Statistics from Altmetric.com
Question: What is the relationship between child physical abuse, emotional abuse and neglect on mental and physical health outcomes?
Outcomes: Any health-related adverse outcome listed on the Global Burden of Diseases, Injuries and Risk Factors Study.
Design: Systematic review and meta-analysis.
Data sources: MEDLINE, EMBASE and PsychINFO were searched up to 26 June 2012. Reference lists of included studies were hand searched and other additional studies were identified through contact with study authors.
Study selection and analysis: Published and peer-reviewed cohort, cross-sectional and case–control studies examining non-sexual child maltreatment as a risk factor for poor health were included if the related health outcomes or behavioural risk factors were among those listed on the Global Burden of Diseases, Injuries and Risk Factors Study. Studies that reported exposure to only combined types of abuse were excluded. ORs were meta-analysed using Meta XL based on a quality-effects model which is a modified form of a fixed-effects inverse variance method that allows additional weighting by study quality. Prespecified subgroup analyses assessed the impact of: gender, geographic location (high income vs low-to-middle income), type of sample (population-based vs non-representative samples), abuse measurement (self-reported vs official records), health outcome assessment (self-reported vs structured clinical interview), prospective versus retrospective assessment of abuse and neglect, and adjustment for confounders. Evidence for causality was assessed using the Bradford Hill criteria.
In total, 124 studies met inclusion criteria, of which 16 were prospective cohort studies and the majority were cross-sectional studies. Most studies (112 studies) were from Western Europe, North America, Australia and New Zealand. Individuals had a higher risk than non-abused individuals having depressive disorders: if they were physically abused (OR 1.54, 95% CI 1.16 to 2.04), emotionally abused (OR 3.06, 95% CI 2.43 to 3.85) or neglected (OR 2.11, 95% CI 1.61 to 2.77). Individuals had a higher risk than non-abused individuals of developing anxiety disorder if they were physically abused (OR 1.51, 95% CI 1.27 to 1.79) emotionally abused (OR 3.21, 95% CI 2.05 to 5.03) or neglected (OR 1.82, 95% CI 1.51 to 2.20). Individuals were more likely to display suicidal behaviour if they were physically abused (OR 3.00, 95% CI 2.07 to 4.33), emotionally abused (OR 3.08, 95% CI 2.42 to 3.93) or neglected (OR 1.85, 95 CI 1.25 to 2.73). Sexually transmitted infections and risky sexual behaviour were more common among those who had been physically abused (OR 1.78, 95% CI 1.50 to 2.10), emotionally abused (OR 1.75, 95% CI 1.49 to 2.04) or neglected (OR 1.57, 95% CI 1.39 to 1.78). Analysis of prospective studies supported the relationships as above, with the exception of the effects of emotional abuse which was not studied in prospective studies, and the relationship between neglect and self-inflicted injury, which did not reach significance. There was some evidence that non-sexual maltreatment was also associated with some chronic diseases, lifestyle risk factor and other physical health outcomes, but findings were not strong or consistent.
The evidence suggests a causal relationship between non-sexual child maltreatment and various mental disorders, suicide attempts, sexually transmitted infections and risky sexual behaviour.
Notes: Inconsistencies were identified for how child abuse and neglect were defined across studies. There were no prospective studies that investigated non-sexual child maltreatment and suicide attempt or ideation.
Child abuse and neglect have long been recognised as particularly deleterious for physical and mental health across the lifespan. Research in this area has proliferated, yet there has not been a comprehensive review of studies of physical abuse, emotional abuse and neglect. Norman and colleagues fill this gap using state-of-the-art meta-analytic methods, including a systematic review which identifies 124 relevant studies. Sexual abuse was explicitly excluded, because it was included in a prior meta-analysis1 thus limiting direct comparisons of the health consequences among all four types of maltreatment.
Notably, the authors report that most published studies still use retrospective designs to determine associations between child maltreatment and later outcomes. Although they identify 16 prospective studies (13%), ten of those were based on the same longitudinal study by Widom and colleagues (please refer to Table 3 in the article on studies with the following first author: (1) Widom, (2) Wilson and (3) Nikulina). Based on their analysis, the authors conclude that there is robust evidence linking all three types of maltreatment and a number of health outcomes. However, an issue not addressed in this article concerns the variability across studies in the assessment and categorisation of maltreatment. For example, retrospective studies are often unclear about how the maltreatment was defined or determined.2 Studies using information from Child Protective Services (CPS) case records have frequently found the type of maltreatment identified by the CPS did not correspond with findings from in-depth examinations of case records. In addition, the majority of cases indicated more than one type of maltreatment.3 This imprecise classification, and high likelihood of co-occurrence of multiple types of maltreatment make it difficult to link a specific type of maltreatment to a specific health outcome.
Given that child maltreatment research involves complex definitional issues and a preponderance of retrospective studies, caution should be used when inferring causality between one particular type of maltreatment and later health outcomes. Our ability to develop prevention or intervention programmes for the consequences of specific types of maltreatment (including sexual abuse) requires more prospective longitudinal research.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online references
Sources of funding: University of Queensland.
Competing interests None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.