A Meta-Analysis of Attachment-Based Interventions for Poor Mothers and Their Children

Research on attachment theory has underscored the importance of early mother-child interaction.  A major premise of attachment theory is that optimal child development is based on the formation of a secure base, which is a function of sensitive dyadic interaction between the primary caregivers and the infant.  A secure base facilitates the development of adaptive representations of the world and gives coherence to self-representation.  Studies show that insecurely attached infants have problematic developmental trajectories throughout their childhood and adolescence.  Adverse factors such as the deleterious effects of poverty can impede the caregivers’ ability to provide the infant secure attachment.  Given the importance of early attachment for children’s current and future functioning, it is crucial to investigate attachment-based interventions that can enhance child functioning.

This paper represents a meta-analysis of interventions focused on improving the attachment relationship between poor mothers and their children.  The search strategy for collecting intervention studies published in English was: Social Services Abstracts, PsycINFO and Medline were searched (search parameters 1990 to 2014) with keywords “attachment intervention” “poor families,” “poverty,”   “insecure attachment,” and “relation-based interventions.”  Twenty-one studies were identified.  Ten studies were excluded as follows: they included middle-class families, did not measure changes in infant attachment, did not have a control group, included serious medical conditions, premature infants, adolescent mothers, and/or maternal mental illness. A detailed coding system was used to rate study design, sample, and intervention characteristics.  Design characteristics included sample size, randomization, and the study’s attrition rate.  Intercoder reliabilities were established (k = 10; mean r = 0.98, range 0.90-1.00). Data analysis was conducted using Comprehensive Meta-Analysis program with attention to whether there were positive changes in infant attachment as a treatment outcome.  The combined samples in all studies were:  N=1281, Treated (n=699) Controls (n=582).  Results showed a medium effect as the odds favored the treatment group  (fixed-effects;  1.925, p<.000; 95% CI [1.511 – 2.452]; random-effects; 2.523 p<.001; 95% CI [1.438-4.428]; Q 47.938,  p<.000).  In follow-up, studies were segregated by treatment type and a sensitivity analysis was performed.  Two of the studies focused on social support yielding an effect size (fixed and random) 4.297, p<.000 95% CI [1.983-9.313]. Four studies targeted maternal sensitivity, effect size (random)2.481, p<.003 95% CI [1.365-4.507].  Four studies consisted of psychodynamic therapy, which resulted in an non-significant effect size of 1.125, p<.650 95% CI [.677-1.870].  One study consisted of the Circle of Security  - effect size 29.333, p<.000.

The findings from this study add to the Evidence-Based Practice repository and serve to enhance practice decisions by providing research evidence for interventions for poor people.  Children with insecure attachment have been shown to be at risk for conduct disorder, delinquency and crime.  These trends, in part, involve a lack of intentionality, limited concern with mental states, lack of sense of self, and limited metacognitive capacity, which are deficits linked to insecure attachment.  In an era when generic therapies no longer exist, this study is an attempt to identify goals and treatment recommendations for prevention and treatment for poor families and their offspring.