Associations between mother-preschooler attachment and maternal depression symptoms: A systematic review and meta-analysis

Abstract

The current study aimed to systematically review and meta-analyze concurrent and longitudinal associations between maternal depression symptoms and mother-child attachment during the preschool period (aged 2 to 7 years) as assessed using the coding systems by Cassidy and Marvin (1992) and Main and Cassidy (1988). The review was pre-registered with PROSPERO (International Prospective Register of Systematic Reviews; Registration number CRD42017073417) and was conducted in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A total of 7,969 records were screened and 18 articles were deemed as eligible for inclusion in the review. Studies were reviewed using qualitative synthesis techniques and, where appropriate, meta-analysis. Qualitative synthesis indicated that mothers of disorganized/controlling children most consistently reported the highest levels of depressive symptoms, both concurrently and longitudinally. The association between disorganized/controlling child attachment and concurrent maternal depressive symptoms was significant (n = 1,787; g = 0.27, 95% CI [0.13,0.40]), and was not moderated by sample type, child gender, or risk of bias. Findings of a relationship between child attachment insecurity and maternal depressive symptoms must be qualified due to significant within-study heterogeneity and publication bias. Results suggest that maternal depressive symptoms may confer risk for disorganized/controlling attachment during the preschool period.

Fig 1. Included study flow chart following PRISMA guidelines.

Fig 2. Risk of bias scores.

Bar graph illustrating proportion of studies that fulfilled each risk of bias consideration.

Fig 3. Forest plot for meta-analysis of maternal depressive symptoms and concurrent attachment insecurity in 2- to 5-year-olds.

Hedge’s g point estimates are depicted by filled squares, with square sizes reflecting the relative weight of each study’s effect size in the analysis. The filled diamond reflects the summary effect size. RE = Random effects model. If a square or error bars cross 0, this indicates no difference between mothers of secure and insecure preschoolers. Squares to the right of zero indicate higher depression levels among mothers of insecure preschoolers, relative to mothers of secure preschoolers.

Fig 4. Contour-enhanced funnel plot for meta-analysis of maternal depressive symptoms and concurrent attachment insecurity in 2- to 5-year-olds.

Each dot represents an included study in the meta-analysis. An over-representation of dots in the shaded (non-white) areas of statistical significance is suggestive of publication bias.

Fig 5. Forest plot for meta-analysis of maternal depressive symptoms and concurrent attachment disorganization in 2- to 5-year-olds.

Hedge’s g point estimates are depicted by filled squares, with square sizes reflecting the relative weight of each study’s effect size in the analysis. The filled diamond reflects the summary effect size. RE = Random effects model. If a square or error bars cross 0, this indicates no difference between mothers of organized and disorganized preschoolers. Squares to the right of zero indicate higher depression levels among mothers of disorganized preschoolers, relative to mothers of secure preschoolers.