A Multilevel Model of Factors Related to Substance Use Behavior in Youth in Foster Care

Substance use (SU) in adolescence is a major public health concern and is associated with a variety of negative outcomes for youth (Lansford et al., 2008; Henry et al., 2008). While research has addressed risk factors related to youth SU in the general population, youth in foster care demonstrate higher rates of SU disorders than youth in the general population (Pilowsky & Wu, 2006; Vaugh et al., 2007); however, little is known about what factors are predictive of SU among foster youth. Further, prior analyses often fail to account for the nestedness inherent within the data on youth residing in out-of-home placement, namely that foster caregivers can vary greatly in their knowledge of youth behavior. Youth in foster care represent a population at higher risk for negative outcomes (Thompson & Auslander, 2007); therefore, focused and effective prevention and treatment tools are needed for this population.

The study examined the role of child characteristics, diagnosis status, and placement type on SU behavior in 260 youth (aged 12 to 18 years old) enrolled in foster care in a multilevel analysis accounting for the influence of Level 2 variance from caregiver reporter. The gender distribution of participants was approximately equal (46% female), with an average age of 15.03 years (Sd=1.90 years). Based on caregiver report 22% of youth had used alcohol and 27% had used drugs. Findings revealed that 15% of observed variation in drug use and 19% of observed variation in alcohol use was due to differences at Level 2, identified as differences across foster parents and residential staff. A model building strategy was employed, starting with age as a predictor and adding in placement type, gender, and then diagnostic status. In models accounting for Level 2 contribution to variance in outcomes, examination of AIC/BIC values and chi-square difference tests revealed a significant predictive relation of placement type, age, and diagnosis status on SU behavior. Gender did not emerge as a significant predictor. These findings suggest factors traditionally known to relate to SU in youth (i.e., age, comorbid diagnoses) as well as factors specific to youth in foster care (i.e., placement type) contribute significantly to variance in SU behavior; consequently, a more specific look at how intervention and prevention strategies may influence SU outcomes for youth residing in foster care placements may be necessary. For example, gender was nonsignificant in this sample suggesting that approaches for prevention or intervention need to target males and females. Further, given emerging evidence for higher rates of SU in youth in residential placements, prevention strategies designed for these youth may be more effective than global strategies for all youth in care.