When policy meets practice: the untested effects of permanency reforms in child welfare

Journal of Sociology and Social Welfare, March, 2006 by Amy D'Andrade, Jill Duerr Berrick

Since the majority of children who enter out-of-home care reunify with their families (Wulczyn, 2004), it makes sense to target this challenging and resource-intensive practice toward those families least likely to reunify. In fact, according to Weinberg and Katz, "... requiring concurrent planning for all cases seriously distorts the model" (1998, p.12). A tool was developed to assist workers in targeting appropriate families, based on practice wisdom accumulated at Lutheran Social Services (Katz & Robinson, 1991). The tool lists conditions describing families believed to have a low likelihood of reunification, and hence who would be most appropriate for concurrent planning ("poor prognosis indicators"), as well as conditions thought to identify families likely to reunify ("family strengths indicators"). The tool has been adopted by many states implementing concurrent planning (D'Andrade, Mitchell, & Berrick, 2003). Table 2 lists poor prognosis indicators from the California version of the Katz tool.

The ability of this targeting tool to accurately and consistently identify families unlikely to reunify is unknown. Certainly, some of the poor prognosis indicators reflect conditions established through research to be negatively associated with reunification. For example, indicators #8 and #18 refer to substance abuse, indicator #12 to criminal activity, and indicator #10 to lack of parental visitation. However, some indicators describe a slightly different condition than the one known to be associated with reunification failure. For example, indicator #8 limits its scope to those parents who are addicted to illegal drugs; indicator #18 appears to refer only to prenatally substance-abusing mothers who received and disregarded medical advice. For a few indicators, the logic associating the indicator with reunification is unclear, such as for indicator #5. While the lifestyle described is certainly illegal, it is not apparent why it would be likely to result in a reunification failure. Critical words in the poor prognosis indicators are also undefined, leaving their meaning open to subjective interpretation. For example, indicator #12 refers to a "recent" history of "serious" criminal activity, but does not specify what time frame is meant or what nature of criminal activity constitutes "serious."

The validity and reliability of the poor prognosis indicators will be important for states and agencies hoping to target concurrent planning toward families less likely to reunify, in order to maximize scarce resources and limit the emotional burden for caregivers. But some have voiced concerns that concurrent planning itself may threaten reunification efforts, if for example social workers fail to provide adequate reunification services due to time constraints (Stein, 2000), or if through inadequate implementation fost-adopt caregivers fail to support natural parents (Weinberg & Katz, 1998).

A further issue then is that very little is known about the effects of concurrent planning. Quantitative evaluation studies are relatively few, and their conclusions arguably equivocal due to design and measurement problems. Several articles report outcomes for the original program in Washington state (Katz, 1990; Katz, 1996 as cited in Schene, 2001), but the lack of any comparison groups makes it difficult to make definitive conclusions about the program's effectiveness. Other studies make use of various types of comparison groups (Brennan, Szolnoki, & Horn, 2003; Kelly & Taylor, 2000; Martin, Barbee, Antle, & Sar, 2002; Monck, Reynolds, & Wigfall, 2003; Schene, 1998) or employ correlational designs to examine implementation of concurrent planning (Martin et al., 2002; Potter & Klein-Rothschild; 2001). These studies have found concurrent planning associated with the following positive outcomes: higher rates of permanency at one year (Potter & Klein-Rothschild, 2001; Schene, 1998); shorter lengths of stay (Martin et al., 2002; Monck et al., 2003; Schene, 1998); fewer placement changes (Monck et al., 2003); lower placement costs (Kelly & Taylor, 2000; Schene, 1998); and improved parental compliance (Martin et al., 2002).