Data were collected through a series of focus groups and 6,607 records of police protective custody (PPC) admissions to the Wichita Children's Home (WCH). Systematic predictive modeling (logistic regression) was used to differentiate between children likely to need continued services under the child welfare system and those who could remain with their families. WCH provided a database on admissions from July 1, 2000, through June 30, 2003. Models were developed to inform policymakers on the most appropriate intake assessment tool for identifying those children most likely to require State custody at the end of the PPC process. Alternatively, the models could also identify children who could safely be dismissed from the PPC process and returned to their families, thereby reserving scarce resources for those who most need them. There was no evidence that the police were abusing the protective custody process. Most of the children entering the protective custody process needed State services at some level; however, some children could be diverted from entering the WCH while the State determines whether services, including custody, are necessary. Based on study findings, this report recommends that the PPC process be streamlined by developing a universal intake process that would involve using one form for PPC by the police, the social workers, and WCH personnel. Another recommendation is to permanently assign a social worker and assistant district attorney to WCH in order to expedite the PPC process. Further, through an active and ongoing recruitment process, the network of emergency foster care homes should be expanded. 7 tables, 3 notes, and 36 references