Commentary: Understanding Research,Policy,and Practice Issues in Adoption Instability

Overview of Articles

Palacios, Rolock, Selwyn, and Barbosa-Ducharne (in press) provide a well-researched overview of the adoption breakdown literature, addressing two important empirical questions: What is the incidence of adoption breakdown and what factors are most often associated with this type of family instability? In addressing the first question, the authors note that the incidence of adoption breakdown depends upon how this term is defined and the methodology used to investigate it. They highlight the important distinction between the removal of a child from the adoptive family prior to legal finalization of the placement (as noted above, typically referred to as adoption disruption) and various temporary or permanent forms of discontinuity in the adoptive parent–child relationship after the adoption has been legally completed. Although Palacios and his colleagues cite a summary of data by Smith (2014b) as providing a reasonable estimate of adoptive family breakdown in the United States—that is, 9.5% adoption disruption and at least 5–10% reentering out of home care after the adoption order—they also note that methodological issues make it difficult to provide truly reliable estimates of these types of family breakdown or compare figures across studies. Moreover, they point out that the incidence of postpermanency discontinuity is linked to the length of time since adoption. They cite data reported by Rolock and Testa (2008) to support their point: At 2 years postadoption, approximately 2% of adoptions had experienced some form of discontinuity, with the figure doubling to 4% by 5 years and more than doubling again to 9% by 10 years postfinalization. Finally, the authors note that the reported rates of adoption breakdown generally are higher in the United States than in Europe, although the basis for this difference is unclear. The takeaway from their excellent literature review is that we still do not have reliable statistics on the incidence of various types of adoptive family instability, especially those that occur in the postadoption period and those that occur without regulation and guidance by the authorities.

The authors also review the literature on those factors most often associated with adoption instability. A triad of factors is identified, including those related to child characteristics, parent/family characteristics, and child welfare and other professional practices (see also reviews by Coakley & Berrick, 2008; Faulkner, Adkins, Fong, & Rolock, 2016). Among child factors, age at adoptive placement is perhaps the most common variable correlated with adoption breakdown; the older the child at placement, the greater the risk for adoption instability. Age at placement, however, is most often viewed as a proxy for experiencing a greater number of preplacement adversities such as neglect, abuse, orphanage life, previous disrupted placements, and exposure to caregiver psychopathology and relationship violence. The Portugal disruption study (Barbosa-Ducharne & Marinho, this issue) added another child factor to those found previously—experiencing postremoval neglect by birth family (e.g., parents did not seek reunification or visit after removal). Many adversities, especially a history of abuse and multiple moves prior to adoptive placement, have also been linked to preadoption and postadoption family maladjustment and breakdown. So too has the presence of behavioral and emotional problems in the child that emerge soon after placement and then escalate as the placement continues, often leading to aggressiveness and violence within the family. On the other hand, the authors note that no consistent pattern has been found between adoption breakdown and placement of siblings together or separately, a common issue confronted in making child welfare placement decisions. The article by Selwyn (this issue) addresses sibling dynamics in at-risk adoptive families in more detail.

A number of parent and/or family factors linked to elevated rates of adoption breakdown were identified in the review article, including parental motives that focus primarily on meeting their own needs as opposed to the child’s needs, ongoing conflicts in the couple relationship, difficulty attaching to the child, inability to cope with children’s problems, and inflexibility in adapting to children’s characteristics. Related to these issues is the failure of parents to develop realistic expectations regarding their child and themselves as caregivers. Unrealistic parental expectations are one of the strongest predictors of adoption breakdown (Reilly & Platz, 2003; Wright & Flynn, 2006). Sometimes the problem is tied to inadequate information about the child’s history or a failure to understand the parenting implications of the history. At other times, it is related to a failure to adapt to new information or the changing needs and behaviors of the child. And in still other situations, it is linked to a lack of knowledge and skills to manage emerging parenting challenges and a resistance to seeking help from others. In fact, help seeking and finding adequate supports by parents were reported by the authors to be a protective factor, stabilizing placements and strengthening family connections. Finally, the authors note that, in most studies (see also Festinger, 1986; McRoy, 1999; Rosenthal, Schmidt, & Conner, 1988) but not all (Reilly & Platz, 2003; Selwyn et al., 2014), when parents have a previous relationship with a child prior to adoption, either as a foster parent or as a biological family member, the adoption is more stable.

Although Palacios and his colleagues report on a number of preadoption and postadoption service variables that have been linked to adoption breakdown rates, Faulkner, Adkins, Fong, and Rolock (2016) suggest that the available evidence is not sufficiently compelling to draw definitive conclusions about some of these service correlates of adoption instability. Nevertheless, those with the strongest support as noted by the review authors are described as follows. Adoptive parent preparation, particularly workers’ sharing relevant background information on the child and explaining their developmental and parenting implications, has been found in some studies to be a protective factor resulting in lower rates of adoption breakdown (Barth & Berry, 1988; Paulsen & Merighi, 2009). In contrast, excessive “stretching” by workers in matching parents and children can lead to increased family difficulties and risk for placement breakdown. Such situations involve workers’ encouraging parents to consider adopting children whose characteristics or previous life experiences are beyond those initially identified by the parents as acceptable for consideration. Regarding postadoption service issues, early identification of problems and the provision of timely support and services strengthen families and reduce placement instability. Conversely, the lack of available and adequate professional services increases the risk of both preadoption disruption and postpermanency discontinuity. So too does high caseworker turnover, less experience of the worker, and the involvement of multiple workers within and across agencies such that there is little consistency in professional involvement with and knowledge about families over time. This problem can undermine awareness of the family’s evolving needs and timely referral to appropriate support services. Ongoing availability and quality of postadoption services have also been linked to adoptive family adjustment and breakdown rates. Although many agencies provide adoption support soon after placement and/or the legal completion of the adoption, problems facing adoptive families often emerge years after placement. In too many cases, postadoption services and support are unavailable in the family’s community or if they are, the professionals often lack adoption clinical competency (Brodzinsky, 2013, 2015; Pinderhughes & Brodzinsky, in press; Smith, 2014b).

Barbosa-Ducharne and Marinho’s article, “Beyond the child’s age at placement: Risk and protection factors in preadoption breakdown in Portugal,” focuses on adoption disruption by comparing 71 disrupted preadoption placements to 71 intact adoptions, matched for age at placement (7.5 years for each group)—a design used in only a few previous studies and one that reveals the possible impact of other variables on adoption stability to be examined independent of placement age (McRoy, 1999; Palacios, Jiménez-Morago, & Paniagua, 2015, Smith & Howard, 1991). The percentage of adoptive placements that disrupted over a 4-year period was 5.8%. Although comparative analyses between the intact and the disruption groups found several differences related to child, parent, and professional practice variables identified in the review article by Palacios et al. (in press), a final regression model composed of six variables explained 54% of the variance and accurately predicted approximately 80% of all adoption success and disruption cases. Adoption disruption was more likely when children experienced postremoval neglect by birth family (e.g., parents did not seek reunification or visit after removal) and exhibited significant behavior problems following placement. Conversely, it was less likely when children were described as more often displaying intense crying and having mild developmental difficulties, which the authors suggested might stimulate parents’ attempts to comfort and help their children, thereby reducing risk of disruption. Placement disruption was also positively associated with rigid parental attitudes and difficulty in acknowledging adoption challenges and specificities. Finally, regarding professional practices, adoption disruption was less common when caseworkers’ records indicated more complete information about parents’ skills and knowledge to deal with adoption challenges; it was more common when families were involved with multiple agencies. Interestingly, the number of worker–family postplacement contacts did not differentiate the two groups, raising concerns by the authors that either the workers did not identify family difficulties in a timely manner or did not attempt to provide more support.

Rolock et al.’s article, “A comparison of foster care reentry after adoption in two large U.S. states,” identified preadoption characteristics associated with postadoption reentry into foster care, a subject that has been rarely studied. Using administrative data, the outcomes of adoptions of youth aged 16 and under who were adopted between 2000 and the end of 2010 in two U.S. states (Illinois and New Jersey) were followed for 5–15 years (depending on their date of adoption), through 2015 or until their 18th birthday, whichever came first. Approximately 5% of these youth had reentered foster care by 2015—an incidence that would likely be higher if all youth had reached the age of 18 by the end of the study. The likelihood of reentering care increased with the length of time since adoption, peaking at 10–11 years postadoption in both states.

This study has many strengths including a large sample from two states, totaling over 38,000 cases. The administrative data were analyzed for each state separately and for the two states combined and enabled analysis of the impact of differences in policy and practice between the two states child welfare systems. For example, Illinois had a higher percentage of children adopted at age 3 or older and significantly more children in care at 3 or more years—two of the four significant variables in the final multivariate survival analysis. Children adopted at age 3 or older were 128% more likely to reenter foster care than those adopted under the age of 3. These differences likely contributed to a higher incidence of foster care reentry in Illinois—6% of adopted youth when compared to 4% in New Jersey. Other variables in the final model that were associated with a higher risk of reentry included the number of foster care placements (each move was associated with a 15% increase in reentry risk) and being African American (a 30% higher risk of reentry than other children). Kinship care, having lived in a group home, and gender were not significantly related to the level of risk. The authors conclude that identifying factors related to increased reentry risk can enable targeting preventive, adoption services to at-risk families.

Selwyn’s study, “Children adopted from care: Sibling relationships in adoptive families in crisis and those that had disrupted,” explored the positive and negative aspects of sibling relationships in adoptions that either had broken down or were in a state of crisis after the adoption court order and whether sibling conflict contributed to the removal of an adopted child. This is a secondary analysis of a larger study on “postorder disruption” or postadoption instability in England and Wales. The original study utilized administrative, survey, and standardized measures and interview data (Selwyn et al., 2014; Wijedasa & Selwyn, 2017). This mixed-methods design enables a more in-depth picture and understanding of the dynamics involved in adoption instability. Based on survey data from 390 families of adopted adolescents, the researchers found that parents reported their adoption outcomes as: going well (37%), highs and lows but mostly highs (30%), major difficulties/in crisis (24%), and left home prematurely (9%).

For the qualitative study in this issue, Selwyn analyzed content related to siblings in interview data from 41 families who had experienced postadoption breakdown and 42 families whose adoptions were in crisis but remained intact. Siblings were defined as any two or more children in the adoptive family. By adolescence, only 22% of these 83 families described normal sibling relationships with a balance of positive and negative interactions, but most sibling relationships were characterized by significant problems, adding immensely to parental stress. The disturbed sibling relationships were primarily characterized by violence and aggression toward a sibling and/or toward a parent, controlling and manipulating a sibling, extreme jealousy and concerns about not being loved equally, sexual abuse, and for a few, overdependence on each other to the point of anxiety when separated. Children with very challenging siblings often felt neglected by their parents, afraid of their sibling, and reluctant to bring friends home; yet, if the difficult sibling left home, the other children were left with complicated and ambivalent feelings—for example, relief, worry, embarrassment in the community, and grief. Overall, only 8 of the 41 adoption breakdowns were directly attributed by parents to conflict among siblings. Parents had asked for the child to be removed when they did not feel able to keep all the children safe. Also, siblings placed together were more likely to experience adoption breakdown than those placed sequentially. Recommendations for practice include more in-depth assessment of sibling dynamics and interventions to improve sibling relationships.

Paniagua, Palacios, Jiménez-Morago, and Rivera’s study, “Adoption breakdown in Spain: A survival and age-related analysis,” focused on the association between adoption breakdown and two child-related variables: duration of placement prior to breakdown and age at adoptive placement. All known cases of adoption breakdown from 2003 to 2012 in the Spanish region of Andalusia were identified using information provided by adoption professionals, with subsequent in-depth case file review by the investigators. Preadoption and formalized adoptions, as well as domestic and intercountry adoptions, were considered. Of the 7,006 adoptive placements reviewed, an overall breakdown rate of 1.3% was found. Child, family, and professional predictors of adoption breakdown were analyzed separately for preadoption placements (PA) and court finalized adoptions (CA).  For the PA group, child’s age at placement was significantly associated with adoption breakdown. Compared to children placed between 2 and 6 years, those placed between 6 and 10 years ended their placements 3 times faster, and those placed after 10 years had a 10-fold increase in the speed of adoption breakdown. Children in the PA group with attachment difficulties and those whose parents manifested unrealistic expectations also showed a shorter time in adoption breakdown. On the other hand, early therapeutic intervention for this group helped sustain the family for a longer period prior to breakdown. Regarding the CA group, only two variables remained significant in the final regression model. First, the speed of breakdown was significantly faster for intercountry adoptions than domestic adoptions (mean time between placement and breakdown was 54 months for intercountry adoptions compared to 108 months for domestic adoptions). Second, the rate of breakdown was significantly and positively associated to child’s age of placement; the older the child at the time of placement the more quickly the adoption ended. The final question addressed by the authors was the relationship between incidence of adoption breakdown and children’s age at placement. Like other studies, the researchers found a clear link between these variables; namely, compared to a very low number of placement breakdowns in children placed from birth to 2 years, there was a significant linear increase in breakdowns for children placed between 2 and 6 years, 6 and 10 years, and beyond 10 years. The takeaway from this study is the important role of the child’s age at placement for both the duration of placement prior to breakdown and the incidence of breakdown. In addition, the study points out the importance of examining predictors of breakdown separately for preadoption cases and postadoption cases.

All of the studies in this issue, except one, include data on postadoption instability. These studies and others confirm that the preteen and teen years are a critical period for adoption instability. In fact, an analysis of U.S. national adoption data indicates that over 80% of adopted children who reenter foster care do so as preteens or teens (Smith, 2014b). During this life stage, adoption issues often become more intense, trauma and loss-related symptoms frequently resurface, the parent–child relationship too often deteriorates, and the risk of instability skyrockets.

Research Recommendations

The articles in this special section raise important conceptual and empirical issues for future researchers to consider. First, the need for more uniform terminology and differentiation in various types of adoption instability is obvious. Comparing studies and developing clear and effective policies and practices require that professionals make sure they are referring to the same phenomena. This is especially true when trying to understand the incidence and predictors of the many different types of adoption instability that have been identified. Studies that group different types of adoption instability together or generalize from one type (e.g., reentry into care) to others make it difficult for professionals to have a clear understanding of these phenomena and to develop appropriate prevention and intervention strategies to deal with the unique circumstances that could be associated with each of them.

A second implication of these articles is the need to improve and expand our methodological approaches to studying adoption instability. The four empirical articles in this special section highlight the value of different research designs for investigating adoption instability. For example, the use of administrative data (e.g., Paniagua, Palacios, Jiménez-Morago, & Rivera, in press; Rolock et al., in press) allows investigators to capture entire populations of child welfare cases in states, regions, or even nations, providing for greater generalization of findings and avoiding some of the inherent difficulties in using convenience samples. At the same time, administrative data typically only allow for a “broad stroke” perspective on placement outcomes but do not provide for the type of fine-grained analyses of other research designs that may be necessary for the development of prevention and intervention strategies, specifically at the case level. Barbosa-Ducharne and Marinho’s study (this issue), in contrast, highlights the value of using a matched sample design that allows researchers to explore the relationship between specific variables and adoption breakdown, independent of other variables—in this case, child’s age at placement. Such methodological control allows for clearer understanding of the role of specific child, parent/family, and practice factors contributing to adoption breakdown. But even this type of methodology has limitations for understanding adoption breakdown at the specific case level than other research approaches. Selwyn (this issue), using qualitative methodology, highlights the value of small scale but intensive interviewing of families as a means of allowing key stakeholders in the adoption breakdown process to provide adoption professionals with information about the complexities of family experiences that are seldom captured in large-scale quantitative research designs. A similar approach has been used by other researchers seeking to understand the lived experiences of former foster youth and adults related to placement breakdowns (Chambers et al., 2017; Rolock & Pérez, 2018; Unrau, Chambers, Seita, & Putney, 2010). Because adoption breakdown occurs for different reasons among different families, information from small-scale, qualitative studies often provide useful guidance to professionals in developing strategies for prevention and intervention.

There is a clear need for more longitudinal research in this area. As noted by Palacios et al. (in press), adoption instability increases as children get older and the longer the duration of the adoption. Studies that only examine families at a specific time in their life cycle are likely to underestimate the true incidence of family breakdown, hindering our understanding of the phenomena and perhaps undermining appropriate efforts for prevention and intervention.

Also, we need to find out what happens to adopted youth who leave their homes. For example, what happens to those who reenter foster care? Are they ever reunited with their adoptive families? An analysis of U.S. Adoption and Foster Care Analysis and Reporting System (AFCARS) data cited earlier (Smith, 2014b) contrasted the outcomes of adopted children who reentered care to foster youth who had not been adopted.  The adopted youth were 3½ times more likely to be placed in nonfamily settings than were other foster children (largely because a higher percentage of older youth are placed in congregate care) and nearly 3½ times more likely to leave care at 18 without legal permanence than other foster children. For adopted children who reentered and then exited care, 36% were reunified with their adoptive families (as compared to 52% of other foster children), and 34% experienced adoption dissolutions. Of those in the latter group, 61% were adopted again (Smith, 2014b), a pattern also noted in the review article in this issue. These findings indicate that youth whose adoptions breakdown are some of the most vulnerable in the child welfare system. They have lost two sets of parents and many siblings and are at high risk of never having a stable, permanent family and an internal sense of security and belonging.

In addition to improved and more varied methodologies, there is a need for more cross-regional and cross-national research. The article by Rolock et al. (in press) points out that in different regions of the same country, adoption breakdown can occur for different reasons and at different rates. As Palacios (2009) has emphasized, adoption occurs and is experienced in a specific context and must be understood within an ecological perspective. We argue that adoption breakdown must also be understood within this type of perspective. Thus, future research needs to examine the similarities and differences in adoption instability within different regions of the same country and from one country to another. In summary, the research implications from these articles suggest that understanding adoption instability is best accomplished through a multimethod, data convergence strategy. When findings converge, using different methods, different samples, and across nations, they provide a firmer basis for generalization and developing more effective prevention and intervention strategies. At the same time, if data suggest that adoption instability and its correlates differ from one region or country to another, it provides important information that allows professionals to tailor prevention and intervention strategies in ways that are appropriate to context.

Finally, there are many disciplines that study adoption adjustment including social work, developmental and clinical psychology, psychiatry, marriage and family therapy, anthropology, demography, and an emerging focus on the biological (e.g., genetic, neurological, and neurochemical) correlates of adoption-related adjustment difficulties (Palacios & Brodzinsky, 2010). Efforts to integrate these various conceptual and empirical perspectives through multidisciplinary studies should lead to a broader and more nuanced understanding of the complexity and interaction of factors affecting adoption adjustment including patterns of adoption instability.

Practice and Policy Recommendations

Facilitating permanence for children is a primary goal of child welfare services. But as noted at the beginning of our commentary, there are different aspects of permanence that need to be considered and supported. Ensuring legal and residential permanence is a critical step, but it does not necessarily guarantee that children will experience psychological permanence. Ongoing conflicts within the adoptive family can seriously damage and sometimes destroy children’s sense of connection, nurturance, trust, and safety in relation to their parents. Furthermore, when children are legally and residentially separated from previous caregivers (e.g., birth parents, foster parents, guardians, or adoptive parents), it does not necessarily mean that psychological or relationship permanence ends. Many children continue to feel strongly and positively connected to previous caregivers even after being separated from them, whether temporarily or permanently. For example, children’s relationships with birth family are often supported by some degree of contact with birth relatives when it is in the children’s best interests, which can be helpful to them in terms of coping with loss and fostering positive self-esteem and identity (Grotevant, Lo, Dunbar, & Fiorenzo, 2017; Grotevant, Perry, & McRoy, 2005; Neil & Howe, 2004; Pinderhughes & Brodzinsky, in press). In other cases, when youth in adoptive or guardianship families end up living outside the legally permanent family, relationships with their adoptive parents or guardians often continue through visitations, involvement in therapy, or some other type of support. Sometimes children even return to the adoptive or guardianship family or, if not, at least have an ongoing relationship with some family members as they move through adolescence and into young adulthood. With these points in mind, we offer the following recommendations in support of all aspects of permanence and to reduce all types of placement instability.

Agencies must eliminate all unnecessary delays in the adoption process. When children come to the attention of child protective services, planning for permanence should be implemented from the very beginning of the decision-making process. Timely assessment of birth parents’ needs and provision of appropriate services for them must be completed to support family reunification, whenever possible. As in other countries, in the United States, kin are preferred placement resources, if there are those who are suitable. Therefore, timely identification of available and appropriate extended family members needs to be completed around the time of the initial placement in order to avoid unnecessary moves and promote kin adoption or guardianship when children cannot be returned to their birth family. When children must be placed in foster care, it is important that the first placement be a good fit for children’s needs and parents’ capacities. All efforts must be made to minimize the need for multiple placements prior to reunification or permanence through adoption or guardianship.

Conclusion

Although adoption has proven to be a very successful option for children needing a permanent and nurturing family when their family of origin is unable or unwilling to provide for them, the challenges facing many adoptive families cannot and should not be ignored. Too many children enter their new families with histories of trauma and early life adversities that contribute to adjustment difficulties and placement instability. To better prepare and serve adoptive families, adoption and mental health professionals need to better understand the risk and protective factors associated with family instability. The articles in this special issue focus much needed attention on an underresearched area in the adoption field and provide valuable insights that hopefully will lead to better adoptive parent preparation and more effective postadoption support and services.