Impact of parental adverse childhood experiences on children’s healthcare utilisation: a systematic review

Abstract

Background

Adverse childhood experiences (ACEs) are linked to poor health and social outcomes, with growing interest in their intergenerational effects. Socioeconomic context strongly shapes both exposure to ACEs and their transmission across generations, yet is often under-represented in research. While many studies have explored how ACEs affect adult health, the impact of parental ACEs on children’s healthcare utilisation has not yet been systematically reviewed.

Methods

We systematically reviewed studies examining associations between parental ACEs or related adversity and children’s use of preventive, primary or secondary healthcare. We searched PubMed, Embase and PsycINFO databases up to December 2024. Title and abstract screening were conducted using an AI-assisted screening tool called Active Learning for Systematic Reviews. Study quality was assessed across design, participant selection, measurement and analysis. Findings were synthesised narratively and visualised with harvest plots, grouped by healthcare level and exposure type.

Results

Out of 8494 records, 15 studies were included. Designs, populations, ACE measures and outcome definitions were heterogeneous. No consistent associations were found in preventive or primary care domains. In secondary care, 11 of 12 analyses found a positive or no association; four high-quality studies showed increased use of emergency, inpatient or psychiatric services among children of parents with ACEs. Methodological variability limited comparability between studies, particularly in ACE definitions.

Conclusions

Parental ACEs may contribute to increased use of secondary healthcare in offspring, though evidence across care levels remains inconclusive. Future studies should aim for more consistent measurement of ACEs and standardised outcome definitions to clarify intergenerational effects on healthcare utilisation.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Adverse childhood experiences (ACEs) can have adverse effects on both mental and physical health. Over the last decade, more has become known about their intergenerational effects. For example, developmental and health effects have been observed in children of parents exposed to ACEs. Also, previous research has demonstrated the effects of ACEs on healthcare utilisation in adults. However, evidence for the effects of parental ACEs on healthcare utilisation of their children has not been systematically reviewed.

WHAT THIS STUDY ADDS

• No association was found between parental ACEs and primary care utilisation, and mixed results were found for preventive care. However, a general trend suggests that parental ACEs increase secondary healthcare utilisation in children.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Higher secondary care utilisation among children of parents with ACEs implicates an increased burden of disease and added stress on healthcare systems. Addressing ACE-related healthcare challenges requires an integrated care approach that incorporates recognition and prevention of ACEs. In practice, this calls for early identification and coordinated family support. Future research should explore which care strategies best meet the needs of these families and how to effectively target upstream risk.

Introduction

Adverse childhood experiences (ACEs) are potentially traumatic events that take place during childhood (0–17 years). Examples of these ACEs are child abuse or witnessing domestic violence, household substance abuse, neglect, psychological problems and divorcing parents.¹ A history of ACEs has been linked to a wide variety of problems, including chronic and mental illnesses, social cognitive impairments and economic disadvantage.2,5 An accumulating body of literature provides growing evidence that ACEs can be transmitted across generations, with parents who have a history of ACEs potentially passing health effects on to their children.⁵

Socioeconomic conditions are key determinants of such intergenerational patterns. Parental socioeconomic position influences both the likelihood of ACE exposure and the available resources to buffer its effects. As recent reviews show, socioeconomic context shapes the prevalence and clustering of ACEs across generations.^{6 7} Even with statistical adjustment, residual confounding and effect modification by socioeconomic mechanisms are likely, given their complexity and multidimensional nature, which complicate the interpretation of associations between parental ACEs and child outcomes. Beyond these social determinants, several mechanisms are thought to play a role in this transmission, including impaired parenting resulting from the social-emotional effects of ACEs, epigenetic changes induced by early-life stress and physiological changes associated with increased allostatic load, further increasing the risk of subsequent illnesses.^{8 9}

The intergenerational transmission of ACEs is an important contributor to the overall burden of adversity on parents, children, healthcare and society. Parental mental health conditions related to ACEs exacerbate long-term negative health outcomes and may increase healthcare use. Moreover, certain healthcare interventions themselves can become traumatic events.¹⁰ Additionally, Shah et al describe that ACEs may lead to suboptimal coping abilities in parents that affect their capabilities to care for their child after hospitalisation.¹¹ This interplay of adversity, caregiving strain and systemic factors further increases the burden on healthcare and society, amplifying the direct effects of ACEs.

According to Hughes et al, ACEs are associated with major health issues and financial costs across European countries.³ However, the actual impact of parental ACEs on healthcare utilisation has never been systematically evaluated. Thus, more extensive research on the impact of parental ACEs and their healthcare consequences is needed for population health management purposes.¹² Hence, this review aims to systematically examine the association between parental ACEs and children’s healthcare utilisation. Understanding this relationship can make way for reducing ACE-related complications in future generations and benefit both communities and patients.

Methods

Data sources and search strategy

This review followed a preregistered protocol (PROSPERO ID: 422942, accessible online via https://www.crd.york.ac.uk/PROSPERO/view/CRD42023422942) and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹³ We searched PubMed, Embase and PsycINFO for literature on parental ACEs and healthcare utilisation published before December 2024, using terms related to ACEs, early life stress and healthcare use (see online supplemental appendix 1). Only English- and Dutch-language reports were included. Case reports, qualitative studies and reviews were excluded. Grey literature was hand-searched via the Data Archiving and Networked Services (DANS) Data Station Life Sciences and British Library E-Theses Online Service (EThOS).

Eligibility criteria

Peer-reviewed studies assessing the association between parental ACEs (experienced by mothers and/or fathers or primary caregivers in their childhood) and healthcare utilisation were eligible for inclusion. Studies had to meet two criteria: (1) the exposure involved parental ACEs and (2) the outcome described healthcare use or access by children. A broad definition of ACEs was used, allowing inclusion of studies using standardised tools (eg, ACE Index and Childhood Trauma Questionnaire), individual events (eg, physical violence, parental substance use) or other adversities (eg, migration). Healthcare utilisation included general practitioner visits, specialist care, routine preventive paediatric check-ups, dental care, treatment and medication.

Study selection

All initially identified records were imported into Active Learning for Systematic Reviews (ASReview) for abstract screening. ASReview combines machine learning with active learning to predict inclusion likelihood.¹⁴ Based on prior evaluation studies, we selected a text-analysis approach that helps the software identify records most likely to be relevant (Naïve Bayes classifier with Term Frequency–Inverse Document Frequency for feature extraction) and applied certainty-based sampling to prioritise relevant records. Screening was halted after 25% of records were reviewed, optimising efficiency while minimising irrelevant results. This was informed by prior validation studies that showed ASReview identified over 95% of relevant records while substantially reducing workload compared with full manual screening.^{14 15} Independent evaluations of ASReview also demonstrate that the approach is reliable and efficient for systematic review screening. Studies show that it can achieve recall and precision comparable to manual screening while reducing the number of abstracts needing review by 40%–60%.^{16 17}

Subsequently, remaining records were imported into Mendeley for full-text screening using a standardised form (Microsoft Forms) to assess eligibility and extract key information. Five reviewers (MFMK, JCK-dJ, MS, AZ and AJMB) independently labelled studies as ‘included’ or ‘excluded’ with reasons; discrepancies were resolved through discussion. All inclusions were verified by the first author. Screening of reference lists from selected studies and prior reviews yielded one additional paper, resulting in 15 included studies. No separate list of excluded studies was compiled. Figure 1 shows the PRISMA flowchart.¹³

Figure 1. PRISMA flowchart. ACEs, adverse childhood experiences; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Data extraction

Data extraction was carried out by the first author in accordance with the review protocol, on which a screening form and data extraction form were based (online supplemental appendix 2). Extracted data included article information, population characteristics (age, ethnicity and type of population), study design, ACE assessment, types of healthcare utilisation and effect estimates.

Assessment of risk of bias

Study quality was assessed using a tool based on three established frameworks (online supplemental appendix 3).18,20 The checklist included five domains: study design, participant selection, outcome selection and measurement and statistical analysis. Each domain was scored 0–2 points, except for statistical analysis, which allowed up to 3 points for more rigorous confounder adjustment. Total scores ranged from 0 to 11 and were grouped into four categories: poor (0–3), fair (4–6), good (7–8) and excellent (9–11). The domains used in this approach were most relevant for the observational designs in our review and allowed consistent scoring across heterogeneous studies. This framework was intended for comparing methodological quality within this review, not for benchmarking against other reviews.

Data synthesis and analysis

Results were synthesised in accordance with the synthesis without meta-analysis (SWiM) reporting guidelines.²¹ For the 15 included papers, harvest plots were used to visualise effects by study. Studies were ranked by quality score, and exposures and outcomes were categorised to summarise the heterogeneous findings. These categories were created postinclusion, based on data availability. Exposures were grouped into two types: studies using the ACE-10 index defined by Felitti et al or studies using other assessments of parental adversity.²² Outcomes were grouped into preventive, primary and secondary care. Preventive care referred to routine services (eg, routine preventive paediatric check-ups and vaccinations), considered preventive in European systems but often delivered in US primary care. Because these levels are related hierarchically, this approach enabled assessment of their mutual influence. Within these groups, we examined heterogeneity by interpreting variation in ACE operationalisation, outcome definitions, population characteristics and quality scores. Reported effect sizes were extracted without transformation and not converted to standardised metrics due to study heterogeneity. Results were synthesised by direction of effects, supported by metrics and CIs where available. All studies were included in the narrative synthesis, with greater emphasis placed on those of higher methodological quality, validated ACE measures and well-defined outcomes.

Results

Table 1 shows the characteristics of the included studies. They were conducted between 2008 and 2024. The majority, 14 out of 15, were conducted in the USA. The total sample size across all studies was 106 775, with a median of 448 (range: 61–93 391). While only a small portion of studies reported parental age at the time of the study, there was a wide range (15–50+ years), and the mean maternal age was 34.4 years. The included publications studied children across all age groups (0–17 years), with a mean child age of 6.8 years. Among the 14 studies that reported ethnicity, seven had a majority Black population and seven had a majority white population. Most studies were cross-sectional observational. The mean study quality was 5.5 (SD 1.9), with a range of 3 to 9. The majority were rated as fair (61.9%), with only one (6.7%) rated as excellent. The findings are summarised in figure 2, ordered by exposure and outcome domain and ranked by quality score. Studied exposures and outcomes were highly heterogeneous. The exposures, outcomes and effect sizes are presented in table 2.

Table 1. Overview of included study characteristics.

Author (publication year)	Country	Sample size	Study design	Quality score (range 0–11)	Predominant ethnicity	Focus population
Ciciolla et al27	USA	164	Prospective observational	Fair (5)	White, 41%	Low-income population
Day et al23	USA	240	Cross-sectional observational	Poor (3)	Black, 81%
Day et al28	USA	1465	Retrospective cohort	Good (8)	Black, 75%
Dennis et al 51	USA	326	Prospective observational	Poor (3)	White, 89.9%	Mothers with chronic pain
Eismann et al24	USA	454	Retrospective cohort	Fair (6)	White, 53.5%
Guyon-Harris et al31	USA	120	Prospective longitudinal	Poor (4)	Black, 56%
Haynes et al 52	USA	1515	Cross-sectional observational	Poor (4)	White, 56%
Lê-Scherban et al5	USA	350	Cross-sectional observational	Poor (4)	Black, 45.1%
Santavirta et al30	Finland	93 391	Retrospective cohort	Good (8)	Not reported	Finnish, born 1945–2010
Shah et al26	USA	1320	Prospective cohort	Good (8)	White, 62%	Hospitalised children
Simons et al34	USA	80	Retrospective observational	Fair (5)	Black, 93%	Drug abusing mothers
Sosnowski et al29	USA	531	Prospective cohort	Excellent (9)	White, 62%
Wamser-Nanney et al33	USA	845	Cross-sectional observational	Fair (5)	Black, 59.5%	At-risk children
Wamser-Nanney et al32	USA	448	Cross-sectional observational	Fair (5)	Black, 48.7%	Maltreated children
Weigert et al25	USA	235	Cross-sectional observational	Fair (6)	White, 48%	Children at PED

PED, paediatric emergency department.

Figure 2. Evidence for effects of parental ACEs on healthcare utilisation. A harvest plot matrix covering three levels of care (columns) and exposures, categorised as original ACE index or any other ACE measure (rows). Each association is represented by a mark in each row for which that study reported relevant results. The height of the bar and the number in the bar indicate the quality of the study. Some papers reported more than one association. *The studied outcome in this research was treatment completion. Hence, a decline in treatment completion should be regarded in the same way as increased healthcare utilisation, as both indicate a higher burden on the healthcare system. The publications studied the same population, but 2 years apart. ACE, adverse childhood experience.

Table 2. Overview of all exposures, outcomes and effect sizes.

Author (publication year)	Parental childhood exposure	Exposure source	Outcome care level	Child outcome studied	Outcome source	Effect size (95% CI, P value)
Ciciolla et al27	ACEs (categorical)	ACE-10	Secondary	NICU admission	Survey	↑ OR 8.70 (1.34–56.65,<0.05)
Day et al23	ACEs (categorical)	ACE-10	Preventive	Influenza vaccination	Survey	↑ OR 1.18 (1.00–1.39, 0.049)
Day et al28	ACEs (categorical)	ACE-10	Secondary	Antibiotics use	Register	→ OR 1.31 (0.86–2.00, NR)
	Other: physical abuse	Survey	Secondary	Antibiotics use	Register	↑ OR 1.90 (1.20–3.20, NR)
	Other: sexual abuse	Survey	Secondary	Infection-related visits	Register	↑ HR 1.27 (1.05–1.53, NR)
Dennis et al51	ACEs (numerical)	ACE-10	Primary	Somatic symptoms	Survey‡‡	→ r 0.02 (NA, NA)
Eismann et al24	ACEs (categorical)	ACE-10	Preventive	Routine preventive check-up	Register	↑ IRR 1.60 (1.04–2.47, 0.01)
	ACEs (categorical)	ACE-10	Preventive	Delayed/missed vaccination	Register	→ IRR 1.17 (0.88–1.56, 0.28)
	ACEs (categorical)	ACE-10	Primary	Sick visits	Register	→ IRR 0.99 (0.94–1.04, 0.65)
Guyon-Harris et al31	Other: sexual abuse	CTQ	Secondary	Emergency department visits	Survey	↑ IRR 1.06 (NR, 0.007)
Haynes et al52	ACEs (categorical)	Survey†	Primary	Having a healthcare provider	Phone	→ χ² (NA, 0.477)
Lê-Scherban et al5	ACEs (numerical)	Survey†	Preventive	Dental examinations	Phone	→ OR 1.14 (0.96–1.37, NR)
	ACEs (numerical)	Survey†	Primary	Having a healthcare provider	Register	↓ OR 0.84 (0.52–1.35, NR)
Santavirta et al30	Other: WWII evacuation	Register	Secondary	Psychiatric hospitalisation	Register	↑ HR 2.04 (1.04–4.01, NR)
Shah et al26	ACEs (categorical)	ACE-10	Secondary	Unanticipated readmission	Register	↑ OR 1.69 (1.11–2.60, 0.02)
Simons et al34	Other: emotional neglect	CTQ	Secondary	Treatment completion	Register	↓ OR 0.11 (NR, <0.001)
Sosnowski et al29	ACEs (numerical)	ACE-10	Secondary	NICU stay	Register	→ OR 0.97 (0.85–1.05, NR)
Wamser-Nanney et al33	Other: victimisation*	Survey	Secondary	Mental healthcare use	Survey	→ OR 1.17 (0.95–1.22, 0.12)
Wamser-Nanney et al32	Other: victimisation*	Survey	Secondary	Mental healthcare use	Survey	→ OR 1.10 (0.99–1.20, 0.17)
Weigert et al25	ACEs (numerical)	ACE-10	Secondary	Emergency department visits	Register	→ IRR 1.10 (1.00–1.02, NR)

Some publications appear more than once, as they report on multiple outcomes. ↑ and ↓ represent positive and negative association, respectively. → indicates no association.

^*Victimisation in these studies was a combination between childhood and adulthood adversity.

^†These surveys contained the ACE-10 questions but were adapted by either adding or removing items.

^‡Somatic symptoms were collected using the validated Children Somatic Symptoms Inventory.22 53 54

ACE-10, ACE questionnaire by Felitti et al; CTQ, Childhood Trauma Questionnaire; NICU, neonatal intensive care unit; WWII, World War II.

Preventive care

Four preventive care outcomes were studied in three distinct reports, as Eismann et al studied two outcomes. The outcomes assessed included vaccination uptake and routine preventive paediatric check-up attendance. Mixed evidence was found in these studies, with some variation based on the type of exposure measured. Day et al reported increased odds of paediatric influenza vaccination as caregiver ACEs increased (OR 1.18; 95% CI 1.00 to 1.39; p=0.049; quality score 3).²³ Eismann et al found that maternal, but not paternal, ACEs were associated with an increased probability of missing routine preventive paediatric check-ups (incidence rate ratio (IRR) 1.60; 95% CI 1.04 to 2.47; p=0.010; quality score 6).²⁴ In the same study, they did not find an association of parental ACEs with delayed or missed immunisations (as recommended by the American Academy of Pediatrics) (IRR 1.17; 95% CI 0.88 to 1.56; p=0.28; quality score 6). Lê-Scherban et al found no association with dental examinations in the past 12 months (OR 1.14; 95% CI 0.96 to 1.37; p=NR; quality score 4).⁵

There were no studies included in which exposures other than the original ACE index were studied in combination with preventive care services. Overall, findings on preventive care utilisation were mixed. Among the studies using categorical ACE exposures, results were inconsistent—one found a positive association, one negative and one no association. The only study using a numerical ACE score reported no association with general preventive care. Two of the four studies had a low to fair study quality, which limits the certainty of findings in this domain. There were also some differences in the studied population: Eismann et al reported predominantly white children (53.5%), Day reported predominantly Black children (81%), and Lê-Scherban reported 45.1% Black parents.

Primary care

Across primary care outcomes, none of the four studies found an association with parental ACEs. There were no studies with other exposures. The outcomes included sick visits or having a usual source of healthcare. In mothers with chronic pain, Dennis et al found no correlation between maternal ACE scores and child somatic symptoms (r=0.02; p=not reported, quality score 3). Eismann et al, using data from an urban paediatric clinic population, reported no association between parental ACEs and sick visits (IRR 0.99; 95% CI 0.94 to 1.04; p=0.65; quality score 6). Haynes et al found no relationship between categorical ACE scores and whether the child had a regular healthcare provider (p=0.477; quality score 4). Finally, Lê-Scherban et al found no association between parental ACE scores and having a usual source of care (OR 0.84; 95% CI 0.52 to 1.35; p=not reported; quality score 4).

Results were consistent in this domain; even though differences in population, quality score and methodology were observed, none of the studies found an association. The consistent absence of associations suggests limited evidence for a relationship of parental ACEs with paediatric healthcare utilisation. However, with three studies of fair quality and one of poor quality, the mean quality of studies in this domain (4.25) was lowest of all domains, so conclusions should be drawn with caution.

Secondary care

10 studies reported on secondary care utilisation, investigating 12 associations. Seven of these associations were positive, indicating a higher likelihood of healthcare utilisation for children of parents with ACEs; four showed no effect, and one indicated reduced odds. Studied outcomes included both hospital-related visits and treatment (completion). In the ACE index group, Shah et al reported higher odds of unanticipated hospital readmission in children of parents with four or more ACEs (OR 1.69; 95% CI 1.11 to 2.60; quality score 8), Weigert et al found an association with more paediatric emergency department visits (IRR 1.013; 95% CI 1.001 to 1.024; p=not reported; quality score 6) and Ciciolla et al reported substantially increased odds of neonatal intensive care unit (NICU) admission for children born to mothers with six more ACEs (OR 8.70; 95% CI 1.34 to 56.65; p<0.05; quality score 6).25,27 Day et al and Sosnowski et al reported no association between parental ACEs and children’s antibiotic use and NICU stay, respectively.^{28 29}

In the six studies reporting on other ACE measures, four positive associations were found. Day et al reported increased odds of antibiotic use in children of caregivers who were physically abused in their childhood—despite observing no association when ACEs were measured cumulatively (OR 1.90; 95% CI 1.20 to 3.20; p=not reported; quality score 8). Moreover, in the same study, they found that caregiver childhood sexual abuse was associated with earlier infection-related clinical encounters (eg, primary care, intensive care admission, emergency department visit, hospital admission) in children (HR 1.27; 95% CI 1.05 to 1.53; p=not reported).²⁸ Additionally, Santavirta et al found an association with more psychiatric hospitalisations in children whose mothers had been evacuated from Finland during World War II (HR 2.04; 95% CI 1.04 to 4.01; p=not reported; quality score 8), and Guyon-Harris et al reported slightly elevated rates of infant emergency department visits when mothers had been sexually abused in childhood (IRR 1.06; 95% CI not reported; p=0.007; quality score 4).^{30 31} Wamser-Nanny et al reported twice on a study in maltreated or at-risk children. They found a trend but no association between parental victimisation history and children’s mental healthcare use in 6-year-old maltreated children (OR 1.10; 95% CI 0.99 to 1.20; p=0.07; quality score 5) and in 8-year-old at-risk children (OR 1.17; 95% CI 0.95 to 1.22; p=0.12; quality score 5). In this study, parent victimisation combines childhood and adult trauma; hence, these results cannot be attributed to childhood adversity alone.^{32 33} Simons et al reported the only negative association: drug-abusing mothers who had experienced emotional neglect and mental health issues were less likely to complete treatment programmes with their children (OR 0.11; 95% CI not reported; p<0.001; quality score 5).³⁴ However, lower treatment completion could increase children’s disease burden, further supporting the trend from three high-quality studies that showed parental ACEs lead to higher secondary healthcare utilisation.

Findings in this domain were mixed, but most associations were positive. There was considerable heterogeneity in study design and outcomes. In the other ACE measure group, all but one study investigated a predominantly Black population. Despite these differences, the overall evidence tentatively points towards a link between parental ACEs and higher secondary care use, though this should be interpreted with caution.

Discussion

While our findings do not provide consistent associations across all healthcare domains, they suggest a positive trend for specific forms of healthcare utilisation. Parental ACEs showed no associations with primary care and yielded mixed results with preventive care, but a more consistent trend towards higher secondary care utilisation emerged.

To our knowledge, no other reviews have examined the association between parental ACEs and children’s healthcare utilisation. In our review, 85% of studies assessing hospital presentation or admission in children reported a greater likelihood of healthcare utilisation. While this evidence is distinct from literature on individuals’ own childhood adversity, findings from adults also show that ACEs are linked to a higher risk of hospitalisation.³⁵ Hence, ACEs might not only be associated with personal healthcare utilisation but also with that of the next generation. Taken together, these strands of evidence suggest that adversity may contribute to increased reliance on acute rather than planned care, with health problems in children of parents with ACEs potentially going unrecognised until symptoms escalate.

The pathways through which ACEs influence child health are likely multifactorial, involving social context, parental coping and behaviour and genetic factors. For example, the observed trend in secondary care, especially hospitalisation, may result from higher disease risk in children of parents with ACEs, possibly from lower engagement with early healthcare services, although the evidence for preventive care in our review was mixed. Parents with ACEs may be unaware of their own or children’s increased health risks, as the intergenerational effects—potentially mediated by mechanisms such as allostatic load—are often gradual, biologically embedded and hard to detect.³⁶ Socioeconomic disadvantage and adversity-related behaviours—such as avoidance of care and reduced coping capacity—may reduce timely healthcare seeking, which may explain the lower treatment completion and more missed routine preventive paediatric check-ups in this review.3437,42

Epigenetic changes have also been proposed as a pathway, but there is no consensus on this. Because of germline reprogramming, critics question the likelihood of epigenetic inheritance.^{43 44} On the other hand, a recent scoping review identified DNA methylation differences in 19 of 22 studies on children of trauma-exposed parents.⁴⁵ These findings are, however, based on observational studies and cannot establish causality. Nevertheless, given the growing body of such evidence and the likelihood that multiple biological and psychosocial mechanisms act together, epigenetic processes may contribute alongside other pathways. Importantly, such biological changes do not occur in isolation but in constant interaction with social environments—poverty, chronic stress and relational adversity are key contexts through which epigenetic processes may be activated or sustained or even lead to syndemics.⁴⁶ A multimodal approach that integrates (neuro)biological, psychosocial and behavioural pathways may therefore provide the most comprehensive understanding of how parental ACEs shape child health.⁴⁷

Another explanation is a possible mismatch between the levels of care. Improved coordination between preventive, primary and specialist services could help prevent unnecessary escalation of care—particularly in ACE-exposed individuals navigating complex care systems. The psychosocial needs underlying the reduced preventive care utilisation could be addressed by intervening at the family level, and early recognition of parental ACEs could help reduce disengagement or delayed care. This focus allows for targeting root causes of healthcare utilisation and improving long-term outcomes, while also prompting reflection on whether conventional service use metrics fully capture the support needs of these families. Barriers such as limited system trust, financial constraints or use of alternative care may contribute to underutilisation of formal services, despite elevated health risks in children of parents with ACEs.

The findings in this research should be interpreted in light of some methodological considerations. First, although the quality of the included studies was generally fair to good, the applied analytics in these studies was highly heterogeneous, which precluded meta-analysis. Moreover, the SWiM-based synthesis revealed substantial methodological heterogeneity across studies, particularly in the operationalisation of parental ACEs. Some studies used numerical ACE scores, others categorical groupings, and multiple classification thresholds were applied. These differences complicate cross-study comparison and may influence effect estimates, with categorical approaches often yielding stronger associations. As noted in the results, where studies using categorical ACEs often produced stronger associations than numerical ones, this highlights the need for greater consensus on ACE operationalisation in future research.

In addition, the use of AI-assisted screening may have introduced bias in study selection. Although active-learning tools such as ASReview can improve efficiency while maintaining high recall, their performance depends on the training data and reviewer decisions made during screening. This may have influenced which studies were prioritised and could have led to inadvertent exclusion of relevant records despite our validation steps.¹⁴ However, screening all 6558 records manually would likely also have introduced human error and subjective bias, reflecting the trade-offs inherent to both approaches.

Moreover, the 10-item index does not capture adversities such as poverty, community violence or peer victimisation, and the binary format may oversimplify complex experiences.⁴⁸ Because morbidity and healthcare use are socially patterned, socioeconomic factors are particularly relevant. In our review, these factors were built into the quality scoring and synthesis—indeed, two-thirds of the studies (10 out of 15) included socioeconomic adjustments. Nevertheless, given the strong social clustering of adversities across social and health domains, future research should go beyond simple adjustment and examine effect modifications or interactions between ACEs and socioeconomic disadvantage. Such approaches align with syndemic frameworks that emphasise how adverse social conditions and health risks cocluster and exacerbate each other in vulnerable populations.⁴⁶ Despite these limitations, the ACE framework remains a frequently used and pragmatic approach in public health research.⁴⁹ Yet, it should be appreciated as a tool that describes probabilities on a population level, rather than as evidence of deterministic causal pathways in individuals. ACEs need to be understood within their social and economic context, as structural conditions shape both exposure and outcomes. Recognising these broader factors is essential for designing effective and equitable prevention strategies.⁵⁰

Additional limitations include the possibility of publication and language bias, as only peer-reviewed English-language studies were included. Moreover, generalisability is limited: 14 of 15 studies were conducted in the USA, and many included ethnically specific or high-risk populations. This may limit the applicability of findings to broader or European contexts. Further research in diverse settings is needed to strengthen the evidence base and inform cross-context health policy.

While there is ample research on the association between ACEs and disease outcomes, evidence on how parental ACEs affect healthcare utilisation in children remains limited. Nonetheless, a trend towards a higher risk of secondary care utilisation was observed. This review highlights the need for further research to clarify these secondary consequences and to inform policy development. Investing in early identification and integrated, proactive care approaches may help reduce the intergenerational burden of ACEs on health systems and society.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.