Sequelae of child maltreatment: Umbrella synthesis of 148 meta-analyses on the mental health correlates.

BACKGROUND

Numerous meta-analyses have established associations between child maltreatment (CM) and mental health difficulties (MH). However, variation exists between meta-analyses regarding the magnitude of these predictions.

METHODS

A systematic, quantitative umbrella synthesis (i.e., meta-analysis of meta-analyses) was undertaken to describe the associations between various types of CM and MH. Meta-analyses were included if they examined CM, including but not limited to retrospective reports in adulthood, and MH at any point. Included forms of CM were: physical abuse, emotional abuse, sexual abuse, neglect, and exposure to intimate partner violence. MH outcomes were: externalising problems, internalising problems, thought problems, suicidal distress, substance misuse, and other psychological difficulties. Searches were run in January 2024. Random effects models were created in R version 4.2.0.

RESULTS

We analysed and combined effect sizes from 148 quantitative meta-analyses, including 668 effect sizes and over 9.5 million data points. CM was associated with all MH outcomes: (1) externalising problems (r = 0.21; 95% CI = 0.18-0.24; k = 32), (2) internalising problems (r = 0.22; 95% CI = 0.20-0.24; k = 46), (3) thought problems (r = 0.24; 95% CI = 0.21-0.27; k = 38), (4) suicidal distress (r = 0.23; 95% CI 0.18-0.28; k = 19), (5) substance misuse (r = 0.19; 95% CI = 0.13-0.26; k = 13), (6) other psychological difficulties (r = 0.24; 95% CI = 0.20-0.28; k = 50). Associations tend to be of similar magnitude for different forms of CM.

CONCLUSION

CM is robustly associated with MH. A parsimonious explanation for these findings would be a common mechanism(s) or a general psychopathology factor conferring high-risk for different mental health difficulties following CM. The results possibly question the conventional wisdom that suggests some forms of maltreatment are intrinsically more harmful to mental health than others. However, further work is required to understand how potentially confounding factors (e.g., age, measurement of CM) influence these associations.