Adverse childhood experiences and chronic health outcomes: evidence from 33 US states in the Behavioral Risk Factor Surveillance System, 2019-2023

Abstract

Background

Recent evidence suggests a significant association between adverse childhood experience (ACE) and chronic health outcomes among U.S. adults. However, there remains a critical need to explore these associations specifically with respect to racial disparities. Early adversity and preexisting health vulnerabilities may interact, compounding the risk of chronic diseases in historically marginalized populations. This study further explored the relationship between ACE and chronic disease, recognizing that ACEs may exert a more pronounced effect in racial and ethnic groups already at elevated risk. To investigate this relationship, subgroup analysis was conducted to explore variations by race and ethnicity.

Methods

We analyzed data from the Behavioral Risk Factor Surveillance System (BRFSS) collected from 33 states between 2019 and 2023. ACE scores were categorized as none, low (1-2), or high (3+). Log-binomial regression assessed the relationship between ACE scores and 17 health outcomes. Subgroup analyses examined variation by race/ethnicity, and geographic patterns were summarized by state. All analyses accounted for age, sex, race/ethnicity, income, and education.

Results

Of 359,507 participants, 24.4% reported high ACE exposure. Emotional abuse, parental separation, and household substance abuse were the most reported ACEs. Individuals with high ACE exposure had higher risks of depression, smoking, coronary heart disease, and other conditions. Racial disparities were evident in the subgroup analysis. While white respondents with high ACE were significantly associated with many health outcomes, other races/ethnicities often demonstrated higher risk ratios when significant. Particularly, AIAN respondents showed the highest national-level risk for conditions such as heart attack, coronary heart disease, and stroke. Geographically, ACE prevalence and health-related outcomes varied by state, with Oregon and Nevada exhibiting the highest mean ACE scores.

Discussion

High ACE scores are associated with chronic disease and mental health issues. These findings highlight significant racial and geographic disparities in ACE exposure and its health impacts. Addressing ACEs holistically by considering state-related factors and predisposed health risks among racial/ethnic groups is an emerging need. State-level policies focused on trauma prevention, particularly for vulnerable racial groups and high-risk geographic areas, may help implement interventions tailored to address the unique associations between ACEs and health outcomes in diverse populations.

Background

Childhood experiences and development are critical components of an overall healthy and prosperous adult population [1]. Multiple research studies have demonstrated that exposure to adverse childhood experiences (ACEs), such as experiencing physical, sexual, or psychological violence, can have long-lasting negative effects into adulthood [2,3,4,5,6,7,8,9]. A common way to study ACEs is to use a scoring system, often ranging from 0 to 8, based on answers to questions related to these topics [4, 10, 11]. Multiple ACE-related studies have concluded that individuals who experience six or more ACEs face an increased risk of premature mortality, dying on average 20 years earlier than those without ACEs [12, 13]. One common system used in these studies is the Behavioral Risk Factor Surveillance System (BRFSS), which conducts annual surveys across the United States (U.S.). BRFSS conducts over 400,000 interviews per year and collects self-reported data for various topics; including ACE, substance use behaviors, and health outcomes such as heart disease, depression, diabetes, and other prevalent chronic diseases [14].

The mechanism between ACE and poor health outcomes is complex and multifaceted; however, there is increasing evidence of childhood exposure to chronic stress influencing the nervous, endocrine, and immune systems [15, 16]. Chronic stress and development issues during childhood can lead to a lifetime of mental health issues, substance abuse, and disease susceptibility, which highlights the importance of monitoring systems and implementation of policies aimed at avoiding or mitigating ACE exposure. One program involved in monitoring ACE effects is the State Trauma and Resilience Network (STRN), which includes 11 states focused on trauma holistically to identify and address the negative effects in an effort to support a healthier population [17]. Member states of STRN share knowledge and strategies toward their common goal to mitigate the effects of trauma and prevent ACE. While no model is universal, establishing a forum for peer learning, sharing research, and discussing effective implementation strategies represents an important collaborative effort to address this burgeoning issue.

While the relationship between ACE and poor health outcomes continues to be investigated, much of the survey data collected from BRFSS (~ 80%) is from the White U.S. population, leaving minorities underrepresented in the data. Limited attention has been given to American Indian and Alaska Native (AIAN) populations, despite their known health disparities and historical exposure to trauma [18]. Notably, North Dakota and South Dakota demonstrate the two highest numbers of AIAN participants in the BRFSS data. Survey data gathered from North Dakota and South Dakota, which are distinctly similar in terms of demographics and geography, may provide unique insights into AIAN health associations with ACE. However, health outcomes associations with ACEs in the Dakotas have not been thoroughly explored.

This study aims to further examine the relationship between ACE and multiple detrimental health outcomes, while considering minority populations and geographic differences. It is imperative to examine how these associations vary across racial and ethnic groups, especially the AIAN populations. Secondarily, we aim to identify state-level patterns in ACE exposure and related health outcomes. BRFSS data from 33 states with available ACE-related data over the past five years (2019-2023) were utilized to evaluate the associations between ACEs and 17 health outcomes. Overall, this research contributes to a more nuanced understanding of how ACEs influence health disparities and underscores the need for targeted interventions that are tailored to specific population needs.

Methods

Data source

This study analyzed cross-sectional data from the BRFSS, overseen by the Centers for Disease Control and Prevention (CDC), collected between 2019 and 2023 across 33 states in the United States [14]. BRFSS is a national telephone survey that collects data on U.S. residents regarding their chronic health conditions, health-related risk behaviors, and use of preventive services in the adult population aged 18 years and older, covering all 50 states, the District of Columbia, and three U.S. territories. Detailed data quality information for each survey year is available in the CDC’s annual Summary Data Quality Reports for 2019 through 2023 [19,20,21,22,23]. The dataset contains information on demographic and socioeconomic factors, chronic health conditions, adverse childhood experiences (ACEs), and mental health. ACE-related questions are part of the optional modules and consist of 11 questions that delve into adverse childhood events experienced before the age of 18. Responses with “Refused” or “Don’t know” were coded as missing for all questions. Respondents with any missing data for ACE were excluded from the analysis. Table 1 shows the summary of our data filtering steps and the number of cases that were excluded based on missing responses to questions used to calculate ACE scoring. This study was exempt from institutional review board approval because CDC did not engage in human subjects research and personal identifiers were not included in the data file.

Table 1 Summary of data filtering. BRFSS data were collected for the years 2019–2023 and processed for ACE scoring. Respondents who were missing ACE data, refused to respond to questions used in ACE scoring, or answered “Don’t Know” to those questions were excluded from the analysis. The respondents who completed the ACE-related questions were utilized in all subsequent analyses

ACE scoring variable

The primary exposure variable was the ACE score (range 0–8). The score was derived from the questions asked of participants about sexual, physical, and emotional abuse, childhood neglect, and household dysfunction. We employed an established approach that has been used by multiple studies, where the participants were categorized into three groups based on their ACE scores: 0 ACEs (None), 1–2 ACEs (Low ACE), and 3 or more ACEs (High ACE) [10, 24, 25]. ACE score was counted as one point per exposure, with related questions cumulatively contributing to a score of 1. This approach resulted in a range of scores from 0 to 8 (Supplementary Table 1).

Health outcome variables

This study considered 17 health outcomes, including arthritis, asthma, chronic obstructive pulmonary disease (COPD), coronary heart disease, depressive disorder, diabetes, kidney disease, skin cancer, other cancers, and stroke, as self-reported. Other health outcomes include BMI above 25, smoking status (defined as a lifetime smoking of at least 100 cigarettes and currently smoking on at least some days), self-rated fair or poor general health, 14 or more days of poor physical and mental health in the past 30 days, and heavy drinking in the past 30 days. Heavy drinking was defined as four or more drinks for females or five or more drinks for males on one occasion. Official BRFSS codes for health outcome categories used are available in Supplementary Table 2.

Sociodemographic variables

We considered categories such as sex, age, education, race/ethnicity, and income as covariates with ACE to control for these sociodemographic factors. While examining racial disparities, we categorized race/ethnicity into White, Black, Asian, Hispanic, AIAN, and Other (Native Hawaiian and Pacific Islanders, Multiracial, and others). Due to the importance of considering racial disparities, we conducted subgroup analyses stratified by race/ethnicity to explore their associations between ACE and health outcomes. Considering the known health disparities of AIAN communities such as rates of diabetes and substance use disorders, we chose to focus on the AIAN respondents from North Dakota and South Dakota. These states are distinctly similar and demonstrate the highest number of responses from this minority population, creating a valuable opportunity to gain insights into how ACEs may influence health outcomes within this community. The predictive power of ACE was explored separately for each of the six racial/ethnic groups.

Statistical analysis

Descriptive analyses were used to summarize ACE scores, demographic characteristics, and the weighted prevalence of each ACE component by socio-demographic groups. Survey weights provided by the BRFSS were adjusted based on the proportion of total pooled samples that each year contributes. We multiplied the BRFSS weights by the ratio of the total ACE-complete samples to that year’s sample size to account for variation in missing ACE data and unequal sample sizes across the five years of data collected from 2019–2023. This adjustment ensured that the data were representative of the overall population across those years.

$${w}_{i,y}^{\text{adjusted}}={w}_{i,y}\times \frac{N}{{n}_{y}}$$

where,

${w}_{i,y}=original\; BRFSS \;weight\; for\; respondent\; i\; in\; year\; y$

${n}_{y }= number\; of\; respondents\; with\; complete\; ACE\; data\; in\; year\; y$

$N= total\; number\; of\; respondents\; with\; complete\; ACE\; data\; across\; all\; five\; years$

Log-binomial regression models were fitted to estimate risk ratios (RR) for each health outcome, adjusting for age group, race/ethnicity, gender, income, and education (Supplementary Table 3). We used Quasi-Poisson regression models with a log link as a stable alternative to log-binomial models, especially in the context of non-rare prevalence of many outcomes. This approach allowed us to estimate relative risk directly while avoiding common convergence issues associated with log-binomial models [26, 27].

The ACE group was included as a categorical predictor, with three levels: None (0 ACEs, reference group), low ACE (1–2 ACEs), and high ACE (3 or more ACEs). This approach allowed for comparisons of health outcomes between individuals with and without ACE exposure. All models were adjusted for age, sex, race/ethnicity, income, and education.

All analyses were performed using R v4.3.0 [28], leveraging packages such as survey v4.4-2 [29] for complex survey design and dplyr v1.1.4 [30] for data manipulation. Figure plotting data onto U.S. maps were generated using the package usmap v0.7.1 [31]. Statistical significance was determined based on a cutoff of Benjamini-Hochberg False Discovery Rate (FDR) < 0.05.

Results

Health outcome associations with ACE

To examine the association between Adverse Childhood Experiences (ACEs) and adult health outcomes, BRFSS data from 2019 to 2023 across 33 states were analyzed. The analysis focused on individuals with complete ACE data and key covariates of interest. Table 1 presents the number of total respondents and the breakdown of missing and included ACE cases for each year from 2019 to 2023 in the BRFSS dataset. The missing percentage represents the proportion of respondents with missing ACE data each year. Across all five years, a considerable number of respondents had incomplete ACE data, with missing values ranging from 65.27% in 2020 to 87.14% in 2023. This highlights a substantial year-to-year variation in data completeness for ACE-related questions.

The final dataset included 359,507 participants, of which 38.9% reported no ACE, 36.7% reported 1–2 ACEs, and 24.4% reported 3 or more ACEs (Table 2). The mean age was 56.5 (SD= 17.5), and 54.9% were female. Race/ethnicity distribution comprised 79% White, 9% Black, 6% Hispanic, 2% AIAN, 2% Asian, and 3% Other. About 54% of the respondents earned more than $50,000 annually. State-level demographics are available in Supplementary Tables 4–36. Emotional abuse was the most reported ACE category in our dataset, with a weighted prevalence of 35%, followed by parental separation (30.9%) and substance abuse by a household member (27.4%) (Supplementary Table 37). The prevalence of each ACE category by state is also presented in Supplementary Tables 38–70.

Table 2 Sociodemographic characteristics of data collected from BRFSS conducted from 2019 to 2023. Survey results are separated by adverse childhood experience score (ACE), including no ACE (none), low ACE (1–2), and high ACE (3+). Alongside response frequency, the prevalence of the three categories of ACE is shown for each sociodemographic group in parenthesis

The geographical distribution of available data points across the U.S. is visualized in Fig. 1A. States such as Florida, Iowa, and Virginia have the highest data availability, exceeding 30,000 participants. In contrast, states represented in gray, such as Washington, California, and Oklahoma, did not contribute data for the analysis. In Fig. 1B, the spatial distribution of mean ACE scores across the U.S. is illustrated. States such as Oregon and Nevada exhibit the highest mean ACE scores, around 2.5–3, while states like North Dakota and Mississippi show lower mean scores, around 1.1–1.5. Overall, participants with 3+ ACEs were at higher risk of all the selected conditions compared to those with no ACEs, with RRs ranging from 1.15 for diabetes to 3.15 for depressive disorder, after adjusting for age, sex, race/ethnicity, income, and education (Table 3).

Fig. 1

Distribution of survey responses and mean ACE scores across the United States in BRFSS 2019-2023. Data were collected from BRFSS for the years 2019 through 2023 and selected for answers related to adverse childhood experiences. A The total number of survey responses relating to ACE information are mapped to each of the 33 states with ACE data available and displayed based on the green color gradient. B ACE scores were calculated from responses and resulting mean ACE scores per state are depicted across the US using a blue color gradient with darker shades indicating a higher average ACE score

Table 3 Prevalence and RRs of health outcomes by ACE exposure level. This table shows the weighted prevalence of selected health outcomes in the BRFSS dataset (2019-2023) and their corresponding RRs for individuals with 1–2 ACEs and 3 or more ACEs when compared to those with no ACEs. Significance was determined based on an FDR cutoff < 0.05

An increased risk of coronary heart disease was associated with both high and low ACE exposure only in North Dakota (Fig. 2). In this state, individuals with high ACE exposure had a 71% higher risk, and those with low ACE exposure had a 32% higher risk, compared to individuals with no ACE exposure. In all other states, the associations were either statistically non-significant or data were unavailable. Depressive disorder had a widespread association with ACE demonstrated by significantly increased RR in all states besides Idaho for both low and high ACE (Fig. 3). In high ACE, this risk was more dramatic, with New Jersey reaching an RR score above 5 and all other included states having a score of 2 or above (Fig. 3A). There is also a significant relationship between certain behaviors and ACE, such as being a current smoker. This relationship was significant in all included states for both low and high ACE groups, except in Kentucky (Fig. 4). The high ACE group consistently showed a higher risk of smoking behavior across states compared to the low ACE group.

Fig. 2

State-level risk ratios for coronary heart disease by ACE exposure level. A represents the state-level adjusted RRs for coronary heart disease among individuals with high ACE exposure (≥ 3 ACEs) and (B) shows RRs for low exposure (1–2 ACEs), using those with no ACEs as the reference group

Fig. 3

State-level risk ratios for depressive disorder by ACE exposure level. A represents the state-level adjusted RRs for depressive disorder among individuals with high ACE exposure (≥ 3 ACEs) and (B) shows RRs for low ACE exposure (1–2 ACEs), using those with no ACEs as the reference group

Fig. 4

State-level risk ratios for current smokers by ACE exposure level. A represents the state-level adjusted RRs for current smoking behavior among individuals with high ACE exposure (≥ 3 ACEs) and (B) shows RRs for low ACE exposure (1–2 ACEs), using those with no ACEs as the reference group

In most states, having fair or poor mental health, 14 days or more of poor physical health, COPD, and poor general health were significantly associated with high ACE scores. States in the West, Midwest, and Southern regions show a consistently significant relationship with various health outcomes in the high ACE group; however, in the low ACE group, there seems to be no clear pattern in how ACE is significantly related to chronic health conditions. Some chronic health conditions like coronary heart disease, skin cancer, diabetes, heart attack, kidney disease, other cancers, and stroke show weak and non-significant associations across multiple states. Maps demonstrating the state risk ratio for remaining health outcome associations with high and low ACE across the U.S. are available in Supplementary Figures 1–28.

To examine the extent to which health outcomes in our study may be confounding, we calculated pairwise Spearman correlations using survey-weighted data across all 17 outcomes (Fig. 5). We observed generally low to moderate correlations among the outcomes, with most coefficients falling below 0.25. A few notable associations such as depressive disorder and ≥ 14 days of poor mental health (ρ= 0.38), fair/poor general health and ≥ 14 days of poor physical health (ρ= 0.47), and between coronary heart disease and heart attack (ρ= 0.45). These findings suggest some interrelatedness among outcomes, which may partially reflect shared underlying risk factors or co-occurring conditions.

Fig. 5

Weighted Spearman correlation matrix of health outcomes. Spearman correlations were calculated for the 17 health outcomes collected from the BRFSS. Shades of red indicate positive correlation while blue indicates negative correlations. Darker shades represent stronger correlations with values ranging from −1 to 1

Race/ethnicity subgroup analysis

To gain further insight into how ACE-related health risks vary across racial/ethnic groups, we examined both the distribution of survey responses and mean ACE scores by race/ethnicity, followed by stratified risk ratio analyses. This subgroup analysis also supports an objective of this study, to explore disparities among American Indian and Alaska Native (AIAN) populations. Figure 6 illustrates the state-level distribution of survey responses and the mean ACE scores by race/ethnicity.

Fig. 6

Ethnicity-related survey responses and mean ACE scores. Survey responses and mean ACE scores were split by ethnicity to uncover any data disparities since the total survey data is heavily imbalanced by the proportion of respondents who identified as white (A). Respondents from the black population were mainly from the southeastern US although mean ACE scores were highest in the western states (B). Hispanic population responses were mainly from Florida and Texas which have a lower mean ACE score than Hispanic data collected from other states (C). AIAN responses were a low proportion of the data but demonstrated a relatively high mean ACE score across all states (D). While Asian survey responses were also a small proportion overall, they have a generally low ACE score compared to other ethnicities within the data (E). All other ethnicities were categorized as Other for the remaining respondents although this group also experienced relatively high ACE scores across all states (F)

The Black community is heavily represented in the southern states, such as Mississippi, Georgia, and Alabama, with around 20,000 participants. Specifically, states like Idaho, Oregon, Nevada, and Utah exhibited the highest mean ACE scores for blacks, while South Dakota, Arkansas, and Mississippi had lower mean ACE scores (1.0–1.5). White respondents were highly represented in states like Iowa, Virginia, and Florida, with around 30,000 participants, but sparsely represented in states like New Mexico and Kentucky. The mean ACE scores for the White population were highest in states like Oregon, Nevada, and Hawaii (2.0–3.0). Hispanics were more represented in the Southwestern states, but those in the Midwest and Southeast had higher mean ACE scores ranging from 2.0–3.0. Asian respondents were uniformly distributed across the states; however, those in the Western states reported higher ACE scores than other parts of the U.S.

To further explore racial disparities, RRs were stratified by race/ethnicity. The AIAN population, as previously mentioned, has exhibited known health disparities such as heart disease and diabetes but has been understudied in terms of ACE. Our study identified a significantly elevated risk for multiple conditions when considering the influence of ACE in AIAN individuals (Table 4). Significant associations were observed for heart attack (RR= 2.47), coronary heart disease (RR= 3.15), stroke (RR= 3.26), asthma (RR= 2.72), COPD (RR= 2.76), depressive disorder (RR= 4.39), kidney disease (RR= 2.23), arthritis (RR= 2.12), fair or poor general health (RR= 1.68), 14 or more days of poor physical (RR= 2.99), and mental health (RR= 3.73). Additionally, being a current smoker (RR= 2.93) and a heavy drinker (RR= 1.88) showed significant relationships among adults with high ACE scores.

Table 4 ACE associated health outcomes risk ratios from AIAN survey response. The BRFSS responses of the AIAN population were examined for specific health outcome associations as a group of interest, particularly in North Dakota and South Dakota

The AIAN population surveyed by BRFSS is highly represented in states like South Dakota (n=1,478) and North Dakota (n=603). This offers a unique opportunity to study a population sample from states with similar demographic and geographic characteristics whereas states like Florida may not be as comparable to the Dakotas in terms of environment. In North Dakota, high ACE scores are significantly associated with depressive disorder (RR= 4.29), and arthritis (RR= 1.85) among the AIAN community, as shown in Table 4. Other outcomes, such as heart attack and 14 or more days of poor mental health, showed elevated RR but were not statistically significant in this subgroup. In South Dakota, the health outcomes that were significantly associated with high ACE scores among AIAN respondents are other cancer (RR= 4.32), and arthritis (RR= 3.52). Other outcomes such as kidney disease (RR= 2.65), depressive disorder (RR= 1.69), and poor mental health days (RR= 2.56) showed elevated RR but did not reach statistical significance. Notably, arthritis is the only condition that was significantly associated with high ACE exposure in both North Dakota and South Dakota, aligning with national-level findings where arthritis (RR= 2.12) was also significantly linked to high ACE scores in the overall AIAN population.

Similarly, the Black population exhibited a significant relationship between high ACE scores and multiple health outcomes, such as stroke (RR= 1.63), coronary heart disease (RR= 2.10), heavy drinking (RR= 2.43), heart attack (RR= 1.58), COPD (RR= 2.24), depressive disorder (RR= 4.20), and current smoking (RR= 2.49) [32]. Regardless of the severity of ACE, the Asian population has the highest risk of experiencing skin cancer (RR= 15.27), depressive disorders (RR= 8.02), poor mental health days (RR= 5.05), and high risk of asthma (RR= 3.72) compared to other groups. Hispanic individuals are about three times as likely to be smokers (RR= 2.78) and have a high risk of depressive disorders (RR= 5.69) and poor mental health days (RR= 4.27). The "Other"racial group shows the highest risk of current smoking (RR= 3.66) and asthma (RR= 3.76). This group also had significant associations with depressive disorder (RR= 6.28) and poor mental health days (RR= 4.36). For the White population, most chronic health outcomes are significantly linked to ACE exposure, except for skin cancer, other cancers, heart attack, and diabetes in the low ACE group. Complete tables of health outcome RR information for each race/ethnic group are available in Supplementary Tables 71–76.

Discussion

The aim of this study was to explore the link between ACE and multiple chronic conditions while investigating how interactions between race/ethnicity and ACE may influence racial health disparities. This study, using the BRFSS dataset, highlights the importance of understanding how early life adversities shape chronic disease risk and health behaviors in adulthood and emphasizes the disproportionate burden of some racial groups, especially the AIAN population. Our findings conform to prior research while providing new insights into the geographic and racial diversity in ACE-related health risks.

Our study identified a significant association between high ACE scores and adverse health outcomes, which aligns with previous research findings that childhood exposure to adverse experiences raises the likelihood of health complications in adulthood [2, 4, 5, 33, 34]. This relationship was evident in our data, with consistent associations observed across a wide range of outcomes, from depressive disorders (RR= 3.15) to physical health conditions like COPD (RR= 2.16), and coronary heart disease (RR= 1.55). These results support the concept that toxic stress impacts both physiological and psychological functions [35, 36]. The most pronounced relationship identified was between ACE and mental health in both high and low ACE groups. This finding corroborates earlier research that demonstrated a link between ACE and the development of mental disorders [6, 7, 9]. Another persistent association was found between ACE and smoking behavior, which resonates with earlier studies confirming the close link between smoking and adverse childhood experiences [3, 37, 38]. A high ACE score was also a significant risk factor for heavy drinking as an adult, supporting previous research findings, which suggest that ACEs may contribute to maladaptive behaviors as a possible coping mechanism [8].

The findings presented show that ACEs affect health differently across racial and ethnic groups. High risks of heart disease and heavy drinking in Black populations may reflect the lasting effects of adversity and systemic challenges. This data supports a report published by the CDC exploring the relationship between various health outcomes and early mortality in this population [32]. Similarly, the increased mental health issues seen in Asian and Hispanic groups highlight the need for support tailored to specific groups.

One of the aims of this study was to investigate the burden of ACE-related health outcomes in the AIAN population. Our analysis revealed that the AIAN population had the highest risks of cardiovascular diseases within our data, including heart attack, coronary heart disease, and stroke, along with the second highest risk of having poor physical health. These findings are consistent with prior studies that linked ACEs to historical trauma and systemic inequalities that inordinately affect AIAN communities [35, 39, 40]. In South Dakota, high ACE scores were associated with a more than 4-fold increase in cancer risk (excluding skin cancer) among AIAN individuals, suggesting that the nexus between early adversity and limited healthcare access may exacerbate health vulnerabilities in this population [41,42,43]. The observed relationship between ACEs and various health outcomes in these populations aligns with existing evidence, reinforcing what is already known about the long-term impact of early-life adversity on health.

Our study also found geographic differences in ACE prevalence and associated health outcomes. States in the West, Midwest, and Southern regions demonstrated a consistent relationship between high ACE scores and chronic health outcomes. For instance, participants in Oregon and Nevada exhibited the highest mean ACE scores, and this correlates with an increased risk of depressive disorder, asthma, poor mental health, and being a smoker. States like Mississippi, South Dakota, and North Dakota exhibited low ACE scores in our dataset; however, these states also had the highest representation of Black and AIAN populations, respectively. These findings point to regional variations and cultural differences in reporting or exposure to ACEs and align with previous research on health disparities faced in regions with high occupancy of Black and AIAN populations [44,45,46,47].

As we discuss how ACE may influence the potential risk for adverse health outcomes, we must also acknowledge that these health outcomes are not entirely independent. Certain outcomes have clear overlapping factors that are not measured by BRFSS such as coronary heart disease and heart attack may both be directly related to an individual’s cardiovascular health [48]. To investigate the confounding between outcomes, we examined the intercorrelations between the 17 health outcomes included in the study using Spearman correlation. While a few outcomes, such as depressive disorder and poor mental health or fair/poor general health and ≥ 14 days of poor physical health, showed a moderate correlation, most relationships were weak (ρ< 0.4) to very weak (ρ< 0.2). This suggests that while some outcomes may co-occur or share underlying factors, the majority are relatively independent. However, residual confounding between health conditions remains a possibility, particularly for variables like BMI and smoking which can act as both outcomes and potential mediators due to their systemic effects on an individual’s health. Future work incorporating longitudinal data may better clarify these complex interdependencies.

Another systemic confounding factor to consider is that the BRFSS data used in this study were collected around the COVID-19 pandemic. The World Health Organization (WHO) declared the disease outbreak of a novel coronavirus as a pandemic from 11 March 2020 until 5 May 2023 [49, 50]. This period directly overlaps with the 2019-2023 data used in this study and may have influenced both data collection processes and reported health outcomes. During this time, various preventative measures were adopted such as masking, practicing good hand hygiene, social distancing, school and workplace closings, and cancellation of public events. While these practices were vital to controlling the spread of coronavirus, these societal-level changes may have also contributed to public stress in uncertain times.

Villas-Boas et al. previously used BRFSS data to assess the trend in reported depression at the onset of the pandemic and in 2021 and noted no significant average changes in depression risk at the onset but estimated a 3% increase in 2021 with meaningful differences across demographic groups [51]. Disproportionate effects across racial and ethnic groups from the pandemic have also been reported [52]. Specifically, in terms of physical activity during the pandemic, only the white population reported a decrease in physical health while black respondents reported an increase in physical activity and excellent general health. Although both white and Hispanic groups reported they were less likely to exercise, Hispanic respondents reported improved general and mental health which may reflect potentially protective cultural factors. Other studies have also noted changes in BMI, as well as drinking and smoking behaviors, with an observed increase in BMI and rate of alcohol consumption but a decreased rate of smoking [53]. As studies continue to examine data from the pandemic period, it is increasingly evident that the pandemic and its aftermath have had a complex and nuanced influence on individual physical and mental health. Pandemic-related factors may also have interacted with ACE exposure in complex ways, potentially amplifying health disparities or altering the strength of observed associations. While not directly assessed here, the broader context of the COVID-19 pandemic likely influenced the presented findings.

Overall, this study has several limitations to consider when interpreting the findings. First, both exposure and outcome variables were collected via self-report, which introduces the potential for recall bias, social desirability bias, and misclassification. The retrospective nature of the ACE module may affect the accuracy of reported childhood experiences, particularly for events occurring early in life. Similarly, self-reported health outcomes, such as those involving mental health, perception of general health, or stigmatized behaviors such as smoking, could be influenced by individual perception, health literacy, and cultural or social factors leading to underreporting or variability in interpretation [54, 55].

Additionally, while our models were adjusted for key sociodemographic variables including sex, age, education, race/ethnicity, and income, unmeasured confounding remains a possibility. Factors such as childhood socioeconomic status, familial mental health history, access to healthcare in early life, or environmental stressors were not available in the BRFSS dataset and may influence both ACE exposure and adult health outcomes, potentially biasing the observed associations. The ACE framework itself, while widely used, primarily captures family-level adversities and may not account for other significant sources of childhood trauma such as community violence or racism. This limitation is particularly important when considering health disparities in marginalized populations, where adverse experiences may not be fully captured by the standard ACE module.

Lastly, the generalizability of our findings is limited by the fact that only 33 U.S. states administered the optional ACE module between 2019 and 2023. The 17 states that were not included in this study may differ systematically in terms of socioeconomic composition, racial/ethnic diversity, health policies, or prevalence of adverse childhood experiences among other factors, from those without ACE data and excluded from this study. As a result, the associations reported here may not fully reflect national patterns.

To build on these findings, future research could expand the ACE framework to include broader forms of trauma that extend beyond family-level adverse experiences. While the cross-sectional nature of BRFSS limits the ability to infer causation or conduct longitudinal analyses, advanced machine learning models could still be applied to identify complex patterns and interactions in the data. Additionally, utilizing other datasets or study designs with longitudinal follow-up could help clarify the association between ACEs and health outcomes, as well as potentially reveal causal pathways. Incorporating additional variables beyond the typical ACE questionnaire could also provide further context to the observed associations, especially for diverse populations where underrepresented forms of trauma might play a critical role.

Conclusions

Despite these limitations, this study supports existing evidence that highlights the significant relationship between adverse childhood experiences and chronic health outcomes in the U.S. We observed geographic and racial/ethnic disparities both in mean ACE scores and in the risk ratios of associated health outcomes. Our findings underscore the need for expanded access to culturally appropriate mental health services to address systemic inequalities. States have an important opportunity to address these issues by developing trauma-informed policies tailored to their populations’ challenges with the goal of avoiding and mitigating ACE exposure.