Caregiver Report of Adverse Childhood Events: Comparison of Self-Administered and Telephone Questionnaires

Abstract

Adverse childhood experiences (ACEs) are traumatic experiences that increase people’s susceptibility to adverse physical health, mental health, and social consequences in adulthood. Screening for ACEs in primary care settings is complicated by a lack of consensus on appropriate methods for identifying exposure to ACEs. It is unclear whether self-report methods could increase disclosure of ACEs as compared to interview-based methods. This study compares data on the prevalence of ACEs from two publicly available surveys conducted on the same population of children’s caregivers: the 2019 Ohio subsample of the web/mail-based National Survey of Children’s Health and the telephone-based 2019 Ohio Medicaid Assessment Survey. We find higher disclosure of caregiver-reported child exposure to ACEs in the telephone interview survey, highlighting the importance of the role of verbal communication in developing a safe and trusting relationship in the disclosure of trauma.

1. INTRODUCTION

Adverse childhood experiences (ACEs) refer to traumatic events occurring within the first 18 years of life, such as divorce or death of a caregiver, exposure to violence in the home, or exposure to violence in the neighborhood (Bright et al., 2015; Felitti et al., 1998). Exposure to ACEs is associated with a dose-dependent increase in toxic stress and various chronic conditions including diabetes, cancer, cardiovascular disease, depression, and substance abuse, as well as decreased educational attainment and employment. (Bernstein et al., 2003; Carlson et al., 2020; Ports et al., 2020; Rariden et al., 2021). The toxic stress from ACEs can negatively affect children’s brain development, immune system function, attention, educational engagement, and ability to form healthy and stable relationships (Loveday et al., 2022; Merrick et al., 2019). Given the profound short- and long-term social and health impacts of ACEs, research efforts in this area have explored multiple methods of screening for ACEs to facilitate their early identification, intervention, or prevention (Jones et al., 2020; Krinner et al., 2021; Loveday et al., 2022; Negriff et al., 2022; Turner et al., 2020).

While the importance of screening for ACEs is understood, there is a lack of consensus and available data on the preferred modality for screening (Dubowtiz et al., 2022; Finkelhor, 2018). There is also ongoing discussion in the field about the ethics of screening and identifying adversity without the capability to provide appropriate interventions (Finkelhor, 2018; Garg et al., 2016). Identifying the screening modality with the greatest potential for disclosure can improve clinicians’ ability to intervene and mitigate the negative effects of ACEs. Single-site studies examining survey modes in assessing sensitive information (apart from ACEs exposure) have noted increased comfort and disclosure with written rather than verbal survey administration (Cullen et al., 2018; Palakshappa et al., 2019). Regarding ACEs screening, single-site studies have demonstrated mixed preference for written versus verbal screening modality (Conn et al., 2018; Kiota et al, 2018). To estimate potential differences in disclosure of ACEs between self-administered and interview-based survey modes, we compared the statewide prevalence of caregiver-reported ACEs among children included in two contemporaneous population-based health surveys, one self-administered and the other conducted via telephone. We hypothesized a higher rate of ACE exposure would be identified in the self-administered questionnaire, as compared to the telephone interview format.

2. METHODS

2.1. Data and Population

We used de-identified public data from two surveys that assessed the physical and emotional health of non-institutionalized children in Ohio. The 2019 National Survey of Children’s Health (NSCH) was a web and mail survey administered nation-wide, from which we extracted the subsample of Ohio respondents. The 2019 Ohio Medicaid Assessment Survey (OMAS) was a statewide survey on health status and health care access among residents of Ohio and was administered by telephone (including a landline sample and a cellphone sample). Methodological details for each survey are described elsewhere (RTI International, 2020; U.S. Census Bureau, 2020). The 2019 NSCH and OMAS were selected for this study because, to our knowledge, these are the two largest surveys collecting data on caregiver-reported ACEs from the same state in the same year with the same question wording and a population-based sampling method but administered via different modes.

In each dataset, we included responses representing children aged 0-17 residing in Ohio, while excluding cases with missing data on ACEs. The NSCH used address-based sampling, while the OMAS used dual frame (cellphone and landline) sampling (Noel-London et al., 2021; Sarkar et al., 2017). In households with multiple children, a single child (“sample child”) was chosen randomly (NSCH) or based on the latest birthday (OMAS) to be the subject of the survey. Both NSCH and OMAS were conducted in English or Spanish. In the NSCH and landline subsample of the OMAS, the survey was completed by a caregiver knowledgeable about the child’s health. In the cellphone subsample of the OMAS, the survey was completed by the cellphone’s owner regardless of their relation to the sample child.

2.3. Variables

In both surveys, respondents were asked if the sample child had ever experienced each of the following ACEs: caregiver divorce or separation; death of a caregiver; caregiver incarceration; violence in the home; violence in the neighborhood; mental illness in the home; substance abuse in the home; and discrimination based on race or ethnicity (Anyigbo et al., 2021). Responses for all items were scored 1 for yes and 0 for no, and the cumulative ACEs score of 0-8 was further categorized as 0, 1, 2, 3, and ≥ 4 ACEs based on prior research (Anyigbo et al., 2021). Another question about ACEs in the 2019 NSCH, regarding difficulty covering basics on the family’s income, was not included in our analysis due to the lack of an equivalent question in the 2019 OMAS. Additional sociodemographic variables were compared between the surveys if they were assessed in a similar manner in each survey and known to be correlated with exposure to ACEs, in both surveys, respondents were asked about the child’s sex, race and ethnicity (Hispanic, Non-Hispanic Black, Non-Hispanic White, or other race and ethnicity), age, and whether the child currently had no health insurance. Respondents to both surveys were also asked about the child’s overall health using a 5-point Likert scale (dichotomized for this analysis as “good”, “fair”, or “poor” versus “excellent” or “very good”).

2.4. Statistical Analysis

Data from each survey were summarized using weighted proportions or weighted means with 95% confidence intervals (CI), including exposure to any versus no ACEs, total number of ACEs as a continuous variable, total number of ACEs as a categorical variable (1, 2, 3, and ≥ 4), and exposure to each specific type of ACE. The statistical significance of differences in these estimates could not be tested directly between the two survey samples due to different sampling and weighting methods but was evaluated by comparing the overlap of the 95% CIs (White et al., 2022). All analyses included survey weights and were adjusted for the complex sampling design of each survey (RTI International, 2020; U.S. Department of Commerce, 2020). Data analysis was performed using Stata/SE 16.1 (College Station, TX: StataCorp, LP).

2.5. Ethics

Since de-identified and publicly available data were used, the study was not considered to be human subjects research and did not require Institutional Review Board approval.

3. RESULTS

The 2019 OMAS included 6,890 children ages 0-17 years, of whom 388 were excluded due to missing data on ACEs or missing survey weights. The 2019 NSCH included 558 children in the same age range who were living in Ohio, of whom 17 were excluded due to missing data. After applying survey weights, the OMAS sample was representative of 2.42 million children and the NSCH sample was representative of 2.49 million children living in Ohio. Demographic characteristics estimated based on the two samples were similar (Table 1; mean age of 8-9 years, mean household size of 4 individuals, 49%-51% female children, 72-73% non-Hispanic White children, 14% non-Hispanic Black children, 6% Hispanic children, 4-5% of children uninsured, and 11-16% of children experiencing poor, fair, or good general health, as opposed to very good or excellent health), further suggesting the two samples were representative of the same pediatric population.

Table 1.

Weighted means and proportions of children’s demographic characteristics.

Variable	OMAS (N=6,502)		NSCH (N=541)a
	Missing data (N)	Mean or proportion (95% CI)	Mean or proportion (95% CI)
Age (years)	<10	8.7 (8.6, 8.9)	8.4 (7.8, 9.0)
Household size	262	4.3 (4.2, 4.3)	4.4 (4.2, 4.5)
Sex	18
Male		0.51 (0.49, 0.53)	0.49 (0.43, 0.54)
Female		0.49 (0.47, 0.51)	0.51 (0.46, 0.57)
Race and ethnicity	72
Hispanic		0.06 (0.06, 0.07)	0.06 (0.03, 0.10)
Non-Hispanic Black		0.14 (0.13, 0.15)	0.14 (0.10, 0.20)
Non-Hispanic White		0.73 (0.71, 0.74)	0.72 (0.66, 0.78)
Other		0.07 (0.06, 0.08)	0.08 (0.05, 0.12)
Insurance status	51
Insured		0.95 (0.94, 0.95)	0.96 (0.92, 0.98)
Uninsured		0.05 (0.05, 0.06)	0.04 (0.02, 0.08)
General health status	<10
Excellent or very good		0.84 (0.83, 0.85)	0.89 (0.84, 0.92)
Good, fair, or poor		0.16 (0.15, 0.17)	0.11 (0.08, 0.16)

^aAll demographic variables in the NSCH subsample have <10 cases with missing data. Exact case counts are not reported to protect respondent confidentiality.

CI, confidence interval; NSCH, National Survey of Children’s Health; OMAS, Ohio Medicaid Assessment Survey.

Comparison of exposure to any ACEs revealed significantly higher prevalence in the telephone-based OMAS (48%; 95% CI: 46%, 49%) as compared to the self-administered NSCH (37%; 95% CI: 31%, 42%). Results were similar when comparing continuous and categorical ACEs variables (Table 2). For example, the total number of ACEs, when examined as a continuous variable, had a weighted mean of 1.0 (95% CI: 1.0, 1.1) based on the OMAS data, as compared to 0.7 (95% CI: 0.6, 0.9) based on the NSCH data. Specific types of ACEs found to be more prevalent in the telephone-based OMAS as compared to the self-administered NSCH included: parental divorce or separation, death of a caregiver, caregiver incarceration, and substance abuse in the home.

Table 2.

Weighted means and proportions of children’s exposure to ACEs in two population-based surveys.

Variable	OMAS (N=6,502)	NSCH (N=541)
	Mean or proportion (95% CI)	Mean or proportion (95% CI)
Any ACEs	0.48 (0.46, 0.49)	0.37 (0.31, 0.42)
Number of ACEs (categorical)
0	0.52 (0.51, 0.54)	0.63 (0.58, 0.69)
1	0.22 (0.20, 0.23)	0.19 (0.15, 0.24)
2	0.11 (0.10, 0.12)	0.07 (0.05, 0.11)
3	0.06 (0.05, 0.07)	0.05 (0.03, 0.09)
4+	0.09 (0.08, 0.10)	0.05 (0.03, 0.08)
Number of ACEs (continuous)	1.0 (1.0, 1.1)	0.7 (0.6, 0.9)
Specific ACEs
Divorce or separation	0.32 (0.30, 0.33)	0.24 (0.20, 0.30)
Death of caregiver	0.04 (0.04, 0.05)	0.02 (0.01, 0.03)
Caregiver incarceration	0.16 (0.15, 0.17)	0.10 (0.07, 0.15)
Violence in the home	0.09 (0.08, 0.10)	0.07 (0.04, 0.11)
Violence in the neighborhood	0.09 (0.08, 0.10)	0.06 (0.03, 0.10)
Mental illness in the home	0.16 (0.15, 0.17)	0.11 (0.08, 0.16)
Substance abuse in the home	0.15 (0.13, 0.16)	0.09 (0.6, 0.13)
Discrimination based on race or ethnicity	0.04 (0.04, 0.05)	0.04 (0.02, 0.07)

ACE, adverse childhood experience; CI, confidence interval; NSCH, National Survey of Children’s Health; OMAS, Ohio Medicaid Assessment Survey.

4. DISCUSSION

Using two state-level surveys examining caregiver-reported (as distinct from self-reported) ACEs among children in Ohio, we found a greater prevalence of ACEs in the survey using telephone-based interviews (48%, 95% CI: 46%-49%), as compared to the survey using self-administered mail or web-based data collection (37%, 95% CI: 31%-42%). Telephone interviews identified higher exposure to a wide range of ACEs, including divorce or separation, death of a caregiver, caregiver incarceration, and substance abuse in the home. These findings could be associated with residual confounding by demographic characteristics (e.g. a different proportion of children in good, fair or poor health, as seen in Table 1), or an unobserved relationship between ACE exposure and willingness to participate in surveys administered via different modes (telephone vs. mail/online). Alternately, these findings could indicate a higher probability of ACE disclosure on verbal as compared to written screening questionnaires. The latter explanation contrasts with our original expectation of higher rates of ACE disclosure on the self-administered survey, and highlights the importance of verbal communication in developing a safe and trusting relationship before disclosure of trauma.

Prior evidence on preference between verbal and written ACE screening in single-site studies has been mixed (Conn et al., 2018; Koita et al., 2018). Other research on the disclosure of sensitive health-related social needs has indicated a higher disclosure rate with written screening (Macias-Konstantopoulos et al., 2021; Palakshappa et al., 2021). In contrast to these findings, our results suggest that verbal communication in the telephone interview may have supported greater disclosure of children’s exposure to ACEs. The OMAS methodology details the training telephone interviewers with healthcare research and call center experience received to ensure fidelity in survey administration but also to normalize screening, address emotional distress, sensitivity and question refusal or avoidance (RTI International, 2020). Verbal communication may aid in establishing rapport, leading to greater disclosure rates (Bodendorfer, et al, 2020). Guidelines on ACEs screening emphasize a trauma-informed approach, empathic dialogue, and a script that normalizes the screening and creates a safe space for disclosure (Gears & Schulman, 2022; Schulman, 2019). Verbal versus written screening may also be considered if caregivers have low literacy or speak a language where the use of an interpreter to verbally ask the questions may be more suitable.

Beyond identifying exposure to ACEs to monitor population health, one potential benefit of verbal screening for ACEs in clinical settings is its utility in building trust in the patient’s clinician and enhancing the patient’s overall healthcare experience (Rariden et al., 2021). This relational component of interview-based screening might be particularly important for trauma disclosure in settings with a higher prevalence of trauma (Crenshaw et al., 2021, Ereyi-Osas et al., 2020). Along with increased disclosure rates, patients are likely to experience more support, validation, and resilience from the clinical discussion that follows ACE screening, especially when done by an empathetic clinician who is trained in trauma-informed care (SmithBattle et al., 2021). The trauma-informed care model encourages adequate training, active listening, verbal communication, developing a safe and trusting relationship, and applying shared decision-making to use the discussion as a bridge to connect to services and resources (Barry & Gundacker, 2023; Marcoux, 2021; Musicaro & Langer, 2022). Having dialogue about trauma can require additional time. In clinical practice, there are continued efforts to strike a balance between screening modalities that efficiently and effectively identify adversity improving the connections to resources, and modalities that avoid emotions such as shame and guilt. However, further research is needed to confirm these potential advantages of verbal over written ACEs screening in clinics serving pediatric patients and investigate further differences in disclosure rates based on the setting of the verbal screening and the role of the interviewer (e.g., physician, nurse, social worker, or other staff member).

Our study revealed an unexpected difference in the disclosure of ACEs based on verbal versus self-administered survey mode, but our conclusions are subject to several limitations. Most importantly, our analysis assumes that the OMAS and NSCH data are both generalizable to the same population (i.e., non-institutionalized children ages 17 and under living in Ohio during 2019). While we have demonstrated similar estimates of population demographics based on questions asked in a consistent manner between the 2 surveys, we excluded questions about other relevant socioeconomic confounders, such as household income and the primary caregiver’s educational attainment, that were not asked in the same way in both surveys. Data available from the public dashboards for each survey suggest that children sampled by OMAS lived with caregivers who had lower income than those sampled by NSCH (weighted estimate of 24% of children living below the poverty line based on OMAS data, compared to 20% based on NSCH data), which could potentially correlate with higher exposure to ACEs (Child and Adolescent Health Measurement Initiative, n.d.; The Ohio Medicaid Assessment Survey Dashboard, n.d.).

Other limitations of our study include the use of data from a single US state where we identified nearly contemporaneous surveys with similar methodology (apart from survey mode) for querying children’s exposure to ACEs, so results may not be generalizable to other states, including states with greater racial or ethnic diversity. We were limited to using ACE questions that were asked in both surveys and could not examine other domains of ACEs covered in expanded ACEs survey study (Cronholm et al., 2015). The NSCH Ohio subsample was also much smaller than the OMAS sample, limiting the precision of the estimates generated from the NSCH and thus potentially missing other statistically significant differences between the two sets of survey estimates. The OMAS had a lower completion rate than the NSCH and included a small proportion of respondents (approximately 5%) who had the option of completing the survey online, as part of a pilot sub-study. The latter subgroup could not be identified or excluded in the public-use data. In both surveys, the measures of ACEs relied on caregiver reports, and it is unknown how well these data would correlate with youth self-report. As noted above, the context of a population-based health survey is different from a clinical setting where ACE screening might be conducted, and it remains to be seen whether these results would be applicable to screening performed in healthcare facilities.

Despite these limitations, our results demonstrate the importance of considering the modality of ACEs screening. While there is a current emphasis on patient-reported questionnaires to assess adversity, our results highlight the need for caution when using written tools exclusively to screen for ACEs. Unlike written questionnaires, verbal communication can be used to develop a safe and trusting relationship, which may increase ACE disclosure. Our analysis indicates a possible improvement in disclosure with verbal rather than written questionnaires, but further work is needed to evaluate the role verbal as compared to written communication in this process.

Highlights.

• Adverse childhood experiences (ACEs) screeners can be verbal or self-administered.

• We used 2 contemporaneous surveys to compare prevalence of caregiver-reported ACEs

• The telephone survey had higher ACEs disclosure rates compared to paper/web survey.

• Findings support the importance of verbal communication to identify ACEs exposure.

Abbreviations:

(ACEs)

Adverse childhood experiences

(NSCH)

National Survey of Children’s Health

(OMAS)

Ohio Medicaid Assessment Survey