ABSTRACT
Introduction
Adverse childhood experiences (ACEs) are associated with poor psychosocial and health outcomes in adulthood. Veterans and females experience ACEs disproportionately. A greater understanding of this disparity may be achieved by examining the relationship between distinct ACE patterns and these demographic characteristics. Therefore, this study examined distinct ACE patterns and their association with Veteran status, sex, and other demographics in a nationally representative sample of U.S. adults to inform interventions tailored to ACE patterns experienced by specific groups.
Materials and Methods
Latent class analysis (LCA) was conducted with data from the National Epidemiologic Survey of Alcohol and Related Conditions-III, a nationally representative structured diagnostic interview conducted from 2012-2013. The target population was the noninstitutionalized adult population living in the USA. The analytic sample was 36,190 (mean age 46.5 years; 48.1% male). Of these participants, 3,111 were Veterans. Data were analyzed between September 2020 and January 2021.
Results
Latent class analysis revealed a four-class solution: (1) “Low adversity” (75.3%); (2) “Primarily household dysfunction” (9.0%); (3) “Primarily maltreatment” (10.7%); and (4) “Multiple adversity types” (5.1%). Compared to “Low adversity,” members in the other classes were more likely to be Veterans (odds ratio (OR)C2vC1 = 1.33, ORC3vC1 = 1.55, ORC4vC1 = 1.98) and female (ORC2vC1 = 1.58, ORC3vC1 = 1.22, ORC4vC1 = 1.65). While lower education and income were also related to higher adversity class membership, Veteran status and sex were the strongest predictors, even when controlling for education and income.
Conclusions
Distinct and meaningful patterns of ACEs identified in this study highlight the need for routine ACE screenings in Veterans and females. As in the current study, operationalizing and clustering ACEs can inform screening measures and trauma-informed interventions in line with personalized medicine. Future work can test if classes are differentially associated with health outcomes.
INTRODUCTION
Adverse childhood experiences (ACEs) are potentially traumatic childhood events or experiences in the home environment that impact safety and stability1 and are disproportionally experienced by Veterans.2 It is well-established that ACEs are associated with numerous subsequent poor health outcomes in adulthood including smoking, overweight/obesity, and mental illness.1,3 Most research examining these associations either consider the total number of ACEs endorsed or use the threshold of four or more ACEs.3,4 These methods assume each ACE type is equally impactful, do not consider patterns among ACE types, and lack information about the inherent heterogeneity of adverse experiences. Latent class analysis (LCA) can be used to overcome these limitations by identifying distinct subgroups (or “classes”) of individuals based on shared ACE patterns. This has been done primarily using data from surveys of two nationally representative U.S. samples.5,6 The handful of studies examining LCAs of ACEs in these samples have identified three to five distinct classes, with one low adversity and several high-adversity groups differing by prevalence rates of ACE types.7–14
Veteran status may be uniquely associated with ACE exposure as, for some, enlisting may provide escape from adverse household environments.15,16 However, it is recognized that many individuals are resilient to adversity and there are many reasons individuals choose to enlist.17 Regarding counts of ACEs, a study examining the current all-volunteer enlistment era found that both males and females with military service reported higher rates of all ACE types than civilians,18 with female Veterans reporting more ACEs than male Veterans.19 Female Veterans also report more ACE types than civilians, suggesting LCA may be useful in exploring patterns of co-occurring ACEs in this population.20 Notably, previous research has not examined such ACE patterns with civilian comparison groups,9,10,21 limiting the applicability of findings and hindering the overall utility for health outcome prediction and intervention development. For example, LCA of ACEs in the most recent National Epidemiologic Survey of Alcohol and Related Conditions (NESARC-III). [NESARC is an interview-based survey conducted in all 50 states at three time periods, called waves (Wave 1 from 2001-2002, Wave 2 from 2004-2005, Wave III from 2012-2013). Waves 1 and 2 consist of the same sample of individuals, while Wave III comprises an entirely new sample and is the sample used in the current study.5 only evaluated ACE classes within specific subsamples such as U.S. military Veterans,9,10 emerging adults,11 and older adults.12 Examining only subsamples impedes relevant comparisons across subpopulations. The one LCA study that used the full NESARC-III included a broader range of adversities in addition to the standard set of ACEs (e.g., childhood poverty).22 Therefore, there is a need to comprehensively examine ACE patterns in male and female Veterans to better understand and address ACEs as potential risk factors for conditions prevalent in Veterans, such as posttraumatic stress disorder (PTSD), depression, and alcohol misuse.18
The current study distinguishes specific patterns of ACEs using the total NESARC-III, which provides an opportunity to examine the replicability of previously identified ACE classes. No studies have comprehensively examined the association of ACE classes with Veteran status. Understanding the differences in ACE patterns between Veterans and civilians could inform our understanding of the mechanisms underlying health disparities among Veterans.23 Further, female Veterans remain underrepresented in research.2 Thus, the aims of the current study were to: (1) Identify unique classes of ACE patterns in the total NESARC-III using LCA; (2) Identify and compare relationships between ACE classes and Veteran status, sex, and other demographics; and (3) Examine the interaction between Veteran status and sex in relation to ACE classes. Greater understanding of ACE patterns in a representative U.S. sample may elucidate intervenable mechanisms between ACEs and poor health outcomes24 in Veterans, and guide targeted health intervention development25 and evaluation.26
METHODS
Data and Sample
Data were obtained from NESARC-III. The target population was the noninstitutionalized U.S. population 18 years or older. Multistage probability sampling was used to collect the data via a computer-assisted personal interviewing system from a final representative sample of 36,309 individuals.5 To conduct the LCA, participants missing values on all ACE items were excluded, resulting in an analytic sample of 36,190. Data were obtained through a Data Use Agreement with National Institute on Alcohol Abuse and Alcoholism and the study was approved by the Research and Development Committee at VA San Diego Healthcare System.
Measures
Adverse childhood experiences (ACEs)
Adverse childhood experiences were assessed and operationalized consistent with the Adverse Childhood Experiences Study,27 resulting in nine indicators. The National Epidemiologic Survey of Alcohol and Related Conditions included variables adapted from the Conflict Tactics Scale,28 Childhood Trauma Questionnaire,29 and questions from Wyatt (1985).30 Respondents rated questions about five types of maltreatment (sexual abuse, physical abuse, physical neglect, emotional abuse, and emotional neglect) and four types of household dysfunction (witnessing interpersonal violence, adult substance use, adult mental health issues, and adult incarceration). Sexual abuse (4 items), physical abuse (2 items), physical neglect (4 items), emotional abuse (3 items), witnessing interpersonal violence (4 items), and emotional neglect (5 items) were assessed using 5-point Likert scales. Household substance use was indicated if respondents answered “yes” to at least one of two items inquiring about problematic (1) alcohol or (2) drug use by an adult in one’s home. Household mental health issues were indicated if participants endorsed at least one of three items regarding whether a parent or adult in the home (1) was treated or hospitalized for a mental illness, (2) attempted suicide, or (3) committed suicide. All domains were dichotomously coded for the current study.10
Sociodemographic variables
Sociodemographic variables included age, sex, race/ethnicity, education, employment, and household income. Age was a continuous variable, and all others were dichotomized, with one as the referent. Sex was coded (0) male, (1) female as the NESARC survey assessed only self-reported sex, and not gender identity. Race/ethnicity was coded (0) White, non-Hispanic, (1) Non-white; education was coded (0) some college education or higher, (1) high school degree or less; unemployment was coded (0) employed, (1) unemployed; lower income was coded as earning (0) ≥$40,000, (1) ≤$39,999, based on median income of NESARC-III.
Veteran status was operationalized as endorsing serving on active duty in the U.S. Armed Forces, Reserves, or National Guard.10 Veteran status was coded as (0) Civilian, (1) Veteran.
Data Analysis
The complex samples module in SPSS 26 was used to calculate descriptive statistics. These analyses accounted for the complex sample design by adjusting for sample weights, clustering, and strata. Latent class analysis was used to identify distinct ACE profiles using Mplus Version 8.6.31 The latent groups examined were based on endorsement of the nine summary binary (yes/no) ACE types. Missing data were accounted for using Full Information Maximum Likelihood. Models were conducted based on 100 or 500 random starts. Relative model fit indices were examined to determine the best fitting model, including Akaike’s information criterion (AIC), Bayesian information criterion (BIC), sample-size adjusted Bayesian information criterion (adj BIC), the Lo–Mendell–Rubin likelihood ratio test (LMR LRT), and entropy.32 Smaller AIC, BIC, and adj BIC values indicate better model fit. The LMR LRT compares the model being examined to a model with k-1 classes. Significant P-values suggest improved model fit relative to a model with k-1 classes. Entropy represents the overall precision of group classification (range = 0-1), with values closer to 1 indicating higher precision. Agreement with theory was considered in determining the number of classes.
Based on the LCA, a nominal variable indicating each participant’s most likely class membership was constructed. Chi-Square Tests of Independence were used to examine the association of class membership with sociodemographics and Veteran status. Multinomial logistic regression was used to examine the odds of class membership as a function of sociodemographics and Veteran status. A sex by Veteran interaction was evaluated to determine whether sex moderated the association between Veteran status and class membership. These analyses also accounted for the complex sample design by adjusting for sample weights, clustering, and strata. Alpha was set at 0.05 for all analyses.
RESULTS
The supplementary table shows fit statistics and class proportions for the latent class models examined. Models up to five classes were examined, whereupon the best log-likelihood value could not be replicated. Considering all fit indices and examining profiles associated with different models, the 4-class model demonstrated the best fit. Figure 1 presents the probability of experiencing each ACE type among participants in each class and the proportion of participants endorsing each of the nine ACEs per class. Probability estimates were used to interpret the classes. Class 1 (“Low adversity;” 75.1%) demonstrated low probability of any form of maltreatment or household dysfunction. Class 2 (“Primarily household dysfunction;” 9.0%) exhibited moderate probabilities of all ACEs and higher probabilities of household dysfunction indicators, especially household substance use. Class 3 (“Primarily maltreatment;” 10.8%) represented high probability for maltreatment indicators and moderate to low probability for household dysfunction. Class 4 (“Multiple adversity types;” 5.2%) demonstrated high probability of all ACE types.
FIGURE 1.
Latent classes and associated endorsement probabilities and frequencies of childhood adversities.
Class 1 = Low adversity; Class 2 = Primarily household dysfunction; Class 3 = Primarily maltreatment; Class 4 = Multiple adversity types.
Table I presents sociodemographic characteristics across the entire sample and for each latent class. Class membership was significantly related to sex, race/ethnicity, employment, income, education, and Veteran status (ps < 0.001). Results of the polynomial logistic regression examining class membership as a function of sociodemographics are in Table II. Compared to individuals in Class 1, those in Class 2 had 33% higher odds of being a Veteran (P = .002), 58% higher odds of being female (P < .0001), and 25% higher odds of unemployment (P = .001); those in Class 3 had 55% higher odds of being a Veteran (P < .0001), 22% higher odds of being female (P < .0001), and 16% higher odds of unemployment (P = .01); and those in Class 4 had 98% higher odds of being a Veteran (P < .0001), 65% higher odds of being female (P < .0001), and 56% higher odds of unemployment (P < .0001). Education and income differentiated Classes 2 and 4 from Class 1 with those with high school education or less having 26% and 49% higher odds and those with an income of less than $40,000 having 27% and 37% higher odds of being in Class 2 and 4, respectively. Class 3 uniquely did not exhibit this association.
TABLE I.
Demographic Characteristics across the Total Sample and within Each Latent Class
| Total Sample (N = 36,190) |
Mean # ACEs (M = 1.29, SE = 0.02) |
≥4 ACEs (11.5%, SE = 0.3%) |
Class 1 (n = 26,679) |
Class 2 (n = 3430) |
Class 3 (n = 4,085) |
Class 4 (n = 1,996) |
||
|---|---|---|---|---|---|---|---|---|
| n, % (SE) | M (SE) | % (SE) | n, % (SE) | n, % (SE) | n, % (SE)/M (SE) | n, % (SE)/M (SE) | χ 2 | |
| Age in years, M (S.E.) | 36,190, 47.1 (0.2) | – | – | 46.9 (0.2) | 43.2 (0.4) | 47.8 (0.4) | 44.6 (0.4) | |
| Female | 20,387, 51.9 (0.3) | 1.38 (0.02) F 1.19 (0.02) M |
13.3 (0.3) F 9.6 (0.3) M |
14,661, 50.3 (0.3) | 2,158, 59.7 (1.0)a | 2,289, 53.2 (0.9) | 1,279, 59.5 (0.9)b | 153.08 |
| Veteran | 3,111, 9.6 (0.3) | 1.44 (0.04) | 13.2 (0.8) | 2,211, 9.2 (0.3)a | 258, 8.4 (0.6)b | 442, 12.8 (0.7)a | 200, 11.6 (1.1) | 63.39 |
| Race/ethnicity | ||||||||
| Non-whitec | 17,052, 33.8 (0.8) | 1.30 (0.02) | 12.5 (0.4) | 12,481, 33.6 (0.8) | 1,687, 35.0 (1.2) | 1,987, 35.1 (1.2) | 897, 31.9 (1.4) | 8.36 |
| White non-Hispanic | 19,138, 66.2 (0.8) | 1.30 (0.02) | 11.6 (0.3) | 14,198, 66.4 (0.8) | 1,743, 65.0 (1.2) | 2,098, 64.9 (1.2) | 1,099, 68.1 (1.4) | 104.36 |
| Black non-Hispanic | 7,733, 11.8 (0.7) | 1.34 (0.03) | 11.8 (0.5) | 5,571, 11.2 (0.7)b | 813, 14.2 (0.9)a | 962, 64.9 (0.9)a | 387, 11.1 (0.9) | |
| Other non-Hispanic | 2,302, 7.3, (0.5) | 1.98 (0.04) | 8.7 (0.7) | 1,804, 7.8 (0.5)a | 139, 4.4 (0.6)b | 255, 6.8 (0.6) | 104, 6.0 (0.7) | |
| Any race Hispanic | 735, 14.7 (0.7) | 1.32 (0.03) | 12.4 (0.5) | 5,106, 14.6 (0.7) | 735, 16.4 (1.1)a | 770, 14.3 (0.8) | 406, 14.9 (1.1) | |
| Unemployedc | 11,121, 34.1 (0.6) | 1.43 (0.02) UE 1.26 (0.03) E |
13.1 (0.4) UE 10.6 (0.3) E |
7,999, 33.1 (0.6)b | 1,042, 34.3 (1.2) | 1,347, 37.4 (1.1)a | 733, 41.1 (1.6)a | 61.64 |
| Household income | ||||||||
| Lower incomec | 19,592, 44.4 (0.7) | 1.43 (0.02) | 14.0 (0.4) | 13,994, 42.6 (0.7)b | 2,046, 51.3 (1.2)a | 2,264, 45.9 (1.2) | 1,288, 55.5 (1.7)a | 194.84 |
| <$20,000 | 9,891, 41.2 (0.6) | 1.51 (0.03) | 15.7 (0.5) | 6,931, 19.3 (0.5) | 1065, 25.2 (0.9) | 1,133, 21.3 (0.9) | 762, 30.5 (1.4)a | 328.63 |
| $20,000-$39,999 | 9,701, 27.2 (0.3) | 1.36 (0.03) | 12.6 (0.5) | 7,063, 23.3 (0.4)b | 981, 26.2 (1.0)a | 1,131, 24.6 (0.8) | 526, 24.9 (1.1) | |
| $40,000-$69,999 | 7,852, 18.3 (0.3) | 1.30 (0.03) | 11.3 (0.5) | 5,810, 22.6 (0.3) | 709, 22.0 (0.8) | 932, 24.8 (0.9)a | 401, 23.3 (1.3) | |
| $70,000+ | 8,746, 13.2 (0.4) | 1.09 (0.02) | 8.3 (0.4) | 6,875, 34.8 (0.7)a | 675, 26.6 (1.0)b | 889, 29.3 (1.1)b | 307, 21.2 (1.5)b | |
| Education | ||||||||
| Less educationc | 15,224, 38.7(0.8) | 1.44 (0.03) | 14.6 (0.4) | 10,956, 73.4 (0.6)b | 1,592, 44.3 (1.1)a | 1,691, 39.0 (1.1) | 985, 48.5 (1.3)a | 138.01 |
| Less than HS | 15,224, 13.0 (0.4) | 1.51 (0.04) | 15.4 (0.8) | 3,854, 12.1 (0.5)b | 566, 15.2 (0.7)a | 648, 14.4 (0.8)a | 394, 18.5 (1.0)a | 420.94 |
| Completed HS | 9,762, 25.8 (0.5) | 1.38 (0.03) | 12.6 (0.5) | 7,102, 25.2 (0.6)b | 1,026, 29.1 (0.9)a | 1,043, 24.6 (0.9) | 591, 30.0 (1.2)a | |
| Some college | 12,065, 33.1(0.5) | 1.38 (0.02) | 12.9 (0.4) | 8,638, 32.1 (0.5)b | 1,230, 35.6 (1.0)a | 1,437, 35.4 (1.0)a | 760, 37.9 (1.4)a | |
| BA or higher | 8,901, 28.1 (0.8) | 1.01 (0.02) | 7.2 (0.3) | 7,085, 30.5 (0.8)a | 608, 20.0 (1.1)b | 957, 25.6 (1.0)b | 251, 13.7 (1.1)b |
Class 1 = Low adversity; Class 2 = Primarily household dysfunction; Class 3 = Primarily maltreatment; Class 4 = Multiple adversity types. χ2 = chi-square test of independence. BA = Bachelor’s degree, E = Employed, F = Female, HS = High school, M = Male, UE = Unemployed. aStandardized residuals > 2; bStandardized residuals <−2; Boldface indicates P <.0001; cDichotomous variable used in primary analyses.
TABLE II.
Summary of Polynomial Regressions Examining Associations between Latent Class and Demographic Characteristics
| Class 2 V Class 1 | Class 3 V Class 1 | Class 4 V Class 1 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | OR | 95% CI | P-value | OR | 95% CI | P-value | OR | 95% CI | P-value |
| Age | 0.98 | 0.98, 0.99 | <.0001 | 1.00 | 1.00, 1.00 | .06 | 0.98 | 0.98, 0.98 | <.0001 |
| Female | 1.58 | 1.43, 1.74 | <.0001 | 1.22 | 1.10, 1.34 | <.0001 | 1.65 | 1.45, 1.88 | <.0001 |
| Veteran | 1.33 | 1.11, 1.60 | .002 | 1.55 | 1.32, 1.81 | <.0001 | 1.98 | 1.51, 2.60 | <.0001 |
| Unemployed | 1.25 | 1.09, 1.42 | .001 | 1.16 | 1.04, 1.31 | .01 | 1.56 | 1.35, 1.79 | <.0001 |
| Lower education | 1.26 | 0.72, 0.88 | <.0001 | 1.05 | 0.87, 1.05 | .34 | 1.49 | 0.59, 0.77 | <.0001 |
| Non-white | 0.91 | 0.82, 1.01 | .076 | 1.11 | 0.99, 1.23 | .06 | 0.78 | 0.68, 0.90 | .001 |
| Lower income | 1.27 | 1.15, 1.41 | <.0001 | 1.05 | 0.96, 1.15 | 0.26 | 1.37 | 1.18, 1.58 | <.0001 |
| Veteran status × Female | 1.14 | 0.75, 1.72 | .54 | 1.03 | 0.67, 1.57 | 0.90 | 1.19 | 0.75, 1.88 | .46 |
Class 1 = Low adversity; Class 2 = Primarily household dysfunction; Class 3 = Primarily maltreatment; Class 4 = Multiple adversity types.
Table III presents the rates of ACE types and classes by Veteran status and sex. Physical Neglect was the most reported ACE in the entire sample and among Veterans. Household Substance Use was the most reported adversity in females overall and female Veterans. Chi-square analyses showed female civilians and female Veterans had significantly higher rates of most ACEs than their male counterparts with the largest discrepancy in Sexual Abuse and Emotional Abuse. Although female Veterans had higher rates of many ACEs and Class 4 membership compared to female civilians, and Veterans had higher rates of many ACEs and adverse classes than civilians, the sex by Veteran interaction was not statistically significant.
TABLE III.
Prevalence of ACE Types and Classes by Sex and Veteran Status
| Civilians | Veterans | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Female % (SE) |
Male % (SE) |
Total % (SE) |
Female v. male | Female % (SE) |
Male % (SE) |
Total % (SE) |
Female v. Male | Civilian female v. veteran female | |
| ACE/Class | n = 20,008 | n = 13,071 | n = 33,079 | χ 2 | n = 379 | n = 2,732 | n = 3,111 | χ 2 | χ2 |
| Household member in prison or jail | 8.1 (0.2) n = 1,782 |
7.4 (0.3) n = 1,060 |
7.8 (0.2) n = 2,842 |
4.54 | 10.7 (2.0) n = 45 |
6.3 (0.5) n = 182 |
6.7 (0.5) n = 227 |
8.71* | 3.49 |
| Sexual abuse | 15.5 (0.4) n = 3,113 |
5.6 (0.3) n = 755 |
11.2 (0.3) n = 3,868 |
806.95** | 19.3 (2.8) n = 78 |
6.9 (0.6) n = 194 |
8.1 (0.6) n = 272 |
55.52* | 3.92 |
| Physical abuse | 18.2 (0.5) n = 3,730 |
17.7 (0.5) n = 2,432 |
17.9 (0.4) n = 6,162 |
1.38 | 28.9 (3.0) n = 104 |
26.6 (1.2) n = 726 |
26.9 (1.1) n = 830 |
0.73 | 28.24** |
| Physical neglect | 25.5 (0.5) n = 5,210 |
29.4 (0.6) n = 3,957 |
27.2 (0.4) n = 9,167 |
59.39** | 30.6 (3.2) n = 113 |
33.6 (1.2) n = 904 |
33.3 (1.2) n = 1,017 |
1.13 | 4.89 |
| Substance use in the home | 26.3 (0.5) n = 5,238 |
22.6 (0.5) n = 2,971 |
24.7 (0.4) n = 8,209 |
62.63** | 34.7 (2.6) n = 126 |
25.2 (1.1) n = 678 |
26.2 (1.0) n = 804 |
12.79
** |
13.11* |
| Mental health issues | 7.8 (0.2) n = 1,504 |
5.8 (0.2) n = 752 |
6.9 (0.2) n = 2,256 |
52.52** | 12.2 (2.3) n = 39 |
6.3 (0.5) n = 183 |
6.9 (0.5) n = 222 |
14.94* | 9.79* |
| Emotional neglect | 11.0 (0.3) n = 2,360 |
8.3 (0.3) n = 1,189 |
9.8 (0.3) N = 3,549 |
69.33** | 12.6 (2.3) n = 50 |
10.9 (0.7) n = 298 |
11.1 (0.7) n = 348 |
0.79 | 36.62** |
| Emotional abuse | 11.4 (0.3) n = 2,319 |
8.3 (0.3) n = 1,162 |
10.1 (0.3) n = 3,481 |
91.14** | 21.6 (2.5) n = 301 |
11.6 (0.8) n = 321 |
12.6 (0.7) n = 398 |
25.03** | 36.62** |
| Witnesses IPV | 14.0 (0.4) n = 2,946 |
9.9 (0.4) n = 1,391 |
12.2 (0.3) n = 4,337 |
125.26** | 17.8 (2.4) n = 71 |
13.3 (0.7) n = 353 |
13.7 (0.7) n = 424 |
4.70* | 4.36 |
| Class 1 | 72.9 (0.5) n =14,426 |
78.7 (0.5) n = 10,042 |
75.4 (0.4) n = 24,468 |
146.61** | 62.1 (3.0) n = 235 |
72.8 (1.0) n = 1,976 |
71.7 (1.0) n = 2,211 |
151.63** | 21.59** |
| Class 2 | 10.3 (0.3) n = 2,107 |
7.6 (0.3) n = 1,065 |
9.1 (0.2) n = 3,172 |
71.31** | 12.8 (1.7) n = 51 |
7.2 (0.5) n = 207 |
7.8 (0.5) n = 258 |
12.17** | 2.60 |
| Class 3 | 11.0 (0.3) n = 2,237 |
9.7 (0.3) n = 1,406 |
10.4 (0.2) n = 3,643 |
14.49* | 15.0 (2.4) n = 52 |
14.2 (0.8) n = 390 |
14.3 (0.8) n = 442 |
0.12 | 5.81 |
| Class 4 | 5.8 (0.2) n = 1,238 |
4.0 (0.2) n = 558 |
5.0 (0.2) n = 1,796 |
146.61** | 10.1 (2.1) n = 41 |
5.8 (0.6) n =159 |
6.2 (0.6) n = 200 |
8.99* | 12.20* |
Class 1 = Low adversity; Class 2 = Moderate maltreatment with substance use; Class 3 = Severe maltreatment with moderate household dysfunction; Class 4 = Severe multi-type adversities. Boldface indicates statistical significance (*P < .01; **P < .001).
DISCUSSION
To our knowledge, this is the first study to comprehensively examine the association of ACE patterns with sociodemographic characteristics and specifically compare profiles between Veterans and civilians. Four unique classes emerged: (1) “Low adversity;” (2) “Primarily household dysfunction;” (3) “Primarily maltreatment;” and (4) “Multiple adversity types.” Veteran status was associated with greater odds of higher adversity than any other characteristic. There was some support this was driven by female Veterans, who had significantly higher prevalence of Class 2 membership than male Veterans. Female sex was associated with membership in higher adversity classes compared to the “Low adversity” reference group. Less education and lower income were associated with membership in Classes 2 and 4, but not with Class 3.
Veterans, regardless of sex, were twice as likely as civilians to be in Class 4. This is consistent with previous research that reported higher prevalence of ACE types in male Veterans than civilians,18 and extends it to suggest both male and female Veterans are more likely to experience patterns of severe childhood maltreatment than civilians. As previously hypothesized for male Veterans,18 joining the military may provide some males and females with a means of escape from dysfunctional home environments. Clinically, these findings suggest military and Veteran healthcare may benefit from ACE screening to inform comprehensive biopsychosocial and personalized treatment decision-making. For example, screening could occur during initial visits to a VA or military healthcare facility so that adversity-related physical and mental concerns can be better recognized and addressed, and trauma-informed care and services can be employed to mitigate the potential harm caused by ACEs.33 This information would provide clinicians insight into potential contributing and maintaining factors for certain health outcomes, and any resilience or protective factors present, to help them better understand their patients’ contexts. Brief ACE screening in healthcare settings also can flag adversity- or trauma-related somatic and mental health symptoms and conditions, and ensure appropriate referrals to services and supports for identified issues. This care approach could provide a holistic streamlined team-based healthcare experience to connect patients with necessary healthcare services as needed.
However, a history of ACEs should not be implied to necessarily cause poor health. In fact, military service could be protective for individuals who have experienced ACEs, for example by providing socioeconomic benefits. Elements of military experience, such as unit cohesion, also may help mitigate the relationship between stressful life experiences and poor mental health outcomes, suggesting resiliency.34 Because ACEs do not ensure poor health outcomes or imply a lack of fitness for duty, such screening should be reserved for VA and other healthcare settings. More research is needed to better understand motivating factors for joining the military, the role of ACEs, subsequent military experiences, and any potential compounding or attenuating effect on physical and mental health in Veterans. Understanding the complex interplay between ACEs, military experience, and PTSD or suicidality, for example, can inform military and Veteran healthcare systems to design tailored prevention and intervention strategies.
Females were more likely to fall within each higher adversity class than males, indicating greater likelihood of experiencing these patterns of ACE types. This is consistent with previous reports of adversity patterns,35,36 and of increased childhood sexual and emotional abuse, and household substance abuse and mental illness in females compared to males.1 Given the significantly higher rates of sexual and emotional abuse in both civilian and Veteran females than males, female membership in higher adversity classes may have been driven by these ACE types. The observed sex differences suggest further research is needed to understand mechanisms underlying associations between sex, ACEs, and health outcomes.
We found higher rates of some ACEs among female Veterans than female civilians or male Veterans, which may drive the association between Veteran status and higher adversity profiles. However, sex did not moderate the relationship between Veteran status and ACE profiles, likely due to the relatively small sample of female Veterans. Because this dataset was designed to be representative of the general population without consideration of the Veteran population, only 9.6% of the total sample were Veterans and only 1% were female Veterans. Thus, the fact that the sex by Veteran interaction was not significant should be viewed with caution and not interpreted as a conclusive absence of an effect. More research is needed to examine how a history of ACEs in females may be implicated in vulnerability to subsequent abuse37 or influence exposure to adverse experiences, like sexual assault15 or military sexual trauma, in adulthood.
Our overall findings confirm ACE patterns previously identified among various subsamples.9,10,12,13 The similarities between previously identified ACE profiles and those in this study demonstrate the replicability of ACE patterns. Specifically, patterns distinguishing Classes 2 and 3 show that exposure to household substance use often co-occurs with a moderate amount of maltreatment and that severe maltreatment often co-occurs with household dysfunction. The unique association of Classes 2 and 4 with less education and lower income potentially reflects a relationship between increased household substance use and socioeconomic status.38 Consistent with prior research,8–10,12,14,22 our identified class structure suggests individuals either experience comparatively minimal ACEs or a combination of multiple adversities.
Limitations
The current study has limitations. First, NESARC-III is cross-sectional and retrospective reports of ACEs may be prone to recall bias. Further, we lacked data on frequency and severity of ACEs. Future prospective studies should focus on the temporal relationship of ACEs of various frequency and severity with time-sensitive demographics (e.g., Veteran status, education, employment, and income). Second, Veteran status was self-reported and because NESARC-III was developed to be representative of the general population and included one question about Veteran status, we were unable to assess military specific factors like combat experience to better describe the sample or examine complex relationships. Future research should examine the role and impact of ACEs in the context of other military specific experiences. Third, NESARC did not distinguish between sex and gender, limiting our analyses to the use of self-reported sex. Future research should differentiate these important aspects of identity. Finally, the association of race/ethnicity as white versus non-white with ACE profiles approached significance, indicating a need for more nuanced investigation of racial/ethnic groups, acculturation, intersectionality, and how these interact with ACE patterns.1 Further, this standard method of assessing race/ethnicity is potentially stigmatizing and does little to elucidate intervenable mechanisms for change.39
CONCLUSIONS
We found three distinct ACE patterns and their variable association with Veteran status, sex, and other sociodemographic characteristics in the large, nationally representative NESARC-III sample. Understanding characteristics that may be associated with certain ACE patterns can facilitate future research to examine underlying mechanisms among sets of experiences, resilience factors, and specific health outcomes. Being Veteran and female were consistently associated with greater odds of membership in all higher adversity classes. This underscores the need for: proactive ACE screenings in healthcare settings for service members, Veterans, and females; longitudinal research with service members and Veterans, especially females; and development and evaluation of streamlined prevention and intervention strategies in healthcare settings for service members and Veterans who may be at risk for physical and mental health conditions linked to ACEs. Using a person-centered approach to operationalize and cluster ACEs could inform personalized and focused interventions that prove most effective for individuals with specific ACE patterns and associated health risk factors.25
Supplementary Material
ACKNOWLEDGMENTS
This manuscript was prepared using a limited access dataset obtained from the National Institute on Alcohol Abuse and Alcoholism and does not reflect the opinions or views of NIAAA or the U.S. Government.
Contributor Information
Mara Tynan, San Diego Joint Doctoral Program in Clinical Psychology, San Diego State University/University of California, San Diego, CA 92120, USA.
Jennalee S Wooldridge, VA San Diego Healthcare System, San Diego, CA 92161, USA; Department of Psychiatry, University of California San Diego, La Jolla, CA 92093, USA; VA San Diego Center of Excellent for Stress and Mental Health, San Diego, CA 92161, USA.
Fernanda Rossi, Center for Innovation to Implementation, VA Palo Alto Health Care System, Menlo Park, CA 94025, USA; Center for Health Policy/Center for Primary Care and Outcomes Research, Stanford University School of Medicine, Stanford, CA 94305, USA.
Caitlin L McLean, VA San Diego Healthcare System, San Diego, CA 92161, USA; Department of Psychiatry, University of California San Diego, La Jolla, CA 92093, USA.
Marianna Gasperi, VA San Diego Healthcare System, San Diego, CA 92161, USA; Department of Psychiatry, University of California San Diego, La Jolla, CA 92093, USA; VA San Diego Center of Excellent for Stress and Mental Health, San Diego, CA 92161, USA.
Jeane Bosch, National Center for PTSD, Dissemination & Training Division, VA Palo Alto Health Care System, Menlo Park, CA 94025, USA.
Christine Timko, Center for Innovation to Implementation, VA Palo Alto Health Care System, Menlo Park, CA 94025, USA; Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94305, USA.
Matthew Herbert, VA San Diego Healthcare System, San Diego, CA 92161, USA; Department of Psychiatry, University of California San Diego, La Jolla, CA 92093, USA; VA San Diego Center of Excellent for Stress and Mental Health, San Diego, CA 92161, USA.
Niloofar Afari, VA San Diego Healthcare System, San Diego, CA 92161, USA; Department of Psychiatry, University of California San Diego, La Jolla, CA 92093, USA; VA San Diego Center of Excellent for Stress and Mental Health, San Diego, CA 92161, USA.
SUPPLEMENTARY MATERIAL
SUPPLEMENTARY MATERIAL is available at Military Medicine online.
FUNDING
Drs. Afari and Herbert, and Ms. Tynan are partially supported by R01DK106415 from the National Institute of Diabetes and Digestive and Kidney Diseases. Drs. Wooldridge and McLean are supported by the VA Office of Academic Affiliations Advanced Fellowship in Women’s Health. Dr. Rossi is supported by the VA Office of Academic Affiliations and Health Services Research and Development Service Research funds. Dr. Gasperi is funded by VA Career Development Award 1IK2CX002107 from the Veterans Affairs Clinical Science Research and Development Service. Dr. Timko is supported by a VA Health Services Research and Development Research Career Scientist Award RCS 00-001. Dr. Herbert is supported by Veterans Affairs Rehabilitation Research and Development Service Career Development Award 1IK2RX002807. The views expressed in this paper are those of the authors and do not reflect the official policy or position of any funding agencies, Department of Veterans Affairs, the United States Government, or any institutions with which the authors are affiliated.
CONFLICT OF INTEREST STATEMENT
None of the authors have any conflicts of interest to declare.
REFERENCES
- 1. Centers for Disease Control and Prevention (CDC) : Adverse Childhood Experiences (ACEs). Available at https://www.cdc.gov/violenceprevention/aces/index.html; accessed April 12, 2021.
- 2. Street AE, Vogt D, Dutra L: A new generation of women veterans: stressors faced by women deployed to Iraq and Afghanistan. Clin Psychol Rev 2009; 29(8): 685–94. [DOI] [PubMed] [Google Scholar]
- 3. Felitti VJ, Anda RF, Nordenberg D, et al. : 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 1998; 14(4): 245–58. [DOI] [PubMed] [Google Scholar]
- 4. Hughes K, Bellis MA, Hardcastle KA, et al. : The effect of multiple adverse childhood experiences on health: a systematic review and meta-analysis. Lancet Public Health 2017; 2(8): e356. [DOI] [PubMed] [Google Scholar]
- 5. Grant BF, Chu A, Sigman R, et al. : Source and accuracy statement: National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III). Available at https://www.niaaa.nih.gov/sites/default/files/NESARC_Final_Report_FINAL_1_8_15.pdf, published 2015; accessed September 10, 2020.
- 6. Centers for Disease Control and Prevention (CDC) : Behavioral risk factor surveillance system. 2019. Available at https://www.cdc.gov/brfss/annual_data/2019/pdf/overview-2019-508.pdf; accessed June 10, 2021.
- 7. Roos LE, Afifi TO, Martin CG, Pietrzak RH, Tsai J, Sareen J: Linking typologies of childhood adversity to adult incarceration: findings from a nationally representative sample. Am J Orthopsychiatry 2016; 86(5): 584–93. [DOI] [PubMed] [Google Scholar]
- 8. Cavanaugh CE, Petras H, Martins SS: Gender-specific profiles of adverse childhood experiences, past year mental and substance use disorders, and their associations among a national sample of adults in the United States. Soc Psychiatry Psychiatr Epidemiol 2015; 50(8): 1257–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. McLafferty M, Ross J, Waterhouse-Bradley B, Armour C: Childhood adversities and psychopathology among military veterans in the US: the mediating role of social networks. J Anxiety Disord 2019; 65: 47–55. [DOI] [PubMed] [Google Scholar]
- 10. Ross J, Waterhouse-Bradley B, Contractor AA, Armour C: Typologies of adverse childhood experiences and their relationship to incarceration in U.S. military veterans. Child Abuse Negl 2018; 79: 74–84. [DOI] [PubMed] [Google Scholar]
- 11. Guastaferro K, Bray BC: Substance use and mental health outcomes during emerging adulthood among individuals with different patterns of child maltreatment. Emerg Adulthood 2020; 8(6): 542–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kim Y, Kim K, Chartier KG, Wike TL, McDonald SE: Adverse childhood experience patterns, major depressive disorder, and substance use disorder in older adults. Aging Ment Health 2021; 25(3): 484–91. [DOI] [PubMed] [Google Scholar]
- 13. Salas J, van Den Berk-clark C, Skiöld-Hanlin S, Schneider FD, Scherrer JF: Adverse childhood experiences, depression, and cardiometabolic disease in a nationally representative sample. J Psychosom Res 2019; 127: 109842. [DOI] [PubMed] [Google Scholar]
- 14. Barboza GE: Latent classes and cumulative impacts of adverse childhood experiences. Child Maltreat 2018; 23(2): 111–25. [DOI] [PubMed] [Google Scholar]
- 15. Sadler A: Health-related consequences of physical and sexual violence: women in the military. Obstet Gynecol 2000; 96(3): 473–80. [DOI] [PubMed] [Google Scholar]
- 16. Woodruff T, Kelty R, Segal DR: Propensity to serve and motivation to enlist among American combat soldiers. Armed Forces Soc 2006; 32(3): 353–66. [Google Scholar]
- 17. Ivany CG, Hoge CW: Adverse childhood experiences and military service. JAMA Psychiatry 2015; 72(3): 296. [DOI] [PubMed] [Google Scholar]
- 18. Blosnich JR, Dichter ME, Cerulli C, Batten SV, Bossarte RM: Disparities in adverse childhood experiences among individuals with a history of military service. JAMA Psychiatry 2014; 71(9): 1041–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Aronson KR, Perkins DF, Morgan NR, et al. : The impact of adverse childhood experiences (ACEs) and combat exposure on mental health conditions among new post-9/11 veterans. Psychol Trauma Theory Res Pract Policy 2020; 12(7): 698–706. [DOI] [PubMed] [Google Scholar]
- 20. McCauley HL, Blosnich JR, Dichter ME: Adverse childhood experiences and adult health outcomes among veteran and non-veteran women. J Women’s Health 2015; 24(9): 723–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Gaska KA, Kimerling R: Patterns of adverse experiences and health outcomes among women veterans. Am J Prev Med 2018; 55(6): 803–11. [DOI] [PubMed] [Google Scholar]
- 22. van Draanen J: Unique roles of childhood poverty and adversity in the development of lifetime co-occurring disorder. SSM Popul Health 2020; 10: 100540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kondo K, Low A, Everson T, et al. : Health disparities in veterans: a map of the evidence. Med Care 2017; 55(Suppl 2): S9–15. [DOI] [PubMed] [Google Scholar]
- 24. Briggs EC, Amaya-Jackson L, Putnam KT, Putnam FW: All adverse childhood experiences are not equal: the contribution of synergy to adverse childhood experience scores. Am Psychol 2021; 76(2): 243–52. [DOI] [PubMed] [Google Scholar]
- 25. Lacey RE, Minnis H: Practitioner review. Twenty years of research with adverse childhood experience scores – Advantages, disadvantages and applications to practice. J Child Psychol Psychiatry 2020; 61(2): 116–30. [DOI] [PubMed] [Google Scholar]
- 26. Lanza ST, Rhoades BL: Latent class analysis: an alternative perspective on subgroup analysis in prevention and treatment. Prev Sci 2013; 14(2): 157–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Dube SR, Felitti VJ, Dong M, Chapman DP, Giles WH, Anda RF: Childhood abuse, neglect, and household dysfunction and the risk of illicit drug use: the adverse childhood experiences study. Pediatrics 2003; 111(3): 564–72. [DOI] [PubMed] [Google Scholar]
- 28. Straus MA: Measuring intrafamily conflict and violence: the Conflict Tactics (CT) scales. J Marriage Fam 1979; 41(1): 75–88. [Google Scholar]
- 29. Bernstein DP, Fink L, Handelsman L, et al. : Initial reliability and validity of a new retrospective measure of child abuse and neglect. Am J Psychiatry 1994; 151(8): 1132–6. [DOI] [PubMed] [Google Scholar]
- 30. Wyatt GE: The sexual abuse of Afro-American and White-American women in childhood. Child Abuse Negl 1985; 9(4): 507–19. [DOI] [PubMed] [Google Scholar]
- 31.Muthén LK, Muthén BO: Mplus User’s Guide. 8th ed. Muthén & Muthén. [Google Scholar]
- 32. Nylund KL, Asparouhov T, Muthén BO: Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling 2007; 14(4): 535–69. [Google Scholar]
- 33. Jones CM, Merrick MT, Houry DE: Identifying and preventing adverse childhood experiences: implications for clinical practice. JAMA 2020; 323(1): 25–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Brailey K, Vasterling JJ, Proctor SP, et al. : PTSD symptoms, life events, and unit cohesion in U.S. Soldiers: baseline findings from the neurocognition deployment health study. J Trauma Stress 2007; 20(4): 495–503. [DOI] [PubMed] [Google Scholar]
- 35. Haahr-Pedersen I, Perera C, Hyland P, et al. : Females have more complex patterns of childhood adversity: implications for mental, social, and emotional outcomes in adulthood. Eur J Psychotraumatol 2020; 11(1): 1708618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. McAnee G, Shevlin M, Murphy J, Houston J: Where are all the males? Gender-specific typologies of childhood adversity based on a large community sample. Child Abuse Negl 2019; 90: 149–59. [DOI] [PubMed] [Google Scholar]
- 37. Ports KA, Ford DC, Merrick MT: Adverse childhood experiences and sexual victimization in adulthood. Child Abuse Negl 2016; 51: 313–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Karriker-Jaffe KJ, Liu H, Kaplan LM: Understanding associations between neighborhood socioeconomic status and negative consequences of drinking: a moderated mediation analysis. Prev Sci 2016; 17(4): 513–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Moscou S: The conceptualization and operationalization of race and ethnicity by health services researchers. Nurs Inq 2008; 15(2): 94–105. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.