What's working for Children: A Policy Study of Adoption Stability and Termination

Among the Institute’s principal findings:

• Terminations Seldom Occur. Concerns that policies promoting adoption would lead to increased terminations generally appear unfounded. In fact, for a variety of reasons, the vast majority of adoptions from foster care remain stable over time.
• Data Collection is Inadequate. An array of problems prevent a thorough understanding of the reality on the ground (of disruption and dissolution rates) or of the impact of various risk factors. These include a lack of uniformity in definitions and inadequate data collection.
• Nontraditional Parents are Effective. Families headed by single, foster, older, lower-income and less-educated parents (as well as by kin) have better stability rates than the average, and therefore provide important opportunities for recruitment and placement from foster care.
• Post-Adoption Services are Vital. In addition to careful matching and preparation before a placement, providing assistance of various kinds after the child is in a home is vitally important in helping to minimize disruptions and promote adoption stability.

While the federal government requires states to collect some data relating to the child welfare system, it does not mandate that they track terminations of adoptive placements and finalized adoptions.¹ Indeed, there is no consensus on definitions of such terminations among practitioners, policy makers or researchers. In the absence of federal or state data-collection requirements, social scientists have examined relatively small, geographically limited, snapshot samples, primarily before ASFA was enacted.

To help fill the information gap, the Adoption Institute conducted a comprehensive study on foster care adoption termination and stability, funded by the David and Lucile Packard Foundation, to draw general conclusions about termination rates and the factors associated with them, and has formulated recommendations to enhance data collection and adoption stability. This information is critical to ensure that thoughtful placements are made from the outset for the 129,000 children currently waiting in foster care to be adopted; to improve the prospects that positive interventions will reduce the incidence of disruption and dissolution; and to provide assistance in family situations with chronic difficulties that do not result in terminations.

The Adoption Institute studied the relevant social science research over the last 20 years, and conducted a survey of 15 states’ information-collection capabilities, yielding important findings about data collection, termination rates and risk factors. The terms “disruption” and “dissolution” are defined in this study, respectively, as the termination of an adoptive placement before the adoption is legally finalized and the termination of an adoption after it is legally finalized.

Recommendations for policy and practice:

• Improve Matching, Preparation and Education. Careful family evaluations should be systematically conducted to enhance the matching process, including thorough strengthsbased assessments as well as tools/protocols such as the Model Approach to Partnerships in Parenting (MAPP) and the Parents’ Resource for Information, Development and Education (PRIDE). Greater use of kin and foster parents also evidently increases stability.
• Develop More-Comprehensive Disclosure Policies. Many terminations (and problems within intact families) occur because parents have unrealistic expectations based on inadequate information; for this reason, among others, outcomes could be improved if states routinely provided comprehensive information in case management, including portable health records.
• Provide Consistent, Reliable Support. Post-adoption services are increasingly viewed as critically important, both to lessen the possibility of termination and to enhance the functioning of intact adoptions; families with children who have histories of severe abuse, neglect or other comparable challenges are particularly vulnerable. Availability of mental-health services and access to temporary residential care, when necessary, also appear to enhance stability.

Perhaps even more important than the incidence of disruption and dissolution are the factors associated with their occurrence. Though states are not collecting the relevant statistical information, researchers have identified factors that represent increased risk for disruption and dissolution: children’s ages, placement history and behavioral issues; parents’ expectations; and systemic factors such as matching and information sharing.

The Adoption Institute’s survey of 15 geographically, demographically and structurally diverse state child welfare agencies found that most of the surveyed states (ranging from 12-14) were capturing basic foster care and adoptive placement statistics, but their capability to collect outcome data on adoption placements and finalized adoptions varied widely. Even when outcome data were available, officials conceded their numbers may not accurately reflect disrupted and dissolved adoptions in their states. Some states that reported their management systems captured placement and outcome data nevertheless failed to provide that information for this survey. Questions therefore remain about the accuracy of the states’ characterization of their systems’ capacity and/or whether this information is regularly generated and reviewed to assess placement efforts. State child welfare agency officials also noted, and an analysis of the data reveals, an important caveat to the information reported: questionable data integrity.

Of the 15 child welfare systems, eight reported in the telephone survey, conducted between December 2001 and February 2002, that they collect data on foster care adoptive placement disruptions, and six reported that they collect data on adoption dissolutions. Of the states reporting capacity to collect data, many were unable or unwilling to provide the information on the data collection forms for this survey. Only four provided disruption data, while five supplied dissolution data; just two submitted both. Unfortunately, it is not clear to the research team whether this reported information represents accurate numbers of disrupted and dissolved adoptions. During interviews, officials from states that tracked this information questioned their own data’s quality and reliability, and attributed their inability to report accurate statistics to a lack of common definitions of disruption and dissolution, inadequate management information systems, and failure of overworked field staff to collect and/or enter data. Officials also reported there is a widely held agency perception that disruption and dissolution rates are low, and therefore that improving data collection is not a priority.

The available data in the Adoption Institute study showed termination rates are indeed low. Nevertheless, greater understanding of what keeps adoptions stable – and what leads to their breakups – will presumably have major positive consequences. Among them are that such knowledge will help in the placement of children into families with greater prospects of stability and, significantly, will help improve daily life for families that have already been formed (since it is widely understood that other reasons for low termination rates include adoptive parents’ resolve to keep their families intact, as well as legal and financial hurdles to termination).

Available data indicate low termination rates:

• State disruption rates are much lower than the estimates reported in earlier empirical studies on disruption; research studies from 1980 to the mid-1990’s found overall disruption rates of 10% to 27%, though rates as low as 3% were recorded for younger populations.
• States reporting disruption information to the Adoption Institute had low rates, ranging up to 8.4%, for adoptive placements in 1999. Dissolution rates were similarly low, from 0.4% to 5.4% for 1998 adoptions and for children entering care in 2000.
• The General Accounting Office, based on disruption data from 20 states and dissolution data from 21 (with 11 providing both) estimated 5% of adoptions planned in 1999-2000 disrupted, while 1% of those finalized in 1999-2000 later resulted in legal dissolution.
• Most of the body of studies on adoption disruption occurred in the 1980s, and as Goerge, et. al’s (1995) large-scale study in Illinois documented, there has been a decline in the disruption rate since that time. Concerns that ASFA would reverse this trend are not supported by data from the GAO or the Adoption Institute’s survey of states.

The Adoption Institute’s research also found no apparent association between characteristics of child welfare systems and data collection. That is, no type of system (state- or county-based) or traits of a system (for instance, whether it completed Statewide Automated Child Welfare Systems or had been awarded adoption bonuses) seemed to make it more or less likely to have accumulated the relevant disruption and dissolution data.

Information gaps left by the states remain unfilled. There is a clear need for academic studies of disrupted placements and dissolved adoptions for children placed after the implementation of ASFA in 1997. And, as noted above, the older studies do not use consistently defined terms and protocols, many do not track children over time, and their samples often are small and limited to a specific geographic area. Among the other problems relating to data collection are that policymakers and practitioners do not have comprehensive, comparable information about disruption and dissolution rates and associated risk factors; most states do not collect disruption and dissolution data, and among the ones that do, it is unclear whether they are accurately capturing the information; the child welfare field in general has not adhered to standard definitions of disruption and dissolution and tools for uniformly measuring them; there is a dearth of empirical studies of disrupted placements and dissolved adoptions for children placed after the implementation of ASFA in 1997; and state collection of disruption and dissolution data is incomplete and the data is therefore questionable.

Conclusion:

Empirical studies have documented specific child- and parent-related factors posing greater risk for adoption stability. While knowledge of these risks cannot prevent termination of all adoptive placements or finalized adoptions, it can promote more effective matching of children and families and more careful attention to addressing known risk factors.

For example, in placing a child with an extensive history of trauma and current severe behavior problems, ideally a family could be found (or trained) with an understanding and acceptance of such children, as well as the skills and resources to parent such a child. At the very least, these risk factors should not be glossed over or ignored in pre- or post-adoption work with the family and child. In addition, systemic factors contribute to adoption instability, so accountability related to these risks needs to be built into child welfare practice – ranging from professional educational programs and agency supervision of workers, to system review standards and performance measures.

Finally, if meaningful empirical examination of adoption outcomes and risks is to occur in the future, attention must be given to data systems and adherence to data collection protocol in the field. It is unlikely that this will occur without federal mandates requiring case-specific reporting of adoptive placements and longitudinal data on adoption outcomes. Such reform is necessary for the permanency goal of adoption to be truly permanent for these children.

¹ The Adoption and Foster Care Analysis and Reporting System (AFCARS) rules mandate that states collect case-specific data on all children in foster care, and data on all adopted children placed by the state child welfare agency, and for whom the state provides adoption assistance, care, or services. Neither Statewide Automated Child Welfare Systems (SACWIS) nor AFCARS requires states to specifically collect foster care adoption disruption or dissolution data. AFCARS asks whether children entering the foster care system have ever been adopted, but this data element can refer to international and private domestic adoptions, in addition to previous foster care adoptions. Additionally, the information does not indicate whether the entry is a temporary displacement or a legally dissolved adoption. As such, this AFCARS element does not represent foster care adoption displacements or dissolutions.