Adverse childhood experiences, morbidity, mortality and resilience in socially excluded populations: protocol for a systematic review and meta-analysis

Abstract

Introduction

Socially excluded populations, defined by homelessness, substance use disorder, sex work or criminal justice system contact, experience profound health inequity compared with the general population. Cumulative exposure to adverse childhood experiences (ACEs), including neglect, abuse and household dysfunction before age 18, has been found to be independently associated with both an increased risk of social exclusion and adverse health and mortality outcomes in adulthood.

Despite this, the impact of ACEs on health and mortality within socially excluded populations is poorly understood.

Methods and analysis

We will search MEDLINE, Cumulative Index of Nursing and Allied Health Literature, Educational Resources Information Center, PsycINFO, Applied Social Science Index and Abstracts and Criminal Justice Database for peer-reviewed studies measuring ACEs and their impact on health and mortality in socially excluded populations.

Three review questions will guide our data extraction and analysis. First, what is the prevalence of ACEs among people experiencing social exclusion in included studies? Second, what is the relationship between ACEs and health and mortality outcomes among people experiencing social exclusion? Does resilience modify the strength of association between ACEs and health outcomes among people experiencing social exclusion?

We will meta-analyse the relationship between ACE exposure and health outcomes classified into six a prior categories: (1) substance use disorders; (2) sexual and reproductive health; (3) communicable diseases; (4) mental illness; (5) non-communicable diseases and (6) violence victimisation, perpetration and injury. If there are insufficient studies for meta-analysis, we will conduct a narrative synthesis. Study quality will be assessed using the MethodologicAl STandards for Epidemiological Research scale.

Ethics and dissemination

Our findings will be disseminated in a peer-reviewed journal, in presentations at academic conferences and in a brief report for policy makers and service providers. We do not require ethics approval as this review will use data that have been previously published.

PROSPERO registration number

CRD42022357565.

Strengths and limitations of this study.

• This will be the first systematic review to examine the impact of adverse childhood experiences (ACEs) on health and mortality outcomes within populations experiencing one or more forms of social exclusion: homelessness, substance use disorder, sex work and criminal justice system contact.

• Our review builds on a prominent 2017 systematic review that examined this association in general the general population but excluded samples of socially excluded populations.

• This review will consider the potential moderating or mediating role of resilience on the association between ACE exposure and health, which has not been a strong focus of previous research on this topic.

• Limitations include that we will only consider cumulative exposure to ACEs, rather than effects of specific types of ACEs on health, to be consistent with the 2017 review.

• We are unable to examine the effects of age at ACEs exposure on health outcomes.

Introduction

Social exclusion results from multiple forms of disadvantage—such as poverty, marginalisation and discrimination—that exclude people from fully participating in mainstream society.¹ Experiences of homelessness, substance use disorder, sex work and criminal justice system contact are associated with extreme social exclusion. Experiencing these forms of social exclusion is associated with substantial difficulties accessing health and social services, difficulties obtaining employment or education, discrimination and stigmatisation.^{2 3} Profound health inequities, including high rates of communicable and non-communicable disease, multimorbidity and premature death are also associated with these forms of social exclusion.^{4 5} For example, a systematic review and meta-analysis of mortality in people experiencing social exclusion identified a standardised mortality rate 7.9 times higher in males and 11.9 times higher in females than in the general population.⁴

Adverse childhood experiences (ACEs), including neglect, abuse and household dysfunction, are also associated with poor health and mortality outcomes.^{6 7} Greater cumulative exposure to ACEs before age 18 has been associated with an increased risk of mental illness, substance abuse, obesity, diabetes and cancer in adulthood.^{8 9} Exposure to a higher number of ACEs is also associated with an increased risk of experiencing social exclusion, including a higher risk of repeated incarceration,¹⁰ engaging in sex work,¹¹ homelessness¹² and substance dependence.¹³

A 2017 meta-analysis of studies conducted in predominantly high-income countries found that 57% of individuals in general population samples reported exposure to at least one ACE and 13% report having experienced four or more ACEs.⁸ However, this review explicitly excluded studies of individuals with experiences of incarceration, homelessness, substance use disorder and other ‘high risk’ socially excluded populations. The burden of ACEs is likely much higher among individuals experiencing social exclusion. For example, a systematic review reported that 90% of individuals who were homeless experienced one or more ACEs and 54% reported four or more ACEs.¹⁴ Another review reported that, among individuals in treatment for substance dependence, 85%–100% had experienced at least one ACE.¹⁵

Within socially excluded groups, ACEs have been associated with heightened risk of health depleting behaviours and disease. One systematic review found that, among studies of individuals experiencing homelessness, exposure to a higher number of ACEs was associated with a higher risk of suicide attempt, mental illness and blood-borne disease.¹⁴ Similarly, another review found that ACEs were associated with earlier onset and higher severity of substance use among individuals with substance dependence.¹⁵

Although much of the ACE literature is deficit-focused, research has shown that positive individual, familial, neighbourhood and structural resilience factors can positively modify the relationship between ACEs and adverse outcomes among young people.¹⁶ Resilience is defined as ‘the process of negotiating, managing and adapting to significant sources of stress or trauma’.¹⁷ It includes both internal (eg, genetics, personality and belief systems) and external (eg, family, community support and environmental) factors that enable individuals to adapt through periods of adversity.¹⁶ In general population samples, internal resilience factors in childhood and adulthood have been found to attenuate negative health effects associated with ACEs.^{18 19} However, there is little evidence on the effect of resilience factors on the association between ACEs and health in socially excluded populations.

Understanding the effects of ACEs on health in socially excluded groups will inform the understanding of how adversity affects health across the lifespan. This includes investigating resilience factors across the lifespan that may attenuate the health depleting effects of ACEs. It also serves to acknowledge the broader social contexts in which ACEs occur, which are important to guide targeted prevention and intervention.²⁰

This review builds on a recent systematic review and meta-analysis of the relationship between cumulative ACE exposure and health outcomes that explicitly excluded studies of socially excluded populations.⁸ It will therefore fill an important evidence gap by synthesising the peer-reviewed evidence on the association between ACEs and health and mortality outcomes in populations experiencing social exclusion. The primary objective of this review is to synthesise and evaluate the peer-reviewed literature to:

1. Calculate pooled prevalence estimates of exposure to any ACEs, multiple ACEs and specific ACEs, within included studies.

2. Quantify the association between cumulative ACE exposure and health outcomes among people experiencing social exclusion.

3. Qualitatively summarise the evidence on the moderating or mediating associations of resilience on the relationship between cumulative ACE exposure and health outcomes among people experiencing social exclusion.

Methods and analysis

We will conduct a systematic review and, where data from the included studies permit, meta-analyses to meet our stated aims. This protocol is reported in accordance with guidelines from the 2015 Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P; see online supplemental appendix).²¹ This review commenced on 2 September 2022 and we anticipate that the review will be completed by 31 August 2024.

Eligibility criteria

Inclusion and exclusion criteria are described in table 1.

Table 1.

Summary of inclusion and exclusion criteria

Component	Inclusion criteria	Exclusion criteria
Participants	Socially excluded populations defined by homelessness, substance use disorder, sex work or criminal justice system contact4	General population samples and selected samples other than those specified
Exposure	Cumulative exposure to a discrete number of ACE	–
Comparator	Studies that compare individuals with a higher number of ACE exposures to those with fewer or no ACE exposures	–
Outcome	Health and mortality outcomes categorised as: Substance use, substance use disorders and substance-related harm. Sexual and reproductive health. Blood-borne and other communicable diseases (excluding STIs). Mental health. Non-communicable diseases. Violence victimisation, perpetration and injury.	Studies that do not measure any health or mortality outcome of interest; studies in which the ACE exposure does not clearly precede the outcome of interest; studies that only measure prognostic biomarkers or psychological states without a diagnostic threshold
Study design	Case–control, cohort and cross-sectional studies	Case reports, case series, qualitative studies, dissertations or any other studies not reporting individual-level outcomes or quantitative data; studies that are not peer-reviewed

ACEs, adverse childhood experiences; STIs, sexually transmitted infections.

Participants

We will include studies whose participants have experienced social exclusion, and where ACE exposure preceded the health and mortality outcomes measured (table 1). We will define social exclusion in accordance with the Lancet inclusion health series^{4 22} to include individuals experiencing homelessness, substance use disorder, sex work and/or criminal justice system contact.^{4 23} Unlike the Lancet inclusion health series, we will also include individuals with alcohol and cannabis use disorders. While cannabis and alcohol use are increasingly legal or decriminalised worldwide, problematic use of these drugs is common and associated with both social exclusion^{24 25} and poor health outcomes.^26–28 Accordingly, these substances have been a focus of public health programmes focussing on inclusion health.²⁹

Exposure measures

Our exposure of interest is cumulative exposure to ACEs. ACEs are typically operationalised in research to include experiences of physical, sexual, emotional or psychological abuse perpetrated by a parent or adult caregiver; neglect (eg, lack of appropriate supervision, abandonment and emotional or material deprivation) and household dysfunction (eg, parent or caregiver mental illness, substance dependence or incarceration; witnessing domestic violence) occurring before age 18 (table 2).³⁰ More recent studies and measurement tools encompass additional ACEs including: peer victimisation or bullying; family homelessness; parent or caregiver separation, divorce or death; separation from family (eg, out-of-home care); serious childhood illness or injury and witnessing or being a victim of neighbourhood violence.

Table 2.

Examples and characteristics of common ACE measurement tools

Measurement tool	Year developed	ACE score	ACEs measured
Original Adverse Childhood Experiences Questionnaire (ACE-Q)6	1998	0–7	Psychological/emotional, physical and sexual abuse; household substance abuse, mental illness, incarceration; violence towards mother
Expanded Original ACE Questionnaire (ACE-EQ)49	2004	0–10	7-item ACE-Q plus physical and emotional neglect, lost biological parent to divorce or abandonment
The Adverse Childhood Experiences International Questionnaire (ACE-IQ)50	2009	0–13	10-item ACE-EQ plus community violence, collective violence, bullying
US Centers for Disease Control ACE Tool51	2010	0–11	7-item ACE-Q plus multiple questions for abuse ACEs considered experience of multiple ACEs, excludes neglect ACEs
Behavioural Risk Factor Surveillance Survey ACE Module52	2014	0–9	7-item ACE-Q, differentiating between household alcohol abuse and substance abuse, plus parental separation/divorce
Suggested revised Adverse Childhood Experiences Questionnaire31	2015	0–14	10-item ACE-EQ plus peer victimisation, peer isolation/rejection, community violence exposure, low socioeconomic status
Childhood Experiences Survey (CES) Expanded53	2017	0–17	10-item ACE-EQ plus financial instability, food insecurity, homelessness, parental absence, parent/sibling death, bullying, violent crime
ACE Abuse Short-Form (ACE-ASF)54	2017	0–3	Emotional, physical and sexual abuse only
The Paediatric ACEs and Related Life-events Screener (PEARLS)55 56	2022	0–10	Same as 10-item ACE-EQ
Childhood Trauma Questionnaire—Short Form57	2003	0–5	25-item scale of severity of exposure to physical and emotional abuse, emotional and physical neglect and sexual abuse. Will only be included if binary classification of exposures are used

ACEs, adverse childhood experiences.

ACE measurement tools are administered to participants aged 18 or younger, or to adults who are asked about their experiences of ACEs before age 18.^{8 31} Typically, respondents are asked to sum the total number of categories of ACEs experienced to generate an overall ‘ACE score’ without specifying which ACEs were experienced (‘deidentified’ ACE score). A higher ACE score is considered to represent a higher degree of childhood adversity. Some ACE measurement tools allow the respondent to report the individual ACEs experienced in addition to the ACE score (‘identified’ ACE score).

Due to the heterogeneity in ACE scales used and the specific ACEs measured, we will include studies that use a measure of ACEs that aligns conceptually with the original ACE questionnaire developed by Felitti and colleagues,⁶ which is considered the genesis of contemporary ACE research.^{6 32}

Studies that measure the association between individual ACEs (such as experiencing sexual abuse) and health or mortality outcomes, but do not include a cumulative measure of exposure to multiple ACEs (eg, ACE score), will not be included in the review. Our focus on cumulative exposure to ACEs, rather than exposure to single ACEs, is to (a) permit assessment of a dose–response relationship in ACE exposure and (b) align with previous ACE research.^{8 32}

Outcome measures

Primary outcomes

The primary outcomes of this study are morbidity and mortality from disease and/or injury. We will categorise health and mortality outcomes into six a priori categories to align with a highly cited review of ACEs and health in the general population.⁸ We modified the classifications used in that review to reflect what is known about the health needs and experiences of socially excluded populations:⁴ physical health status was divided into two distinct categories for communicable and non-communicable diseases, we removed a ‘weight and physical exercise’ category and recategorised obesity to the non-communicable disease category, and we included injury-related morbidity and mortality alongside outcomes relating to violence victimisation and perpetration.

We will examine health outcomes in the following six a priori categories:

1. Substance use, substance use disorders, and substance-related harm.

2. Sexual and reproductive health (eg, adolescent pregnancy, early sexual initiation, number of sexual partners and sexually transmitted infections (STIs)).

3. Blood-borne and other communicable diseases (excluding STIs).

4. Mental health (including mental health disorders, self-harm and suicidality).

5. Non-communicable diseases.

6. violence victimisation, perpetration and injury.

Data on outcomes of interest will be extracted and, where possible, grouped using International Classification of Diseases, Tenth Revision (ICD-10) codes³³ provided in the online supplemental appendix.

We will include studies that measure relevant outcomes using physician diagnosis, validated survey measures with clinical cut-offs, diagnostic biomarkers, health and mortality records or self-report. We will not include studies that only measure psychological states without a diagnostic threshold (eg, depressive symptoms or psychological distress), or prognostic biomarkers (eg, serum levels of C-reactive protein).³⁴

Secondary outcomes

Within included studies, we will summarise and evaluate evidence on the potential for resilience factors to modify the association between ACE exposure and its association with health outcomes and mortality in individuals who have experienced social exclusion. Resilience is considered to be an important moderator or mediator in development of disease after exposure to ACEs.^35–39 We will define resilience as ‘the process of negotiating, managing and adapting to significant sources of stress or trauma’.¹⁷ Several validated measures of resilience are used that contain considerable heterogeneity in their operationalisation of resilience.^{40 41} To compensate for this limitation in the study of resilience, we will define resilience to include any factors measured that may theoretically support people exposed to ACEs. This includes measurement of resilience using validated surveys. This inclusive approach to defining resilience will allow the broader inclusion of resilience and other protective factors that may be amenable to policy or practice intervention.

Within included studies, we will summarise and evaluate evidence on the potential for resilience factors to moderate and/or mediate the association between ACE exposure and its association with health outcomes and mortality in individuals who have experienced social exclusion.

Our database search will not include terms specific to resilience. We will only extract data on resilience and its moderating or mediating effects on the association between ACEs and health in studies that meet our other inclusion criteria.

Study design

Our review will include peer-reviewed observational studies that report quantitative data for at least one of our primary outcomes of interest (ie, mortality and health outcomes within at least one of the six categories of interest). We will include prospective and retrospective cohort studies, cross-sectional studies if we can determine temporality of ACE exposure and measurement of the outcome of interest, case–control studies and secondary analyses of randomised control trials. We will not include studies that analyse only aggregate data (eg, ecological studies and reviews), studies that are not peer-reviewed (eg, grey literature and conference abstracts), dissertations, editorials or opinion pieces, studies that only report qualitative data or case reports and case series.

Search strategy and data management

Search strategy

We will search MEDLINE, Cumulative Index of Nursing and Allied Health Literature (CINAHL), Educational Resources Information Center (ERIC), PsycINFO, Criminal Justice Abstracts and Applied Social Science Index and Abstracts (ASSIA). Our search strategy has been developed in collaboration with a research librarian and with advice from the authors of a previous systematic review examining the impact of ACEs on health outcomes in the general population.⁸ The search strategy formatted for MEDLINE is included in the online supplemental appendix. Reference lists of all included studies will be checked for relevant studies not identified in the database search.

The search will be restricted to studies published after May 1998 to coincide with the seminal ACE study.⁶

Study selection

Studies identified in the database searches will be imported into EndNote V.20⁴² and duplicates will be removed based on two iterations: (1) same title, author and year of publication, and (2) same title, author and journal. The remaining citations will be uploaded into Covidence software⁴³ and a second check for duplicates will be done prior to screening. Titles and abstracts will be independently screened by two reviewers based on eligibility criteria (table 1). Reviewers will screen 10% of titles and abstracts before evaluating if the criteria need to be modified to increase clarity or accuracy. Following this evaluation, all remaining studies will be independently screened by two reviewers. We will calculate Cohen’s alpha to assess inter-rater agreement. Full-text articles of relevant studies will be independently screened by two reviewers based on eligibility criteria. Any disagreements during title and abstract or full-text screening will be resolved through discussion with a third reviewer.

Studies will be assessed for inclusion or exclusion using the criteria in table 1.

Data extraction

We will extract all data on cumulative exposure to ACEs, including the total number of ACEs measured, mean number of ACE exposures and exposure to different levels of ACEs (eg, proportion of study population exposed to 0, 1, 2, 3, …, n ACEs).

When identified (ie, item level) ACE data are available, we will extract data within four subgroups (table 3).

Table 3.

Four subgroups of ACEs

ACE subgroup	Examples of included ACEs
Abuse	Physical, sexual, emotional and psychological abuse by a parent or adult caregiver
Neglect	Abandonment, lack of adequate supervision, emotional or material deprivation
Household dysfunction	Parental or caregiver mental illness, substance use disorder, incarceration and domestic violence
Other	Peer victimisation or bullying; family homelessness; parental or caregiver separation, divorce or death; separation from family (eg, out-of-home care); serious childhood illness or injury and witnessing or being a victim of neighbourhood violence

ACEs, adverse childhood experiences.

Data will be extracted using an Excel sheet developed by the research team. The extraction sheet will include basic study information (author(s), publication year, journal, country, year(s) of recruitment, DOI and study design), information on the sample (age, sex and type of social exclusion), exposure to ACEs (ACE measurement tool used, identified ACEs (yes/no), prevalence of distribution of ACE exposure mean ACE score), health and mortality outcomes of interest (outcomes measured and how outcomes were measured) and measures of association used to report on the relationship between ACEs and health and mortality outcomes. If a study provides crude and adjusted measures of association, we will extract both measures in accordance with Cochrane guidelines.⁴⁴

Data will be extracted by one reviewer and checked by a second. Any discrepancies will be resolved through consensus with a third reviewer. Data from graphs or figures that are not listed in tables or in-text will be extracted using WebPlotDigitizer.⁴⁵

Risk of bias assessment

Bias in each study will be assessed using the MethodologicAl STandards for Epidemiological Research (MASTER) scale.⁴⁶ The MASTER scale is a 36-item scale that assesses the quality of studies by the presence or absence of 36 methodological safeguards against bias. A relative risk of bias score is calculated as a proportion of bias safeguards met relative to the study with the highest MASTER scale score included in the review. We will report absolute and relative bias scores. A higher relative risk of bias score represents a greater number of bias safeguards reported by a study and a lower probability of bias.⁴⁶ Bias for each study will be assessed independently by two reviewers. Any conflicts in evaluation of bias will be resolved through consensus with a third reviewer.

We will neither include nor exclude studies based on quality score but will conduct sensitivity analyses to evaluate the effect of study quality on overall findings of our review.

Statistical analysis

We will present summary statistics describing the characteristics of included studies, including the years of data collection, population(s) studied; country; study design; sample size; mean age in years; sex (% female participants); ACE measurement tools used; number of ACEs measures; prevalence of different levels of cumulative ACE exposure; health and mortality outcomes examined; measures of association; and quality assessment results.

Three review questions will guide our data analysisfigure 1. First, what is the prevalence of ACEs among people experiencing social exclusion in included studies? Second, what is the relationship between ACEs and health and mortality outcomes among people experiencing social exclusion? And third, what is the role of resilience in moderating or mediating the relationship between ACEs and health and mortality outcomes?

Figure 1.

Analytic framework. This figure depicts the analytic framework that will guide our review. Numbers indicate the three research questions we will address. Arrows indicate the presumed links between the population, exposure and outcomes.

If there are sufficient studies, we will conduct a meta-analysis to calculate a pooled prevalence of ACE exposure by total number of ACEs and prevalence of different levels of cumulative ACE exposure.^{6 8 14} If there are sufficient data, we will conduct separate meta-analyses to calculate a pooled OR for each of the six a priori health and mortality outcome categories, comparing individuals with varying levels of exposure to ACEs. In line with previous reviews of ACEs and health in the general population,⁸ we will compare between individuals who have been exposed to four or more ACEs to those experiencing no ACEs.

We will use an I² test to assess heterogeneity in effect size. If there is an insufficient number of studies for any outcome category, we will conduct narrative synthesis according to the Synthesis Without Meta-analysis (SWiM) guidelines.⁴⁷ We will use Egger’s test to assess the potential for publication bias in our pooled estimates⁴⁸ and leave-one-out sensitivity analyses to test for small study effects.

If there are sufficient data, we will conduct a meta-regression to explore potential sources of heterogeneity across studies and their impact on pooled estimates.⁴⁸

We will not assess the relationship between ACEs and health and mortality outcomes within specific socially excluded populations. As previous studies have found considerable overlap across categories of social exclusion, there are likely to be individuals who belong to multiple populations that experience different forms of social exclusion.^{4 5} As a result, differences between studies of different socially populations would likely reflect differences in recruitment methods rather than population differences, which could bias our analyses.

If there are sufficient studies providing empirical evidence on the role of resilience, we will conduct appropriate meta-analyses. Otherwise, quantitative evidence on resilience will be summarised using narrative methods.⁴⁷

Patient and public involvement

Our review will involve no new data collected from any participants nor any participant recruitment. This work is a collaboration between academics and a community-based advocacy and service delivery organisation that supports young people and their families to access effective treatment for substance use and mental healthcare and become meaningfully engaged in decision making that impacts their lives.

Ethics and dissemination

This study does not require ethics approval because it will incorporate and synthesise data that have already been published. Our findings will be disseminated through a peer-reviewed journal article and will be presented at national and international conferences. We will develop a plain language summary for dissemination to relevant health and social service providers including the study’s primary funder, Youth Support and Advocacy Service (YSAS).

Ethics statements

Patient consent for publication

Not required.