The impact of adverse childhood experiences on age of diabetes diagnosis and associations with race and ethnicity

Diabetes diagnosis, age of diabetes diagnosis, and survey completion

For this study, we included BRFSS participants reporting a diagnosis of diabetes who responded to the ACEs survey module. After identifying individuals meeting the inclusion criteria, we extracted data from the question, “How old were you when you were told you had diabetes?” Because ACEs are linked to comorbidities later in life, we excluded individuals who reported receiving a diabetes diagnosis before the age of 18 [1]. Individuals lacking responses for any question were excluded from the analysis. Gestational diabetes was not included.

Adverse childhood experiences

The BRFSS ACE Module consists of 13 questions – 11 that evaluate ACEs and 2 that evaluate Protective and Compensatory Experiences (PACEs) before the age of 18 and can be located within the codebook available at https://www.cdc.gov/brfss/data_documentation/index.htm. This module is available in 16/50 US states, including Alabama, Arkansas, Iowa, Kansas, Maine, Mississippi, Nevada, New Hampshire, New Jersey, New York, North Dakota, Ohio, Oregon, South Carolina, Virginia, and Wisconsin. For this study, we utilized the first 11 questions, which cover the domains shown in Table 1 [20]. We re-coded each item to be a binary variable – either having occurred (1) or not (0) because some responses included a degree of frequency: never happened, occurred once, or more than once. We then summed the number of different ACEs that each participant reported experiencing.

Control variables

Due to the data collection methods of BRFSS, all variables including sociodemographic variables were derived from self-reported data. These variables included sex, age, race, and ethnicity. Sex was reported as male or female. Age was reported in three age groupings: 45–54, 55–64, and 65 years and older. Racial groups, provided in the imputed variable in BRFSS, included “White non-Hispanic,” “Black,” “Asian,” “American Indian/Alaska native” (AI/AN), “Hispanic,” and “other race.” Education was also extracted from BRFSS and included “did not graduate high school,” “graduated high school,” “attended college or technical school,” and “graduated from college or technical school.”

Statistical analysis

For all prevalence estimates and analyses, we employed a survey design and sampling weights provided by BRFSS. First, we report the prevalence of individuals with diabetes overall and then report the sociodemographics profile among those with diabetes by the number of ACEs experienced – categorized as 0, 1–3, and 4+, and we report the mean (standard deviation [SD]) age of diagnosis of diabetes by demographics. Next, we visualized the distributions of the reported age of diabetes diagnosis utilizing histograms. We also constructed a histogram contrasting the distribution of age at T2DM diagnosis among those with 0 ACES and 4+ ACES. Finally, we utilized binary and multivariable regression analyses to determine the association between the age of diabetes diagnosis and cumulative ACEs with a significance threshold of 0.05. Statistical analyses were conducted utilizing Stata 16.1 (StataCorp, LLC, College Station, TX). This study was not submitted for ethics review to an institutional review board (IRB) oversight because it did not meet the regulatory definition of human subject research as defined in 45 CFR 46.102(e) of the Department of Health and Human Services’ Code of Federal Regulations. This study adhered to the STrengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines.

ACEs and age of diabetes diagnosis

From the bivariate regression analysis (Table 3), we found that compared to individuals reporting no ACEs, those with 1–3 ACEs had an age at diagnosis 2.21 (SE=0.48, p<0.001) years earlier, and those with four or more ACEs were diagnosed 6.37 (SE=0.62, p<0.001) years earlier (Figures 1 and 2). From the adjusted model, these values were 2.15 (SE=0.47, p<0.001) and 6.37 (SE=0.61, p<0.001) years for individuals with 1–3 and 4+ compared to individuals with no ACEs. Compared to White individuals, the cumulative difference in age at diagnosis for individuals varied across ethnoracial groups; however, it was consistently earlier for all groups. The diagnosis age was 3.79 (SE=0.58, p<0.001) years earlier for Black individuals, 8.90 (SE=2.33, p<0.001) years earlier for Asian individuals, 4.69 (SE=1.88, p=0.013) years earlier for AI/AN individuals, 7.92 (SE=1.02, p<0.001) years earlier for Hispanic individuals, and 1.66 (SE=1.77, p=0.35) earlier for ‘other race’ individuals. When compared to individuals who did not graduate HS, individuals who graduated HS were diagnosed 0.44 (SE=0.94, p=0.64) years later, and individuals who attended college/technical school were diagnosed 0.04 years later (SE=0.95, p=0.97), whereas individuals graduating college/technical school were diagnosed 0.43 (SE=0.96, p=0.65) years earlier. Finally, women were diagnosed 0.42 (SE=0.45, p=0.34) years later than men.

Figure 1:

A histogram of the age at diabetes diagnosis among all included participants.

Figure 2:

Contrasting histograms of the age at diabetes diagnosis among those with zero adverse childhood experiences (ACEs) compared to those with 4+ ACEs.

We also assessed the summative difference in age of diabetes diagnosis by ethnoracial group and ACE category (e.g., 0 ACEs, 1–3 ACEs, and 4+ ACEs) from Table 3. When compared to White, non-Hispanic individuals with no reported ACEs, Black individuals were diagnosed 5.49 (SE=0.9, p<0.001) years earlier, Asian individuals were diagnosed 8.95 (SE=2.87, p=0.002) years earlier, Hispanic individuals were diagnosed 9.36 (SE=2.21, p<0.001) years earlier, and ‘other race’ individuals were diagnosed 1.73 (SE=2.53, p=0.495) years earlier. In contrast to the other ethnoracial minority groups, AI/AN individuals with 0 ACEs were diagnosed 2.59 (SE=2.16, p=0.229) years later compared to White individuals.

White individuals with 1–3 ACEs were diagnosed with diabetes 2.74 (SE=0.54, p<0.001) years earlier than White individuals with no reported ACEs. Additional ethnoracial groups with 1–3 ACEs were similarly diagnosed earlier. Black individuals within this group were diagnosed 5.15 (SE=0.88, p<0.001) years earlier with AI/AN individuals being diagnosed 7.64 (SE=2.17, p<0.001) years earlier, and ‘other race’ individuals being diagnosed 3.94 (SE=2.84, p=0.166) years earlier. Of note, Asian and Hispanic individuals with 1–3 ACEs saw earlier diagnoses of over a decade with Asian individuals being diagnosed 12 (SE=4.87, p=0.014) years earlier, and Hispanic individuals being diagnosed 10.63 (SE=1.44, p<0.001) years earlier.

Finally, when assessing individuals with 4+ ACEs, we found that White, non-Hispanic individuals were diagnosed 6.75 (SE=0.72, p<0.001) years earlier compared to White non-Hispanic individuals with no reported ACEs. This trend continued with Black individuals being diagnosed 10.44 (SE=1.23, p<0.001) years earlier, Asian individuals being diagnosed 15.65 (SE=5.56, p=0.005) years earlier, AI/AN individuals being diagnosed 12.25 (SE=3.24, p<0.001) years earlier, Hispanic individuals being diagnosed 12.96 (SE=1.41, p<0.001) years earlier, and ‘other race’ individuals being diagnosed 9.10 (SE=3.86, p=0.019) years earlier.

Implications and recommendations

Current guidelines for T2DM screening, starting at age 35, may critically miss a large portion of individuals before this targeted age [44], 45]. Similarly, studies have shown that interventions targeting proximal determinants of adversity may reduce the prevalence of T2DM in adults [46], showing that earlier screenings for both ACEs and diabetes may benefit chronic disease prevention. A systematic review by Loveday et al., [47] published in 2022, demonstrates the efficacy of pediatric screening for exposure to ACEs in detecting and promoting early intervention before the diagnosis of disease. It is recommended that these screenings occur within a primary care setting to allow for consistent interaction and a trusting provider/patient relationship [48]. In addition, screening both the child and the caregiver for ACEs at varied intervals during well-child visits might address both the impact of intergenerational trauma and the impact of ACEs on health outcomes [49]. Physicians must be prepared for action when the occurrence of specific ACEs are disclosed by the patient or patient’s guardian and the potential signs of ACEs when they are not disclosed but are apparent [50], [51], [52], [53]. Early identification can assist in directing individuals with increased ACE burdens to responsive plans and psychoeducation on known protective factors [49]. Furthermore, evidence of a graded association between ACEs exposure and adolescents’ mental health levels has been identified – particularly regarding substance use disorder [54]. Therefore, connecting adolescents with mental health resources is recommended to prevent substance use in adolescents with four or more ACEs [54].

Unfortunately, the diagnosis of T2DM continues to rise in both youth and adults and is more common within lower socioeconomic populations as are ACEs [55], 56]. The impact of ACEs may be further compounded for this population due to food insecurity, poorer health literacy, and limited exercise opportunities [57]. Therefore, even with mental health therapy targeted toward ACEs, interventions promoting healthy lifestyle changes may be warranted [58]. Given the nature of ACEs, policies should be wide-ranging and include early initiatives to strengthen family economic stability such as tax credits, increasing existing social support systems such as the Supplemental Nutrition Assistance Program (SNAP), and child care subsidies [59]. Additionally, previous evidence has suggested that Medicaid expansion is associated with a reduction in reports of child neglect, the most prominent form of maltreatment in the United States [60], 61]. Other policies could include family-friendly workplace initiatives for parents including paid family leave as well as flexible scheduling [59]. The current child welfare system in the United States is generally response-driven, meaning that families do not come into contact with its various agencies until after an allegation of maltreatment has occurred [62]. As such, additional policies should expand social support services to allow for the bolstering of child welfare services to allow for comprehensive support services that families can access.

It is important to note that the current research findings cannot support a direct biological link between specific identities and the diagnosis of T2DM. Instead, the findings highlight that each unique ethnoracial group may be experiencing different disparities regarding the development and diagnosis of the disease due to present-day social determinants of health that are rooted in historical mistreatment and oppression [25], [26], [27], [28], [63], 64]. This is an important distinction because groups have different forms of historical trauma as well as culturally protective factors. As such, a one-size-fits-all approach to mitigating the development of T2DM may not be beneficial for every group. Given this, further research and policy work are needed that work specifically with these communities to determine best practices at the community level.

Because the tenets of osteopathic medicine place a special emphasis on somatic, visceral, and psychological medicine – and prioritize “whole-person,” patient-centered care considering numerous factors when treating a patient, not just empirical values – osteopathic clinicians should be aware of the presences of abuse, trauma, or family dysfunction to optimally serve their pediatric patients. Thus, when assessing an individual for T2DM, physicians should incorporate the patient’s lived experience and hardships when diagnosing and formulating a treatment plan [13].

Limitations

Limitations within BRFSS data include the limited number of individuals who responded to the diabetes questions and completed the ACEs modules; however, the resulting sample was sufficiently large to power our analyses. Because there was a limited number of states that shared the ACEs module responses with the CDC to incorporate them into publicly available data, generalization among the entire US population may be impaired. Additional populations not surveyed included households without landlines or phone access and those who live in a group setting, such as nursing homes, military bases, or prison. Another limitation is that the reported age of diagnosis did not discriminate between diabetes types; however, we did find that among individuals reporting diabetes, only 2.9 % of the sample reported being diagnosed before the age of 18, which is in line with the reported prevalence of Type 1 diabetes mellitus and would have very limited impact on our results. With concern to ACE screening, and acknowledging that it can be a vital public health tool, previous evidence has called into question the applicability of utilizing ACEs to determine disease risk at the individual level [65]. Additionally, our study is correlational, rather than causal, and longitudinal cohort studies may be necessary to further assess the linkages under investigation here. Furthermore, due to the self-reporting nature of the BRFSS data set, a tendency to report a healthier lifestyle may occur. Lastly, it is important to note that the United States Preventive Services Task Force (USPSTF) does not recommend screening for ACEs due to insufficient evidence on the benefits and harms of universal screening in primary care. Screening is recommended by the USPSTF for depression, anxiety, and intimate partner violence, which overlap with ACEs.