| This study examines the importance of child level of functioning in understanding the outcomes of Colorado's mental health family preservation services. Three primary issues are addressed: (1) the nature of mental health IFPS programs, (2) measurement of children's level of functioning, and (3) correlates of placement prevention.;Using a sample of 5000 child mental health recipients, an instrument validation study is conducted of the Colorado Client Assessment Record, which assess nine domains of functioning. Exploratory and confirmatory factor analyses indicate that a nine-factor model best fits the data, a finding which is consistent with the instrument's conceptual structure.;The nature of the family preservation intervention is examined by qualitative inquiry into intervention fidelity across the eight service sites. Site visits included semi-structured interviews with line staff and program administrators. Strong similarities are found in: (1) service philosophy, and (2) intervention focus. Differences exist in program interface with host agencies and local communities. Seven of the eight sites are reliably implementing the intervention model.;Using a sample of 316 family recipients of family preservation services, correlates of placement prevention are examined using logistic regression. Four sets of variables are examined: child and family characteristics, children's level of functioning, presenting situation and service targets. During the first three months following service, children's socio-legal functioning, previous group home placement and the presence of mental health presenting problems increase the odds of placement, while increased numbers of followup services decrease the odds of placement. Between 3 and 6 months post-intervention, projected placement in foster care and a child behavior presenting problem reduce the odds of placement. Service targets of alcohol monitoring and time management increase the odds of placment dramatically. Time management is found to represent a cluster of variables related to parental incapacity and provision of basic family needs.;These results are useful for modification of services to better respond to identified risk factors. However, use of results in structuring access to services is not warranted, due to limits of the predictive abilities of the models. Implications for program design and for future research are discussed. |
