Associations between adverse childhood experiences, attitudes towards COVID-19 restrictions and vaccine hesitancy: a cross-sectional study.

OBJECTIVES: Adverse childhood experiences (ACEs) can affect life-course health and well-being, including risk-taking behaviour and trust. This study explored associations between ACEs and trust in health information on COVID-19, attitudes towards and compliance with COVID-19 restrictions and vaccine hesitancy.
DESIGN: National cross-sectional telephone survey using a sample of landline and mobile numbers stratified by Health Board, deprivation quintile and age group.
SETTING: Households in Wales during national COVID-19 restrictions (December 2020 to March 2021). PARTICIPANTS: 2285 Welsh residents aged ≥18 years. MEASURES: Nine ACEs; low trust in National Health Service (NHS) COVID-19 information; supporting removal of social distancing and mandatory face coverings; breaking COVID-19 restrictions; and vaccine hesitancy (rejection or uncertainty of vaccination).
RESULTS: Increasing ACE counts were independently related to low trust in NHS COVID-19 information, feeling unfairly restricted by government and ending mandatory face coverings. High ACE counts (4+ vs 0 ACEs) were also associated with supporting removal of social distancing. Breaking COVID-19 restrictions increased with ACE count with likelihood doubling from no ACEs to 4+ ACEs. Vaccine hesitancy was threefold higher with 4+ ACEs (vs 0 ACEs) and higher in younger age groups. Thus, modelled estimates of vaccine hesitancy ranged from 3.42% with no ACEs, aged ≥70 years, to 38.06% with 4+ ACEs, aged 18-29 years.
CONCLUSIONS: ACEs are common across populations of many countries. Understanding how they impact trust in health advice and uptake of medical interventions could play a critical role in the continuing response to COVID-19 and controlling future pandemics. Individuals with ACEs suffer greater health risks throughout life and may also be excluded from interventions that reduce infection risks. While pandemic responses should consider how best to reach those suffering from ACEs, longer term, better compliance with public health advice is another reason to invest in safe and secure childhoods for all children. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

A large national sample surveyed during a period of national COVID-19 restrictions.

Although not unusual for unsolicited telephone surveys, the participation level was 36.4%, creating a potential for a self-selection bias among respondents.

Prevalence of adverse childhood experiences (ACEs) reported was consistent with other comparable population surveys, including those undertaken face to face.

ACEs were self-reported and measured retrospectively, and therefore, may have been misremembered or otherwise misreported.

Outcomes investigated both measures of trust and preference for different health regulations and restrictions as well as measures of behaviour.

Introduction

In many countries, the control of COVID-19 has relied on public acceptance of, and compliance with, restrictions on travel, work, socialising and public behaviour.1 Medical advice provided through governmental and health professional bodies has formed the principal mechanism for encouraging social isolation, mask wearing and other COVID-19 prevention measures. Although restrictions are often reinforced with fines and other judicial measures, their implementation still depends heavily on public support.2–4 Moreover, despite some discussion on mandatory vaccination, the success of this emergent COVID-19 control measure also relies on individuals having confidence in and complying with health messaging.5 Consequently, it is critical for COVID-19 control to understand what factors differentiate individuals who may or may not trust health information, adhere to behavioural advice or accept offers of vaccination. Such understanding can inform the development and targeting of future measures to maximise behavioural compliance and vaccine uptake in different population groups.

Adverse childhood experiences (ACEs) include child maltreatment (physical, psychological, sexual and neglect) and other sources of chronic trauma in childhood, such as growing up in a household affected by domestic violence, substance use and other criminal justice problems.6 Multiple studies have shown strong relationships between experiencing more types of ACEs and the development of health-harming behaviours such as smoking, harmful alcohol use and illicit drug use, as well as increased involvement in antisocial behaviour and violence.7 8 Although suffering ACEs is not deterministic, higher exposure to ACEs is related to a greater likelihood of developing chronic health conditions such as cancer, cardiovascular disease, type 2 diabetes and respiratory diseases.7 9–11 Thus, individuals with ACEs may be at greater risks of COVID-19-related morbidity and mortality through higher vulnerability resulting from behaviours such as smoking12 and conditions such as obesity13 and diabetes.14 ACEs are also associated with substantive increases in poor mental health.7 15 Moreover, more limited research suggests maltreatment during childhood may leave individuals with lower levels of trust including in health and other public services.16 17 What is less well studied is whether a history of ACEs impacts compliance with advice and instruction from public health and healthcare systems. Around half of adults in Europe and North America have experienced at least one ACE with estimates suggesting around a quarter have suffered multiple ACEs.18 Consequently, it is important to understand and address any impact of ACEs on compliance with COVID-19 controls in order to avoid repercussions both for the health of those with ACEs and for infection risks in their local communities.

Here, we examine relationships between a history of childhood adversity and current levels of trust in health systems information, support for and compliance with COVID-19 control restrictions, and intention to be COVID-19 vaccinated. We hypothesise that, independent of sociodemographics, exposure to more ACEs will be associated with less trust in health systems, lower support for governmental restrictions intended to control COVID-19 transmission and higher vaccination rejection rates (termed here vaccine hesitancy). We examine these relationships through a national anonymous telephone survey of adults in Wales. Finally, we explore how measures to influence public behaviour might better support those who have suffered ACEs with respect both to COVID-19 and preparing for other future pandemics.

Methods

Data collection

A national telephone survey of Welsh residents aged 18 years and over was conducted between December 2020 and March 2021. Although pilot data were collected on 15 December and 16 December, final survey data collection all occurred within a period of consistent national COVID-19 restrictions in Wales. Thus, a national lockdown including orders to stay at home and mandatory closure of non-essential retail, hospitality sectors and gyms was established 20 December 2020 with relaxation of restrictions beginning predominantly from 13 March 2021.19 Mixing of two households indoors was permitted for just 25 December 2020 but no data collection occurred on this day. A minimum target sample of 2000 was set to capture adequate individuals across ACE categories, with a minimum of 200 respondents in the highest ACEs category (4+).17 A professional market research company (MRC) was commissioned to undertake sampling and data collection. Landline and mobile telephone contacts were obtained from a commercial sample provider stratified by Welsh Health Board area, residential deprivation quintile (using Welsh Index of Multiple Deprivation (WIMD)20) and age group, to attain a sample broadly representative of the age, deprivation and geographical profile of the Welsh population.

Study inclusion criteria were Welsh resident aged 18 years or over and cognitive ability to participate in a telephone interview. Potential participants were given a verbal description of the study including its purpose and voluntary, anonymous and confidential nature. Participants were informed they could skip or decline questions, withdraw at any point and that a decision to stop would not affect their rights, health treatment or service provision. Informed consent was recorded using opt-in consent. A web address was provided to participants containing further study information and links to appropriate support services. All study materials were provided in English and Welsh and participants could complete the interview in either language. Telephone calls were made across all days of the week between the hours of 9:00–21:00 hours on weekdays and 10:00–16:00 hours on weekends, and interviews took on average 20 min to complete.

Contact was made with 6763 individuals, of whom 98 (1.4%) were ineligible, 4062 (60.1%) declined and 2603 agreed to participate in the study. Of those who agreed, 277 did not meet the age quota in their area and 2326 completed the questionnaire, with 64.7% of respondents being female. Thus, the participation rate was 36.4% (2326/6388) of eligible individuals who met the quota sampling, or 34.9% (2326/6665) of all eligible participants. The sample used for analysis here was limited to participants who answered all questions of interest (N=2285).

Study questionnaire

The study questionnaire included questions on participant demographics, ACEs, health conditions, trust in information on COVID-19 from the National Health Service (NHS), and attitudes towards COVID-19 restrictions and vaccination. All measures were self-reported. The full questions and response options used to measure ACEs and the outcomes included in this study are provided in online supplemental appendix table A1.

Nine ACE types before the age of 18 years (physical, verbal and sexual abuse; parental separation; exposure to domestic violence; and living with a household member with mental illness, alcohol abuse, drug abuse or who was incarcerated) were measured using an adapted version of the Centers for Disease Control and Prevention short ACE tool.21 In line with international literature,7 responses to the nine ACE questions were used to calculate an ACE count (0 ACEs, 1 ACE, 2–3 ACEs, 4+ ACEs). Such categorisation has enabled: comparative examination of individuals exposed to lower, mid, and higher counts of ACEs; a more consistent approach to analyses between ACE studies; and combined analyses of findings from different studies.7 Low trust in NHS COVID-19 information was measured by a question asking how much participants would trust information on COVID-19 from the NHS (scale 0=not at all, 10=completely; low <6). Feeling unfairly restricted a lot by government was identified by a response of ‘yes, a lot’ to a question asking if, during the pandemic, participants felt they had been unfairly controlled by the national restrictions imposed by the government. Beliefs that mandatory face coverings should go and social distancing should end were measured with questions asking if face coverings in shops should continue to be a legal requirement (qualifying response ‘no’) and if social distancing should remain in place or be removed (qualifying response ‘be removed’), respectively. Participants were asked if, during lockdown or local COVID-19 restrictions, they had always followed the advice, bent or broken the rules occasionally, or largely ignored the rules; those providing either of the latter two responses were categorised as break restrictions at least occasionally. Vaccine hesitancy was identified by responses of ‘no’ or ‘unsure’ to a question asking if participants would want to receive a COVID-19 vaccination. Participants were categorised as having had COVID-19 if they responded ‘yes’ to a question asking if they thought they have had, or currently have, coronavirus; and as having had a chronic disease if they reported having ever been told by a doctor or nurse that they had any of the following conditions: cancer, type 2 diabetes, heart disease (coronary heart disease, heart attack or stroke) or respiratory disease (chronic bronchitis, emphysema, chronic obstructive pulmonary disease, asthma).

Sex (male; female; other), ethnicity (self-defined using UK census categories) and postcode of residence were also collected. For the purposes of anonymity and consistent with previous studies, respondents’ age was collected in 5-year age groups but combined into 10-year age categories (18–29, 30–39, 40–49, 50–59, 60–69, 70+) in order to ensure sufficient numbers in each category for analysis. Due to low levels in non-white categories, ethnicity was recategorised (white, other). Postcode was categorised into deprivation quintile by the MRC using the WIMD (1=least deprived to 5=most deprived).

Statistical analysis

Statistical analyses used SPSS V.27. Cross-tabulations and χ2 tests were used to measure relationships between outcome variables, and to examine initial relationships between outcome variables and ACEs and other participant characteristics (age, sex, ethnicity, deprivation, COVID-19 infection and chronic disease). Independent associations between ACEs and outcomes were measured using logistic regression, controlling for other participant demographics. Having had COVID-19 was included in the model as it was hypothesised that individuals who report this may feel protected from the virus.22 Similarly suffering from a chronic disease was included in the model as individuals with a chronic disease may feel more at risk of the virus. Finally, the estimated adjusted proportions (estimated marginal means; EMMs) reporting breaking restrictions at least occasionally and vaccine hesitancy in different ACE categories and age groups were generated from the final logistic regression models.

Patient and public involvement

The study did not involve patients. Study findings are being made publicly available to participants and the general public through the production of study reports and open access journal articles. The study webpages provided contact details for the research team if any individual wished to directly request publications.

Results

Approximately half of participants reported having experienced no ACEs (51.86%) with proportions in the 1 ACE, 2–3 ACEs and 4+ ACE categories being 21.40%, 16.46%, and 10.28%, respectively. A breakdown of participant demographics by ACE count is shown in online supplemental appendix table A2. Respondents’ views of having low trust in NHS COVID-19 information and being unfairly restricted a lot by government were associated with higher levels of favouring the immediate cessation of social distancing and mandatory face coverings, breaking restrictions and vaccine hesitancy (table 1). For example, 42.11% of those reporting low trust in NHS COVID-19 information also reported vaccine hesitancy, compared with just 5.62% of those without such low trust.

Table 1

Relationships between views on fairness of restrictions, trust in NHS COVID-19 information and support for COVID-19 control and vaccination measures

		Low trust in NHS COVID-19 information (%)	Unfairly restricteda lot by government (%)	Social distancing should end(%)	Mandatory face coverings should go (%)	Break restrictions at least occasionally (%)	Vaccine hesitancy (%)
All (n=2285)		5.82	9.41	5.91	5.82	25.86	7.75
Low trust in NHS COVID-19 information	No		7.81	4.51	4.32	24.86	5.62
	Yes		35.34	28.57	30.08	42.11	42.11
	X2		111.387	130.480	151.552	19.426	233.296
	P value		<0.001	<0.001	<0.001	<0.001	<0.001
Unfairly restricted a lot by government	No*	4.15		3.91	3.77	24.40	5.80
	Yes	21.86		25.12	25.58	40.00	26.51
	X2	111.387		157.517	169.061	24.732	116.950
	P value	<0.001		<0.001	<0.001	<0.001	<0.001

See online supplemental appendix table A1 for full wording of all questions and classification of responses.

*Includes those who responded ‘yes, a little’.

NHS, National Health Service.

Low trust in NHS COVID-19 information

Individuals with higher ACE counts were more likely to have low trust in NHS COVID-19 information along with individuals from more deprived quintiles of residence (table 2). Other sociodemographics and a history of either chronic disease or COVID-19 infection were not significantly associated with low trust. When using logistic regression to control for confounding relationships, ACEs and deprivation were the only significant predictors of trust in NHS COVID-19 information (table 3).

Table 2

Adverse childhood experiences (ACEs), sociodemographics, other individual characteristics and associations with compliance with, trust in and support for COVID-19 control measures

		n	Low trust in NHSCOVID-19 information(%)	Unfairly restricteda lot by government (%)	Social distancing should end (%)	Mandatory face coverings should go (%)	Break restrictions at least occasionally (%)	Vaccine hesitancy(%)
ACE count	0	1185	4.05	7.26	4.39	3.46	20.93	4.98
	1	489	5.73	8.59	5.73	6.34	28.02	7.16
	2–3	376	7.45	12.23	5.59	7.45	31.12	10.11
	4+	235	12.34	17.45	14.47	14.04	37.87	19.15
	X2		26.817	28.154	35.999	43.081	39.321	58.625
	P value		<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Age(years)	18–29	174	6.90	16.09	10.92	9.77	47.70	18.39
	30–39	239	7.11	12.97	9.21	12.97	35.98	15.90
	40–49	371	6.20	10.78	9.16	8.89	28.84	9.16
	50–59	543	5.16	9.21	6.45	5.16	26.34	7.00
	60–69	447	4.70	7.83	3.13	3.36	23.04	4.70
	70+	511	6.26	6.07	2.15	1.76	13.50	2.74
	X2		2.839	21.525	39.054	54.389	100.389	75.027
	P value		0.725	0.001	<0.001	<0.001	<0.001	<0.001
Deprivation quintile	Least 1	495	4.44	9.29	4.65	4.24	26.67	4.44
	2	509	3.73	5.70	4.13	4.32	26.13	5.50
	3	490	5.10	8.16	6.33	5.51	28.37	7.76
	4	437	8.70	11.67	7.55	6.86	25.63	9.84
	Most 5	354	8.19	13.84	7.63	9.32	21.19	12.99
	X2		16.440	19.909	8.485	13.207	5.838	27.466
	P value		0.002	0.001	0.075	0.010	0.212	<0.001
Sex	Male	806	6.95	10.67	7.69	8.06	27.79	8.56
	Female	1479	5.21	8.72	4.94	4.60	24.81	7.30
	X2		2.887	2.322	7.131	11.438	2.412	1.156
	P value		0.089	0.128	0.008	0.001	0.120	0.282
Ethnicity	White	2254	5.77	9.36	5.90	5.81	25.87	7.59
	Other	31	9.68	12.90	6.45	6.45	25.81	19.35
	X2		0.853	0.450	0.017	0.023	0.000	5.926
	P value		0.356	0.502	0.897	0.880	0.994	0.015
HadCOVID-19*	No	1837	5.50	8.49	5.50	5.39	24.39	6.80
	Yes	448	7.14	13.17	7.59	7.59	31.92	11.61
	X2		1.777	9.245	2.833	3.180	10.656	11.625
	P value		0.182	0.002	0.092	0.075	0.001	0.001
Chronic disease†	No	1488	5.51	9.68	6.45	7.06	29.50	8.40
	Yes	797	6.40	8.91	4.89	3.51	19.07	6.52
	X2		0.747	0.360	2.267	11.887	29.452	2.556
	P value		0.387	0.548	0.132	0.001	<0.001	0.110

*Having had COVID-19 was self-reported, see the Methods section.

†Chronic diseases included cancer, type II diabetes, heart disease and respiratory diseases, see the Methods section for details. Full wording of all questions is provided in online supplemental appendix table A1.

NHS, National Health Service.

Table 3

Logistic regression analysis of relationships between adverse childhood experiences (ACEs), sociodemographics and other individual characteristics and compliance with, trust in and support for COVID-19 control measures

		Low trust in NHSCOVID-19 information			Unfairly restricted a lot by government			Social distancing should end			Mandatory face coverings should go			Break restrictions at least occasionally			Vaccine hesitancy
		AOR	95% CIs	P value	AOR	95% CIs	P value	AOR	95% CIs	P value	AOR	95% CIs	P value	AOR	95% CIs	P value	AOR	95% CIs	P value
ACE count	0	Ref		<0.001	Ref		0.002	Ref		<0.001	Ref		<0.001	Ref		<0.001	Ref		<0.001
	1	1.46	0.90 to 2.37	0.123	1.13	0.76 to 1.67	0.549	1.22	0.76 to 1.98	0.410	1.80	1.10 to 2.94	0.019	1.34	1.05 to 1.73	0.020	1.29	0.83 to 2.02	0.254
	2–3	1.82	1.11 to 2.99	0.017	1.50	1.02 to 2.21	0.042	1.01	0.60 to 1.72	0.963	1.76	1.06 to 2.93	0.029	1.48	1.13 to 1.94	0.004	1.56	1.01 to 2.43	0.047
	4+	3.22	1.94 to 5.36	<0.001	2.19	1.44 to 3.33	<0.001	2.89	1.79 to 4.68	<0.001	3.57	2.14 to 5.94	<0.001	2.01	1.47 to 2.76	<0.001	3.11	2.00 to 4.82	<0.001
Age(years)	18–29	Ref		0.819	Ref		0.112	Ref		<0.001	Ref		<0.001	Ref		<0.001	Ref		<0.001
	30–39	1.05	0.48 to 2.30	0.897	0.78	0.44 to 1.37	0.378	0.88	0.45 to 1.71	0.706	1.51	0.79 to 2.88	0.214	0.66	0.44 to 0.98	0.041	0.82	0.48 to 1.40	0.465
	40–49	1.08	0.52 to 2.24	0.847	0.71	0.41 to 1.20	0.198	0.96	0.52 to 1.77	0.900	1.13	0.60 to 2.14	0.702	0.49	0.34 to 0.72	<0.001	0.51	0.30 to 0.88	0.015
	50–59	0.92	0.45 to 1.88	0.819	0.63	0.38 to 1.05	0.074	0.67	0.37 to 1.23	0.197	0.67	0.35 to 1.27	0.219	0.44	0.31 to 0.63	<0.001	0.41	0.24 to 0.69	0.001
	60–69	0.89	0.42 to 1.89	0.753	0.55	0.32 to 0.96	0.034	0.32	0.15 to 0.66	0.002	0.46	0.22 to 0.96	0.038	0.38	0.26 to 0.56	<0.001	0.29	0.16 to 0.53	<0.001
	70+	1.29	0.62 to 2.69	0.495	0.45	0.25 to 0.79	0.006	0.23	0.10 to 0.50	<0.001	0.27	0.11 to 0.64	0.003	0.22	0.15 to 0.34	<0.001	0.18	0.09 to 0.36	<0.001
Deprivation quintile	Least 1	Ref		0.024	Ref		0.012	Ref		0.275	Ref		0.157	Ref		0.031	Ref		0.010
	2	0.82	0.43 to 1.53	0.527	0.59	0.36 to 0.96	0.034	0.87	0.47 to 1.61	0.664	1.04	0.56 to 1.94	0.903	1.01	0.75 to 1.34	0.970	1.30	0.73 to 2.34	0.375
	3	1.14	0.63 to 2.06	0.666	0.84	0.54 to 1.32	0.451	1.36	0.77 to 2.39	0.285	1.28	0.71 to 2.34	0.413	1.07	0.80 to 1.43	0.631	1.76	1.01 to 3.05	0.046
	4	1.87	1.08 to 3.24	0.026	1.19	0.77 to 1.82	0.433	1.54	0.88 to 2.71	0.131	1.56	0.86 to 2.81	0.143	0.89	0.66 to 1.21	0.468	2.09	1.21 to 3.61	0.008
	Most 5	1.64	0.91 to 2.95	0.100	1.33	0.85 to 2.07	0.209	1.36	0.75 to 2.46	0.318	1.91	1.06 to 3.45	0.033	0.64	0.45 to 0.89	0.009	2.44	1.41 to 4.23	0.001
Sex*	Male	1.42	0.99 to 2.04	0.059	1.35	1.01 to 1.82	0.044	1.89	1.32 to 2.72	0.001	2.29	1.59 to 3.31	<0.001	1.28	1.05 to 1.57	0.016	1.37	0.99 to 1.91	0.060
Ethnicity*	Other	1.44	0.42 to 4.93	0.559	1.00	0.34 to 2.94	1.000	0.70	0.16 to 3.00	0.626	0.63	0.14 to 2.73	0.534	0.70	0.30 to 1.61	0.395	1.78	0.69 to 4.56	0.230
Had COVID-19*†	Yes	1.18	0.77 to 1.81	0.443	1.37	0.99 to 1.91	0.058	1.07	0.71 to 1.63	0.745	1.04	0.66 to 1.55	0.949	1.22	0.96 to 1.54	0.104	1.33	0.93 to 1.90	0.121
Chronic disease*‡	Yes	1.07	0.73 to 1.58	0.722	0.99	0.72 to 1.36	0.958	0.92	0.61 to 1.38	0.691	0.55	0.35 to 0.87	0.010	0.68	0.54 to 0.85	0.001	0.94	0.65 to 1.35	0.731

*Reference categories for sex, ethnicity, had COVID-19 and chronic disease were female, white, not had COVID-19 and not had a chronic disease respectively. Full wording of all questions is provided in online supplemental appendix table A1.

†Having had COVID-19 was self-reported, see the Methods section.

‡Chronic diseases included cancer, type II diabetes, heart disease and respiratory diseases, see the Methods section for details.

AOR, adjusted odds ratio; NHS, National Health Service.

Unfairly restricted a lot by government

Just under 1 in 10 people reported feeling unfairly restricted (table 1). This rose with ACE count, with the proportion among those with four or more ACEs being more than twice as high as in those with none (table 2). Younger individuals were also more likely to report feeling unfairly restricted, along with those who were resident in more deprived quintiles and those who reported having had COVID-19 (table 2). When using logistic regression, independent relationships between feeling unfairly restricted and increasing ACE count remained, although differences between the no ACE and one ACE categories failed to reach significance. Logistic regression showed younger age and being male were also significantly related to feeling unfairly restricted (table 3).

Social distancing should end

Supporting the removal of social distancing increased more than threefold from those with no ACEs to those with four or more (table 2). Ending social distancing was also significantly more supported by those who were younger and male. Ethnicity, deprivation, or having had COVID-19 or a chronic disease were not significantly associated with support for ending social distancing. Using logistic regression, having more ACEs was still significantly associated with favouring ending social distancing but only having four or more ACEs remained significantly different from no ACEs (table 3). Those aged 60 years or over were significantly less likely to support ending social distancing (compared with those aged 18–29 years) with males also substantially more likely than females to support social distancing ending (table 3).

Mandatory face coverings should go

Support for ending mandatory face coverings increased fourfold between those with no ACEs and those with four or more ACEs (table 2). Younger individuals, those resident in more deprived quintiles, males and those who had not had a chronic disease were more likely to support mandatory face coverings ending. In logistic regression, ACE counts continued to show a positive relationship with support for ending mandatory face coverings. This was significant even with a single ACE compared with those with no ACEs (table 3). Younger ages, being male and not having had a chronic disease remained significantly associated with ending mandatory face covering measures. However, differences by age were only significant between the 18–29 years and 60+ years groups (table 3).

Break restrictions at least occasionally

Overall, around one in four respondents broke COVID-19 restrictions at least occasionally (table 1). In bivariate analyses, proportions having broken restrictions increased with ACE count and decreased with age but were not significantly related to deprivation, sex or ethnicity. Those who reported having already had COVID-19 and those without a history of chronic disease were more likely to have broken restrictions (table 2). When controlling for relationships between variables, breaking restrictions remained strongly related to ACE count with the likelihood of such behaviours being twice as high in those with four or more ACEs compared with those with none (table 3). Breaking restrictions also remained significantly associated with younger ages and not having suffered from a chronic disease, with deprivation also marginally significant (with less restriction breaking in the most deprived quintile; table 3).

Vaccine hesitancy

Around 1 in 13 individuals surveyed reported vaccine hesitancy (table 1). However, this increased around fourfold between those with no ACEs and those with four or more (table 2). Younger age groups were also more likely to report vaccine hesitancy along with those living in more deprived quintiles, those of other than white ethnicity and those who had already had COVID-19 (table 2). Applying logistic regression, having more ACEs remained significantly associated with vaccine hesitancy, although the difference between the no ACEs and one ACE category was not significant. Younger age remained strongly related to vaccine hesitancy along with being resident in more deprived quintiles. Ethnicity was not significantly related to vaccine hesitancy once ACEs, age and deprivation had been accounted for (table 3).

For breaking restrictions and vaccine hesitancy we also generated estimated levels (EMMs) in order to provide absolute measures of prevalence of breaking restrictions and vaccine hesitancy by ACE and age categories (figures 1 and 2). For having broken restrictions at least occasionally, estimated levels ranged from 10.67% (95% CIs 6.72% to 16.53%) in those aged 70+ years with no ACEs to 51.95% (95% CIs 38.34% to 65.27%) in those aged 18–29 years with four or more ACEs (figure 1). Similarly for vaccine hesitancy, levels ranged from 3.42% (95% CIs 1.66% to 6.93%, no ACEs, aged 70+ years) to 38.06% (95% CIs 24.08% to 54.35%, 4+ ACEs, aged 18–29 years; figure 2). Within any single age group, ACE count contributed to a steep increase in predicted breaking of restrictions and vaccine hesitancy. For instance, for vaccine hesitancy, in those aged 30–39 years, there was a rise from 13.95% (95% CIs 7.83% to 23.62%) in those with no ACEs to 33.48% (95% CIs 20.75% to 49.18%) in those with four or more (figure 2). CIs for all data points are provided in online supplemental appendix table A3.

Figure 1

Adjusted mean percentage of individuals having broken COVID-19 restrictions at least occasionally, by age and adverse childhood experience (ACE) count.

Figure 2

Adjusted mean percentage of individuals with vaccine hesitancy, by age and adverse childhood experience (ACE) count.

Discussion

Voluntary compliance with public health advice has played a central role in reducing the viral transmission of COVID-19. In this study, approximately a quarter of participants admitted to at least occasionally breaking the rules (table 1) while a minority supported immediately ending social distancing and face coverings (5.91% and 5.82% respectively; table 1); regulations in place at the time of this study.19 Critically, 7.75% of individuals would not immediately agree to a COVID-19 vaccination. Vaccine hesitancy, as well as breaking or ending current restrictions, were related to sociodemographics with younger age groups in particular reporting more restriction breaking and higher vaccine hesitancy (table 2, figures 1 and 2). As reported elsewhere, males were also more likely to break restrictions and favour an end to those in place (table 3).3 23 Lower trust in NHS COVID-19 information and feeling unfairly restricted by government were also related to vaccine hesitancy and restriction breaking (table 1). However, while interrelations between trust in public bodies and compliance with guidance has been studied elsewhere,2 24 far less attention has paid to the life-course factors that may contribute to lower trust in health and state systems and potential rejection of related regulations and medical interventions.

Critically, most individuals surveyed, including those with ACEs, supported and followed COVID-19 restrictions (table 2, figure 1). However, results identify individuals with a history of childhood adversity having less trust in NHS COVID-19 information and being more likely to favour removal of control measures (tables 2 and 3). Lower trust in NHS COVID-19 information tripled between those with no ACEs and those with four or more and feeling unfairly restricted by government more than doubled (table 2). Such increases are consistent with other findings here that individuals with four or more ACEs were two times more likely to break restrictions at least occasionally compared with those with no ACEs when controlling for relationships with sociodemographic factors and history of COVID-19 infection or chronic disease (table 3). Studies elsewhere suggest individuals with ACEs are more likely to have developmental and behavioural factors that increase the risk of ill health across the life-course;25 potentially leaving them more susceptible to infection and ill health from COVID-19 (eg, through smoking, cancer).7 12 26 Consequently, understanding why individuals with ACEs may be more likely to reject virus control measures is vital to protecting their health.

A number of outcomes previously associated with exposure to ACEs may contribute to links between greater ACE exposure and lower compliance with and support for COVID-19 interventions. Higher ACEs are associated with lower acceptance of delayed gratification with greater preference for short-term returns at the expense of potentially greater return in the longer term.27 28 ACEs have also been associated with lower prosocial behaviours and sense of belonging,17 29 although such effects are not well studied in adults, nor whether they affect consideration of how personal behaviour may impact the well-being of others in local communities. However, a history of ACEs is known to be associated with other anti-social behaviours, including violence.7 Higher exposure to ACEs is also associated with poorer mental well-being and alcohol and drug use18 with the latter especially having known associations with the adoption of wider risk-related behaviours.30 31 Finally, ACEs have been associated with having lower trust both in other individuals and public services,16 17 32 a finding consistent with results here whereby lower trust in COVID-19 information from the NHS increased from 4.05% with no ACE to 12.34% in those with four or more (table 2).

With vaccination at the centre of COVID-19 control strategies going forward, higher levels of vaccine hesitancy in those with more ACEs is an important consideration. In those aged 18–29 years, modelled vaccine hesitancy more than doubled from an estimated 16.52% in those with no ACEs to 38.06% in those with four or more (figure 1, online supplemental appendix table A3). In this study 48.14% of individuals had at least one ACE and 10.28% had four or more. Such figures are consistent with those reported in other studies (eg, England,33 USA,9 New Zealand34) suggesting that ACEs are a feature of the life-course of a substantive proportion of the population. Consequently, unaddressed high levels of vaccine hesitancy in this group represent a significant risk to the health of those with a history of ACEs and potentially also to those in their local communities. Our results suggest that ACE-informed and trauma-informed approaches may be an important consideration when considering compliance with infection control restrictions and in improving uptake of medical interventions such as COVID-19 vaccination. Although little work has been undertaken specific to COVID-19, increased compliance from those with ACEs may benefit from a greater emphasis on safety and trustworthiness. Thus, strategies may consider use of alternative spaces and settings, avoiding ones which may potentially be associated with previous negative experiences for some individuals (eg, healthcare). They may also require different channels for information provision to account for lower trust in public services. Moreover, awareness and training for those contacting individuals, potentially with a history of trauma, may allow them to support those still wavering, for instance, with vaccine compliance.35

Consistent with exposure to ACEs not being deterministic of outcomes such as trust or behaviour, most individuals with ACEs followed restrictions and supported vaccination. Risks of negative outcomes in those exposed to ACEs are reduced through, for instance, exposure to sources of resilience.17 36 37 Thus, access to a supportive adult, connectedness with local communities and support managing behaviour and emotions in childhood are all related to reducing risks of poor outcomes from ACEs across the life-course.38–40 During the pandemic, available sources of resilience for children may have fallen and exposure to ACEs risen in some communities;41 42 harming children and potentially increasing future risks of poor life-course outcomes and rejection of virus control restrictions. Policies and interventions that prevent ACEs and build resilience are increasingly well evidenced and include better parenting support, legislation to protect children in the home and policies to reduce issues such alcohol misuse.43 44 While such interventions may not immediately impact adult views and support for pandemic restrictions, they may encourage trust and support for public services in children and in the longer term increase community resilience to transmission of future infections.

There were a number of important limitations with this study. Compliance was 36.4% of those answering the telephone. Although this is similar to many phone surveys, including during COVID-19,45 46 we do not have any measures of whether responses would have differed in those refusing to participate or not answering calls. The survey used self-reported measures of ACEs and COVID-19-related behaviour. Individuals may have either exaggerated, forgotten or chosen not to disclose childhood adversities or compliance with COVID-19 restrictions. However, levels of ACEs reported were comparable to those previously collected in the UK including through face-to-face interviews.33 While the survey included over 2000 individuals, women were overrepresented in the final sample. However, sufficient data were available to include sex in all data models in order to identify differences between sexes and to control for sex-related differences when examining relationships between outcomes of interest and other independent variables. The sample did not provide adequate numbers for detailed analyses by individual ethnicity types, limiting analyses to just binary white and other categories. However, even with a low sample size and all black, Asian and other minority groups combined into a single category, odds of vaccine hesitancy, for instance, were substantially higher than in the white population (1.78, 95% CIs 0.69 to 4.56); although this failed to reach statistical significance (table 3). While the low level of ethnic minority participants reflects Wales having only 5.6% of adults from black, Asian and other minority ethnic groups,47 this could be rectified in further studies with oversampling in such communities and may result in the identification of other important differences between ethnicities. Analysis employed a categorical approach to variables including ACE count and age. While this allowed non-ordinal comparisons between categories, potential differences between individuals within categories may have been obscured. Finally, while the survey was conducted during a period of national lockdown, individuals’ responses may have been affected by the timing of their interview (eg, near the start or end of the lockdown period). However, individuals from all different sociodemographic groups were sampled throughout the entire data collection period.

Conclusions

There is an immediate and ongoing need to understand how best to maximise uptake of COVID-19 vaccinations and compliance with public health restrictions aimed at reducing the spread of COVID-19 or any other infectious agents that may provide a threat to public health. Coping with trauma resulting from at least one current or previous ACE is common in the populations of many countries with proportions having experienced multiple ACEs frequently reaching ten percent or more of the population.7 Such individuals are already known to have greater health risks across the life-course. Results here, suggest such individuals may have more difficulty with compliance with public health control measures and consequently require additional support. A better understanding of how to increase their trust in health systems and compliance with health guidance is urgently required. Without consideration of how best to engage such individuals, some risk being effectively excluded from population health interventions, remaining at higher risks of infection and posing a potential transmission risk to others. Increasing the appeal of public health information and interventions, such as vaccination, to those who have experienced ACEs should be considered in health protection responses. Longer term, however, achieving better compliance with pandemic and other public health advice is another reason to invest in safe and secure childhoods for all children which are free from ACEs and rich in sources of resilience. Such measures appear likely not only to reduce health-harming behaviours and ill health across the life-course but may also reduce the spread of COVID-19 or other infectious threats to public health that may materialise in subsequent decades.