"I Had Most of the Stuff I Needed Except a Decent Foster Home": Predictors of Youth Absences from Foster Care

Background/Purpose: Running away from foster care is a known risk factor associated with poor intermediate and long-term outcomes, such as placement stability and permanency, respectively. However, few studies have documented the nuances of runaway behavior to reveal opportunities for intervention. In seeking to build deeper knowledge on runaways, this study sought to understand: (1) whether subpopulations may be identified as more likely to runaway from foster care, and (2) whether specific critical life events serve as risk or protective factors influencing likelihood of runaway experiences.

Methods: A mixed-method approach was used to address study objectives. Quantitative analyses answers the first question on subpopulations using a secondary data sample of all youth, ages 12-17, in foster care with one agency in July 2021 (N=1,127). Latent Class Analysis investigated whether latent classes representing youth subpopulations were characterized by family history, family connections, and system experiences. Next, binary and multinomial logistic regressions (LRs) examined whether class membership significantly predicted youth runaways. Finally, qualitative interviews were conducted with 20 youth in out-of-home placement. Interviews were audio recorded, transcribed, and coded in Dedoose.

Results: Question one results indicated that four latent classes predicted the likelihood of runaways: Class 1: High Environmental Stability/High Familial and Environmental Connections (i.e., low system involvement and highest probabilities of being in their preferred school environment and being with sibling(s)). Class 2: Moderate Connected/Moderate Environmental Risk (i.e., moderate/high system involvement and average of 14 placement, most in home settings; highest probability of removal reasons due to inadequate housing). Class 3: Low Environmental Stability/Low Community Connection/High Familial Connection (i.e., average of 30 placements; low probability of preferred school; high probability of developmental, medical, and mental health issues; high instances of acute psychiatric stays; more females; removed mainly due to neglect). Class 4: Low Stability and Low Environmental Connections (i.e., substance use and mental health concerns; highest probability of deceased/incarcerated parent; older males). Multinomial LR found class membership, age, and number of siblings were significant predictors of runaway behaviors.

Question two results revealed five overarching themes. 1. Absence of current connections was linked to historical family bonds/attachments. 2. Fear and/or lack of control of own circumstances perpetuated by system and/or placement constraints increased runaway decisions. 3. Supportive placements extending beyond basic needs and promoting belonging and a sense of normalcy mitigated runaway risk factors. 4. Maintaining family bonds was a strong protective factor influencing youth decisions to run-away or return. 5. System shortcomings failed to protect against runaway risk.

Implications/Conclusion: Youth placement stability is a key predictor of positive long-term outcomes for youth in foster care. Decisions to runaway are complex, involving many co-occurring and overlapping factors. Understanding these factors allows programs to identify youth needs early on and provide the supports necessary to mitigate risk. Key recommendations for practice include providing support for more frequent and extended visitation, application of shared parenting practices, promoting youth voice in case decisions, review and relaxing of standard placement rules, and additional screening for risk paired with wraparound supports.