new tab

3.2. Occurrence of maltreatment types singularly and in combination with other types

Individuals with physical abuse were least likely to report additional maltreatment types (over 45% of individuals reporting physical abuse reported no other maltreatment type; Fig. 1). In contrast, only 18% of those with emotional abuse reported it as the only type of maltreatment. Familial sexual abuse was also highly overlapping with other types of maltreatment, with 50% of individuals reporting familial sexual abuse also reporting 2 or more other types of maltreatment.

Fig. 1.

Fig. 1.

Open in a new tab

Percent of individuals with each maltreatment type across maltreatment score categories (1, 2, or 3+ maltreatment types).

3.3. Associations of individual maltreatment types with health outcomes

All maltreatment types, except familial sexual abuse, were linked to depression (Fig. 2a, Table 3a). Associations ranged from a 6 percentage-point greater risk in those with physical abuse (95% CI: 1%, 12%) to a 19 percentage-point greater risk in those with emotional abuse (95% CI: 12%, 27%). For eating disorder diagnosis, RDs ranged from 4 percentage points (95% CI: −0.01, 0.08) for physical abuse, to 10 percentage points for familial sexual abuse (95% CI: 0.01, 0.18).

Fig. 2.

Fig. 2.

Open in a new tab

Risk differences for associations of each type of maltreatment with (a) mental health, (b) substance use, and (c) weight status outcomes adjusted for age, socioeconomic status, sex, ethnicity/race, and baseline BMI.

Table 3a.

Adjusted* risk differences for mental health outcomes for those with a given maltreatment exposure relative to those without the exposure, Project EAT (n = 1782).

Mental health outcomes
Maltreatment exposure Depressive mood Any eating disorder diagnosis (current or past)
N (row %) with outcome RD (95% CI) N (row %) with outcome RD (95% CI)
Maltreatment score 0 136 (11.6%) REF – 53 (4.5%) REF –
1 80 (22.2%) 0.08 (0.03, 0.14) 15 (4%) 0.01 (− 0.03, 0.04)
2 44 (30.8%) 0.19 (0.10, 0.27) 28 (19.7%) 0.14 (0.06, 0.22)
3 31 (29.7%) 0.12 (0.03, 0.22) 13 (12.5%) 0.08 (0.00, 0.16)
Familial sexual abuse No 262 (15.5%) REF – 92 (5.5%) REF –
Yes 29 (31.2%) 0.01 (− 0.07, 0.10) 16 (17.1%) 0.10 (0.01, 0.18)
Non-familial sexual abuse No 252 (15.3%) REF – 86 (5.2%) REF –
Yes 39 (28.7%) 0.10 (0.02, 0.18) 23 (16.4%) 0.08 (0.02, 0.15)
Familial physical abuse No 213 (15.1%) REF – 78 (5.6%) REF –
Yes 78 (21%) 0.06 (0.01, 0.12) 30 (8.1%) 0.04 (− 0.01, 0.08)
Familial emotional abuse No 226 (14.1%) REF – 83 (5.2%) REF –
Yes 65 (37.3%) 0.19 (0.12, 0.27) 25 (14.4%) 0.09 (0.03, 0.15)
Emotional neglect No 228 (14.6%) REF – 82 (5.2%) REF –
Yes 62 (28.7%) 0.12 (0.05, 0.18) 27 (12.3%) 0.07 (0.01, 0.12)
Open in a new tab
*

Adjusted for age, socioeconomic status, sex, ethnicity/race, and baseline BMI

Tobacco and marijuana use also showed strong associations with maltreatment exposures (Fig. 2b, Table 3b). For smoking, RDs ranged from 13 percentage points for emotional neglect (95% CI: 0.05, 0.21) to 23 percentage-points for non-familial sexual abuse (95% CI: 0.14, 0.32). For marijuana use, RDs ranged from 8 percentage points for familial sexual abuse (95% CI: −0.03, 0.20) to 16 percentage points for non-familial sexual abuse (95% CI: 0.07, 0.26).

Table 3b.

Adjusted* risk differences for substance use outcomes for those with a given maltreatment exposure relative to those without the exposure, Project EAT (n = 1782).

Substance use outcomes
Maltreatment exposure Cigarettes Marijuana
N (row %) with outcome RD (95% CI) N (row %) with outcome RD (95% CI)
Maltreatment score 0 281 (24%) REF – 206 (17.8%) REF –
1 116 (32.3%) 0.13 (0.07, 0.19) 97 (26.8%) 0.12 (0.06, 0.18)
2 55 (38.6%) 0.23 (0.13, 0.32) 43 (30.5%) 0.22 (0.13, 0.31)
3 41 (40.1%) 0.23 (0.12, 0.34) 19 (19.3%) 0.13 (0.02, 0.24)
Familial sexual abuse No 456 (27.1%) REF – 346 (20.7%) REF –
Yes 37 (39.7%) 0.19 (0.08, 0.31) 20 (21.7%) 0.08 (− 0.03, 0.20)
Non-familial sexual abuse No 435 (26.6%) REF – 328 (20.2%) REF –
Yes 58 (42%) 0.23 (0.14, 0.32) 38 (27.8%) 0.16 (0.07, 0.26)
Familial physical abuse No 352 (25.1%) REF – 276 (19.8%) REF –
Yes 141 (38.2%) 0.17 (0.11, 0.23) 90 (24.5%) 0.09 (0.03, 0.15)
Familial emotional abuse No 430 (26.8%) REF – 322 (20.2%) REF –
Yes 63 (36.9%) 0.14 (0.05, 0.22) 44 (25.6%) 0.12 (0.04, 0.20)
Emotional neglect No 417 (26.8%) REF – 308 (19.9%) REF –
Yes 76 (35.2%) 0.13 (0.05, 0.21) 58 (27%) 0.14 (0.06, 0.21)
Open in a new tab
*

Adjusted for age, socioeconomic status, sex, ethnicity/race, and baseline BMI.

Weight status showed far more muted associations (Fig. 2c, Table 3c). For BMI ≥ 25 kg/m2, RDs were close to the null for non-familial sexual abuse, physical abuse, and emotional abuse. Emotional neglect was associated with a 7 percentage-point increase in risk of BMI ≥ 25 kg/m2 (95% CI: 0.00, 0.14), while familial sexual abuse was associated with a 9 percentage-point increase in risk, although the confidence interval was very wide (95% CI: −0.02, 0.20). The pattern for obesity was quite different: Emotional abuse and neglect were the only maltreatment types associated with elevated obesity risks, with 12 percentage-point (95% CI: 0.05, 0.20), and 11 percentage-point (95% CI: 0.04, 0.18) increases in risk, respectively.

Table 3c.

Adjusted* risk differences for weight status outcomes for those with a given maltreatment exposure relative to those without the exposure, Project EAT (n = 1782).

Weight status outcomes
Maltreatment exposure BMI ≥ 25 kg/m2 BMI ≥ 30 kg/m2
N (row %) with outcome RD (95% CI) N (row %) with outcome RD (95% CI)
Maltreatment score 0 687 (61.7%) REF – 301 (27%) REF –
1 222 (64.9%) 0.01 (− 0.05, 0.07) 114 (33.5%) 0.04 (− 0.01, 0.09)
2 89 (68.6%) 0.04 (− 0.05, 0.13) 50 (38.4%) 0.04 (− 0.04, 0.12)
3+ 63 (65.9%) 0.06 (− 0.04, 0.16) 38 (40.5%) 0.11 (0.01, 0.21)
Familial sexual abuse No 996 (62.6%) REF – 467 (29.4%) REF –
Yes 64 (73.2%) 0.09 (− 0.02, 0.20) 36 (40.8%) 0.02 (− 0.07, 0.11)
Non-familial sexual abuse No 979 (63.2%) REF – 459 (29.6%) REF –
Yes 81 (62.4%) 0.03 (− 0.06, 0.11) 44 (34.3%) 0.01 (− 0.06, 0.09)
Familial physical abuse No 832 (62.4%) REF – 388 (29.1%) REF –
Yes 228 (65.9%) 0.00 (− 0.06, 0.06) 115 (33.4%) 0.02 (− 0.03, 0.07)
Familial emotional abuse No 953 (62.6%) REF – 434 (28.5%) REF –
Yes 107 (68.4%) 0.03 (− 0.04, 0.11) 69 (44.3%) 0.12 (0.05, 0.20)
Emotional neglect No 930 (62.7%) REF – 425 (28.6%) REF –
Yes 130 (66.4%) 0.07 (0.00, 0.14) 78 (40%) 0.11 (0.04, 0.18)
Open in a new tab
*

Adjusted for age, socioeconomic status, sex, ethnicity/race, and baseline BMI.

3.4. Associations of cumulative maltreatment with health outcomes

There were elevated risks of depressive mood and eating disorder diagnosis (Table 3a), as well as of cigarette smoking and marijuana use (Table 3b), associated with 1, 2, or 3+ maltreatment types relative to none. However, there was no apparent further increase associated with going from 2 to 3+ types. For weight status (Table 3c), associations were more muted, with the largest increase in risk for BMI ≥ 30 kg/m2 among those with 3+ maltreatment types.

Results of sex–maltreatment interaction analyses suggested that the exposure-outcome associations were largely similar for males and females. In the small number of cases where a difference was detected, males tended to have a stronger association (Supplemental Table 1). Specifically, males showed stonger associations of non-familial sexual abuse with depression and BMI ≥30, of emotional abuse with depression and marijuana, and of emotional neglect and the maltreatment score with depression. Emotional abuse had a stronger relationship with BMI ≥30 in females than in males.

4. Discussion

Over 30% of this cohort reported one or more types of maltreatment, with those from more marginalized socioeconomic and ethnic/racial backgrounds reporting the highest prevalence. This overall prevalence is lower than in prior studies, which have found more than half of individuals have experienced one or more ACEs (Merrick et al., 2018). These differences may be explained by the fact that we focused only on abuse and neglect, and did not capture other common ACEs such as household dysfunction. We also used a more conservative cut-off for emotional abuse, resulting in a prevalence in our sample of 10% versus 34% in the BRFSS analysis. Exposure to most types of maltreatment in childhood was strongly associated with depressive mood, eating disorder diagnosis, cigarette smoking, and marijuana use, consistent with a large body of literature showing that maltreatment and other ACEs are strongly predictive of mental health and health behaviors later in life (Merrick et al., 2019). The modest associations found for BMI ≥ 25 kg/m2 and BMI ≥ 30 kg/m2 are in contrast to two meta-analyses in 2013 finding that exposure to abuse (physical or sexual) in childhood was associated with 1.4 times the risk of BMI ≥ 30 kg/m2 in adulthood and that physical abuse was associated with 1.5 times the risk of BMI ≥ 30 kg/m2 (Danese and Tan, 2014; Hemmingsson et al., 2014). It is possible that the differences in results between our study and the prior literature reflects the relatively young age of our sample; prior work has suggested that abuse-related differences in weight status emerge in young adulthood or later (Danese and Tan, 2014; Noll et al., 2007).

We found that non-familial sexual abuse, emotional abuse, and emotional neglect appeared to be most strongly and consistently associated with the range of outcomes examined. Our identification of emotional abuse as an important predictor of later health is consistent with previous findings in similar cohorts (Mason et al., 2015) with respect to weight status and eating behaviors. Emotional neglect has been less frequently examined in regard to later life health outcomes, but a large body of work indicates that emotional neglect has a deleterious effect on child development (Power et al., 2020); our findings indicate that these early developmental effects may have a persistent impact on later life mental health and behavior. The emergence of non-familial sexual abuse as a strong predictor of the health outcomes studied was more surprising. We would have anticipated that sexual abuse by a family perpetrator would have proved a stronger predictor of negative outcomes, but a notably stronger association of familial sexual abuse was only seen for BMI ≥ 25 kg/m2, although the wide confidence intervals overlapped with estimates for other types of maltreatment. It will be important to examine both familial and non-familial sexual abuse in other samples to see if this finding replicates before drawing conclusions about which type of abuse is more impactful for later health outcomes.

It is notable that, for eating disorders and substance use, there appear to be fewer differences in associations across different types of maltreatment than were observed with depression and weight status. Depression and BMI ≥ 30 kg/m2 both show substantial elevations in risk specifically for emotional abuse, suggesting that there may be a particular mechanism by which emotional insults in childhood affect weight via depressed mood. Together, these findings possibly indicate that, while eating disorders and substance use may be coping strategies used in response to a broad array of stressors, changes in weight status may be specific to the sustained impact of emotional abuse on psychological distress.

Strengths of this study include its longitudinal design, large sample size, and selection from a non-clinical community sample. The examination of multiple types of child abuse and neglect, including less-examined measures such as non-familial sexual abuse and emotional neglect, contributes to a growing understanding of how different maltreatment types may shape later-life health. The range of health outcomes included also provides a more holistic picture of the adult well-being of individuals with a history of abuse and/or neglect.

Limitations of this study include the attrition of the sample over longitudinal follow-up. Inverse probability weighting used in these analyses was designed to make the findings generalizable to the original baseline, population-based sample with regard to key sociodemographic measures (e.g., sex, ethnicity/race, parent education). If attrition is correlated with both the study exposures of childhood maltreatment and with the health outcomes of interest, then there may be selection bias in exposure-outcome associations. A second potential limitation is that retrospective self-reports of physical and sexual abuse may suffer from recall bias; however, retrospective recall is often the only feasible way to assess abuse and neglect in large epidemiologic cohorts. Further, is not clear the extent to which retrospective report is less accurate than as prospective assessments, which also suffer from under-reporting and measurement error (Baldwin et al., 2019).

In summary, our study adds to a body of work identifying potential long-lasting impacts of childhood abuse and neglect. Key contributions of our study include the identification of non-familial sexual abuse and emotional neglect as important predictors of mental and behavioral health outcomes. It is notable that Child Protective Services, the institution primarily tasked with addressing child maltreatment, intervenes most often in cases of physical neglect and physical abuse (Child Welfare Information Gateway. Child Maltreatment 2019: Summary of Key Findings. U.S. Department of Health and Human Services, 2021). Although these types of maltreatment may pose immediate harms to children, our study findings suggest a disconnect between the priorities of Child Protection and potential long-term implications of maltreatment exposures such as emotional abuse and neglect. Our findings emphasize the need to prevent maltreatment. Recent studies have found that economic policies may be promising public health tools for maltreatment prevention (Berger et al., 2017; Ginther, n.d.; Raissian and Bullinger, 2017); these types of population-based approaches warrant additional research and action. In cases where maltreatment cannot be prevented, it will be important to develop and test trauma-informed treatments to mitigate the potential effects of early trauma on poor health.

Supplementary Material

table

Funding

Data collection for the study was supported by Grant Number R01HL116892 from the National Heart, Lung, and Blood Institute (PI: Dianne Neumark-Sztainer). The authors’ time to conduct and describe the analysis reported within this manuscript was supported by Grant Number R35HL139853 from the National Heart, Lung, and Blood Institute (PI: Dianne Neumark-Sztainer) and by Grant Number R01HD090053 (PI: Susan Mason) through the National Institute for Child Health and Human Development. Rebecca Emery’s time was supported by the National Center for Advancing Translational Sciences under TL1 TR002493 (PI: Fulkerson) and UL1 TR002494 (PI: Blazar). Dr. Mason also acknowledges support from the Minnesota Population Center (Grant Number P2CHD041023 from the National Institute for Child Health and Human Development). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute, the National Institute for Child Health and Human Development, the National Center for Advancing Translational Sciences, or the National Institutes of Health.

Footnotes

CRediT authorship contribution statement

Susan M. Mason: Conceptualization, Methodology, Software, Formal analysis, Writing – original draft. Rebecca L. Emery: Methodology, Writing – original draft, Writing – review & editing. Jessica Friedman: Methodology, Writing – original draft, Writing – review & editing. Stephanie Hanson: Methodology, Writing – original draft, Writing – review & editing. Sydney Johnson: Methodology, Writing – original draft, Writing – review & editing. Dianne Neumark-Sztainer: Writing – review & editing, Funding acquisition.

Declaration of Competing Interest

The authors have no conflicts of interest or financial disclosures to report.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ypmed.2022.107234.

Data availability

Data will be made available on request.

References

  1. Abajobir AA, Najman JM, Williams G, Strathearn L, Clavarino A, Kisely S, 2017. Substantiated childhood maltreatment and young adulthood cannabis use disorders: a pre-birth cohort study. Psychiatry Res 256, 21–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Adelmann PK, Harrison P, Hedger S, 2000. 1998 Minnesota Student Survey: Risks, Protectors, and Preteen Behaviors. Minnesota Department of Human Services Performance Measurement & Quality Improvement Division, St. Paul, MN [Google Scholar]
  3. Afifi TO, Henriksen CA, Asmundson GJG, Sareen J, 2012. Childhood maltreatment and substance use disorders among men and women in a nationally representative sample. Can J Psychiatry Rev Can Psychiatr 57 (11), 677–686. [DOI] [PubMed] [Google Scholar]
  4. Anda RF, Croft JB, Felitti VJ, Nordenberg D, Giles WH, Williamson DF, et al. , 1999. Adverse childhood experiences and smoking during adolescence and adulthood. JAMA 282 (17), 1652–1658. [DOI] [PubMed] [Google Scholar]
  5. Aydin B, Akbas S, Turla A, Dundar C, Yuce M, Karabekiroglu K, 2015. Child sexual abuse in Turkey: an analysis of 1002 cases. J. Forensic Sci 60 (1), 61–65. [DOI] [PubMed] [Google Scholar]
  6. Baldwin JR, Reuben A, Newbury JB, Danese A, 2019. Agreement between prospective and retrospective measures of childhood maltreatment: a systematic review and meta-analysis. JAMA Psychiatry 76 (6), 584–593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Berger LM, Font SA, Slack KS, Waldfogel J, 2017. Income and child maltreatment in unmarried families: evidence from the earned income tax credit. Rev. Econ. Househ 15 (4), 1345–1372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bernstein DP, Fink L, Handelsman L, Foote J, Lovejoy M, Wenzel K, et al. , 1994. Initial reliability and validity of a new retrospective measure of child abuse and neglect. Am. J. Psychiatry 151 (8), 1132–1136. [DOI] [PubMed] [Google Scholar]
  9. Cammack AL, Haardörfer R, Suglia SF, 2019. Associations between child maltreatment, cigarette smoking, and nicotine dependence in young adults with a history of regular smoking. Ann. Epidemiol 40, 13–20.e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Campbell JA, Walker RJ, Egede LE, 2016. Associations between adverse childhood experiences, high-risk Behaviors, and morbidity in adulthood. Am. J. Prev. Med 50 (3), 344–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Caslini M, Bartoli F, Crocamo C, Dakanalis A, Clerici M, Carrà G, 2016. Disentangling the association between child abuse and eating disorders: a systematic review and meta-analysis. Psychosom. Med 78 (1), 79–90. [DOI] [PubMed] [Google Scholar]
  12. Child Welfare Information Gateway. Child Maltreatment 2019: Summary of Key Findings. U.S. Department of Health and Human Services, 2021. Administration for Children and Families, Children’s Bureau
  13. Danese A, Tan M, 2014. Childhood maltreatment and obesity: systematic review and meta-analysis. Mol. Psychiatry 19 (5), 544–554. [DOI] [PubMed] [Google Scholar]
  14. Dong M, Giles WH, Felitti VJ, Dube SR, Williams JE, Chapman DP, et al. , 2004. Insights into causal pathways for ischemic heart disease: adverse childhood experiences study. Circulation 110 (13), 1761–1766. [DOI] [PubMed] [Google Scholar]
  15. Dubowitz H, Roesch S, Arria AM, Metzger R, Thompson R, Kotch JB, et al. , 2019. Timing and chronicity of child neglect and substance use in early adulthood. Child Abuse Negl 94, 104027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, et al. , 1998. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. Am. J. Prev. Med 14 (4), 245–258. [DOI] [PubMed] [Google Scholar]
  17. Finkelhor D, Shattuck A, 2012. Characteristics of Crimes against Juveniles [Internet] Available from: http://www.unh.edu/ccrc/pdf/CV26_Revised%20Characteristics%20of%20Crimes%20against%20Juveniles_5-2-12.pdf.
  18. Font SA, Maguire-Jack K, 2016. Pathways from childhood abuse and other adversities to adult health risks: the role of adult socioeconomic conditions. Child Abuse Negl 51, 390–399. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Freyd JJ, 2008. Betrayal Trauma. In: Reyes G, Elhai JD, Ford JD (Eds.), Encyclopedia of Psychological Trauma John Wiley and Sons, New York, p. 76. [Google Scholar]
  20. Ginther DK, 2022. Do State TANF Policies Affect Child Abuse and Neglect?, 34. [Google Scholar]
  21. Hemmingsson E, Johansson K, Reynisdottir S, 2014. Effects of childhood abuse on adult obesity: a systematic review and meta-analysis. Obes. Rev 15 (11), 882–893. [DOI] [PubMed] [Google Scholar]
  22. Hernan M, Robins J, 2020. Causal Inference: What if Chapman & Hall, Boca Raton. [Google Scholar]
  23. Humphreys KL, LeMoult J, Wear JG, Piersiak HA, Lee A, Gotlib IH, 2020. Child maltreatment and depression: a meta-analysis of studies using the childhood trauma questionnaire. Child Abuse Negl 102, 104361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Infurna MR, Reichl C, Parzer P, Schimmenti A, Bifulco A, Kaess M, 2016. Associations between depression and specific childhood experiences of abuse and neglect: a meta-analysis. J. Affect. Disord 15 (190), 47–55. [DOI] [PubMed] [Google Scholar]
  25. Kandel DB, 1982. Epidemiology of depressive mood in adolescents: an empirical study. Arch. Gen. Psychiatry 39 (10), 1205. [DOI] [PubMed] [Google Scholar]
  26. Kisely S, Abajobir AA, Mills R, Strathearn L, Clavarino A, Gartner C, et al. , 2020. Child maltreatment and persistent smoking from adolescence into adulthood: a birth cohort study. Nicotine Tob Res Off J Soc Res Nicotine Tob 22 (1), 66–73. [DOI] [PubMed] [Google Scholar]
  27. Kobulsky JM, Yoon S, Bright CL, Lee G, Nam B, 2018. Gender-moderated pathways from childhood abuse and neglect to late-adolescent substance use. J. Trauma. Stress 31 (5), 654–664. [DOI] [PubMed] [Google Scholar]
  28. Kristman-Valente AN, Brown EC, Herrenkohl TI, 2013. Child physical and sexual abuse and cigarette smoking in adolescence and adulthood. J Adolesc Health Off Publ Soc Adolesc Med 53 (4), 533–538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Larson N, Neumark-Sztainer D, Story M, van den Berg P, Hannan PJ, 2011. Identifying correlates of young adults’ weight behavior: survey development. Am. J. Health Behav 35 (6), 712–725. [PMC free article] [PubMed] [Google Scholar]
  30. Larson N, Chen Y, Wall M, Winkler MR, Goldschmidt AB, Neumark-Sztainer D, 2018. Personal, behavioral, and environmental predictors of healthy weight maintenance during the transition to adulthood. Prev. Med 113, 80–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lash TL, 2017. The harm done to reproducibility by the culture of null hypothesis significance testing. Am. J. Epidemiol 186 (6), 627–635. [DOI] [PubMed] [Google Scholar]
  32. Lewis T, Kotch J, Proctor L, Thompson R, English D, Smith J, et al. , 2019. The role of emotional abuse in youth smoking. Am. J. Prev. Med 56 (1), 93–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Li L, Pinto Pereira SM, Power C, 2019. Childhood maltreatment and biomarkers for cardiometabolic disease in mid-adulthood in a prospective British birth cohort: associations and potential explanations. BMJ Open 9 (3), e024079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lissau I, Sørensen TI, 1994. Parental neglect during childhood and increased risk of obesity in young adulthood. Lancet Lond Engl 343 (8893), 324–327. [DOI] [PubMed] [Google Scholar]
  35. Little RJA, 1986. Survey nonresponse adjustments for estimates of means. Int Stat Rev Rev Int Stat 54 (2), 139–157. [Google Scholar]
  36. Mason SM, MacLehose RF, Katz-Wise SL, Austin SB, Neumark-Sztainer D, Harlow BL, et al. , 2015. Childhood abuse victimization, stress-related eating, and weight status in young women. Ann. Epidemiol 25 (10), 760–766.e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Merrick MT, Ford DC, Ports KA, Guinn AS, 2018. Prevalence of adverse childhood experiences from the 2011–2014 behavioral risk factor surveillance system in 23 states. JAMA Pediatr 172 (11), 1038–1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Merrick MT, Ford DC, Ports KA, Guinn AS, Chen J, Klevens J, et al. , 2019. Vital signs: estimated proportion of adult health problems attributable to adverse childhood experiences and implications for prevention - 25 states, 2015–2017. MMWR Morb. Mortal. Wkly Rep 68 (44), 999–1005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Mills R, Alati R, Strathearn L, Najman JM, 2014. Alcohol and tobacco use among maltreated and non-maltreated adolescents in a birth cohort. Addict Abingdon Engl 109 (4), 672–680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Nelson J, Klumparendt A, Doebler P, Ehring T, 2017. Childhood maltreatment and characteristics of adult depression: meta-analysis. Br. J. Psychiatry J. Ment. Sci 210 (2), 96–104. [DOI] [PubMed] [Google Scholar]
  41. Neumark-Sztainer D, Wall M, Larson NI, Eisenberg ME, Loth K, 2011. Dieting and disordered eating behaviors from adolescence to young adulthood: findings from a 10-year longitudinal study. J. Am. Diet. Assoc 111 (7), 1004–1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Neumark-Sztainer D, Wall MM, Chen C, Larson NI, Christoph MJ, Sherwood NE, 2018. Eating, activity, and weight-related problems from adolescence to adulthood. Am. J. Prev. Med 55 (2), 133–141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Noll JG, Zeller MH, Trickett PK, Putnam FW, 2007. Obesity risk for female victims of childhood sexual abuse: a prospective study. Pediatrics 120 (1), e61–e67. [DOI] [PubMed] [Google Scholar]
  44. Norman RE, Byambaa M, De R, Butchart A, Scott J, Vos T, 2012. The long-term health consequences of child physical abuse, emotional abuse, and neglect: a systematic review and meta-analysis. PLoS Med 9 (11), e1001349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pignatelli AM, Wampers M, Loriedo C, Biondi M, Vanderlinden J, 2017. Childhood neglect in eating disorders: a systematic review and meta-analysis. J Trauma Dissociation Off J Int Soc Study Dissociation ISSD 18 (1), 100–115. [DOI] [PubMed] [Google Scholar]
  46. Power C, Li L, Pinto Pereira SM, 2020. An overview of child maltreatment (neglect and abuse) associations with developmental trajectories and long-term outcomes in the 1958 British birth cohort. Longitud Life Course Stud 11 (4), 431–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Quick V, Wall M, Larson N, Haines J, Neumark-Sztainer D, 2013. Personal, behavioral and socio-environmental predictors of overweight incidence in young adults: 10-yr longitudinal findings. Int. J. Behav. Nutr. Phys. Act 25 (10), 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Raissian KM, Bullinger LR, 2017. Money matters: does the minimum wage affect child maltreatment rates? Child Youth Serv. Rev 72, 60–70. [Google Scholar]
  49. Sherwood NE, Wall M, Neumark-Sztainer D, Story M, 2009. Effect of socioeconomic status on weight change patterns in adolescents. Prev. Chronic Dis 6 (1), A19. [PMC free article] [PubMed] [Google Scholar]
  50. Thomas C, Hyppönen E, Power C., 2008. Obesity and type 2 diabetes risk in midadult life: the role of childhood adversity. Pediatrics 121 (5), e1240–e1249. [DOI] [PubMed] [Google Scholar]
  51. Williamson DF, Thompson TJ, Anda RF, Dietz WH, Felitti V, 2002. Body weight and obesity in adults and self-reported abuse in childhood. Int. J. Obes 26 (8), 1075–1082. [DOI] [PubMed] [Google Scholar]
  52. Yoon S, Dillard R, Kobulsky J, Nemeth J, Shi Y, Schoppe-Sullivan S, 2020a. The type and timing of child maltreatment as predictors of adolescent cigarette smoking trajectories. Subst Use Misuse 55 (6), 937–946. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Yoon S, Shi Y, Yoon D, Pei F, Schoppe-Sullivan S, Snyder SM, 2020b. Child maltreatment, fathers, and adolescent alcohol and marijuana use trajectories. Subst Use Misuse 55 (5), 721–733. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

table
NIHMS1866941-supplement-table.docx (89.3KB, docx)

Data Availability Statement

Data will be made available on request.

RESOURCES