Long-Term Effects of Institutional Care and Enhanced Attachment Relationships on Close Adolescent Friendships

Abstract

This study examined whether early institutional rearing and attachment security influence the quality and quantity of friendships at age 16 in 138 participants, including children abandoned to institutions in Bucharest, Romania, who were randomized to care as usual (n = 45, 26 female), or foster care (n = 47, 25 female), and a never-institutionalized group (n = 46, 18 female). Adolescents in the foster care group with secure attachment to their foster mothers at 42 months were comparable to never-institutionalized adolescents in having more friends and more positive behaviors with their friend during dyadic interactions, compared to the foster care group with insecure attachment and care as usual group. Interventions targeting early child-caregiver attachment relationships may help foster the ability to build positive friendships in adolescence.

Figure 1.

CONSORT diagram.

Figure 2.

Results comparing secure and insecure subgroups within the FCG to the CAUG and NIG. Note. *p< .05. +p< .06. Medians and means (diamonds) with standard errors are shown in box plots. All analyses adjust for sex differences. n in number of friends =137 (45 CAUG, 25 insecure-FCG, 22 secure-FCG, 45 NIG). Analyses on friendship quality and observed behaviors exclude those paired with acquaintances or strangers and further adjusts for whether the peer had a history of institutionalization: n in friendship quality=112 (37 CAUG, 16 insecure-FCG, 15 secure-FCG, 44 NIG), n in observed behaviors= 114 (36 CAUG, 19 insecure-FCG, 17 secure-FCG, 42 NIG.

Figure 3.

Indirect effects of age of placement on number of close friends and observed off-task behaviors through attachment security within the foster care group. Note. Standardized estimates are shown. All models adjust for sex differences: n=46 in analysis for number of close friends. The model testing observed off-task behaviors excludes those paired with strangers and acquaintances and further adjust for whether the peer had a history of institutionalization: n=35 in analysis of observed behaviors.