Abstract
Purpose:
The two goals of this article are the analysis of the duration of adoptive placements ending in breakdown and the role of age at placement in the breakdown experience.
Method:
All known cases of adoption breakdown during a whole decade in Andalusia, a Spanish region, were studied. Preadoption and formalized adoptions, domestic and intercountry adoptions were included. Data were analyzed using survival analysis, Cox regression, χ2, and rate ratio analyses.
Results:
The duration of adoptive placements ending in breakdown, significantly shorter in intercountry adoptions, is associated with a configuration of characteristics in the child, the adoptive parents and adoptive family life, and professional intervention. Among child-related factors, age at placement is of special relevance for the breakdown experience.
Conclusions:
Placements involving older children last less and break down more frequently, but are not condemned to failure. They need to be better supported with protective factors compensating the risks.
As part of the special section on adoption breakdown, this article adds to the increasing efforts to gain a better understanding of the factors and processes involved. Within the diversity of approaches to studying adoption breakdown summarized by Palacios, Rolock, Selwyn, and Barbosa-Ducharne (this issue), the study to be reported herein refers to adoption breakdown in all adoptive placements occurring in a region of Spain during the decade 2003–2012, no matter the pre- or post-formalization stage in the adoption process, the domestic or international origin of the child, or any other of the subjects’ characteristics. The duration of the adoptive placements before separation and its associated factors will first be explored. Subsequently, given the prominent role played by the child’s age at placement on the breakdown experience, this variable will be analyzed more in depth. To the best of our knowledge, this is one of the first studies to consider in detail the duration of life together before final separation and the first survival analysis with Spanish breakdown cases.
Although there is some information about the duration of the adoptive placements ending in breakdown (e.g., Selwyn, Meakings, & Wijedasa, 2015), this refers mainly to the length of time from placement (or from the court decision) to breakdown, with no detailed analysis of the factors associated with this duration. On the contrary, there is a fair amount of research and consensus around the fact that, more than with any single factor, adoption breakdown is linked to an accumulation of circumstances related to a triad including the child, the adoptive parents and adoptive family life, and the professional interventions around these children and families (Palacios et al., this issue).
As shown in the review article in this special section, when adoption breakdown literature explores the factors associated with breakdown, the relationship between older age at placement and adoption breakdown is one of the more consistent findings, with less agreement regarding almost all other variables, with the exception of behavioral problems. Given the importance of age at placement in the extant literature and the particular emphasis on this variable in this article, it needs further consideration. As discussed by Palacios et al. (this issue), age at placement is often a proxy for the accumulation of adversity before adoption. Typically, children adopted at older ages have been exposed to maltreatment, toxic stress, and institutionalization during a more protracted period of time than those adopted younger, and this prolongation is associated with children’s more troubled emotions and behaviors.
Although the critical role of age at placement presents research unanimity, the basic consensus is that placements involving older children are more at risk of breakdown than placements involving younger ones. However, each study has a different approach to the definition of “older children.” Some studies used age as a continuous variable (e.g., Goerge, Howard, Yu, & Radomsky, 1997); as an example, Smith, Howard, Garnier, and Ryan (2006) concluded that, with all other variables held constant, for each 1-year increment in placement age, there was a 6% increased likelihood of disruption. Other researchers used age groups instead (e.g., Orsi, 2015); as an example, Selwyn et al. (2015) found that in their adopted children’s group, compared with the reference group of 0–1 year at placement, the breakdown hazard ratios (HRs) were 2.94 for placements 1–2 year, 6.16 for 2–4 years, and 13.45 for 4+ years, without any further differentiation within this last group. Since in their daily practice adoption professionals and agencies in Spain usually refer to age groups and child protection administrative files are also organized by age groups, we will take this approach in our study (see Method section).
The role of age at placement on the incidence of breakdown is frequently analyzed within the disrupted group, comparing the incidence among children placed younger and older, generally confirming the higher proportion of cases in the older group (e.g., Goerge et al., 1997). In our study, we have taken a step further, going outside the disrupted group and comparing the breakdown cases with the intact ones in terms of age at placement. Our research question is not only of all the breakdown cases, what is the proportion of children placed older compared to the proportion of those placed younger? In addition, our question is of all existing adoptions (i.e., all potential breakdown cases), what is the proportion of intact and disrupted cases for each age at placement group?
Our study took place in Andalusia, a region in Southern Spain that represents around 20% of the total country population, with a demography of adoption similar to the rest of the country. The public adoption agency in the region commissioned the study of adoption breakdown in the decade 2003–2012, whether domestic or international, including both prelegalized adoptions (PA) and court confirmed adoptions (CA; Palacios, Jiménez-Morago, & Paniagua, 2015; Paniagua, Jiménez-Morago, & Palacios, 2016). Since previous research on the incidence of adoption breakdown (Child Welfare Information Gateway, 2012; Donaldson Adoption Institute, 2004; Smith, 2014) has shown a different profile for each adoption stage (PA vs. CA), their trajectories (Kaplan–Meier survival analysis) and the factors associated with the duration of the placements (Cox regressions) will be considered separately for the analysis of placement duration.
Our first hypothesis in this study concerns the duration of adoptive placements ending in breakdown. We expect that, similar to what happens in the case of breakdown incidence, the duration of the placements will be associated with a configuration of factors (in the child, the adoptive family, and the professional interventions) more than with any isolated factor, although we expect age at placement to play a significant role among child-related factors. In the hypothesis for our second goal, we expect that age at placement will be significantly associated with the breakdown incidence, as in all the existing literature.
Method
Subjects
During the decade under consideration (2003–2012), there were 7,006 adoptive placements in the region, with 93 identified breakdown cases, a total breakdown rate of 1.3%. In 40% of the cases, the children involved were members of an adopted sibling group, with all siblings leaving the adoptive home in 57% of these cases. To reduce the burden for adoption professionals in data collection, it was decided that only information regarding the more problematic child in the sibling group (according to judgment of the professionals involved) would be collected. Very often, the narrative accounts in the files focused only on the sibling with more difficulties. All the analysis regarding the duration of placement and its associated factors refer to 33 PA and 36 CA cases, with 69 breakdown cases in total (representing 74% of all the disruptions in the region during the decade). For the analysis of the role of age at placement in the breakdown incidence, all 93 breakdown cases will be considered in the second part of the Results section.
Measurement and Procedure
All adoptions in Spain must go through a public agency responsible for placing children in families and, eventually, for moving them back to care in the case of breakdown. The public adoption agency makes all the administrative decisions and keeps the case records. When the preadoptive placement is considered satisfactory, the case is taken to court for legal completion. Adoption professionals always work in a team with a social worker and a psychologist. In the agency files, information is available on the adopted child’s birth parents, the adopted child, the adoptive parents and the family life after placement, and the professional interventions carried out with the adoptive parents and the adopted child. The files contain two types of information. There is information used for administrative purposes, with fields for variables such as the gender of the adopted child, whether the adoption was domestic or international, age at placement in the family, if the adoption is by a single parent or by two parents, and time between the placement and the court adoption decision. The other information is of a narrative nature and contains the caseworkers’ written account of events and interventions. In terms of data analysis, it is easy to retrieve information regarding the first set of variables, as they can be easily extracted from the data files. However, these files do not contain a field for the existence of breakdown; this information cannot be easily inferred from other administrative information and can only be identified from the caseworkers’ narrative accounts. In this study, we had to identify the breakdown cases through a request submitted to all adoption professionals in the region, analyzing thereafter both the administrative and the narrative accounts contained in the corresponding files.
A data collection document with four sections was devised: characteristics of the children (gender, age, domestic or international origin, single or sibling adoption, etc.), their birth parents (reasons for removal, sociodemographic profile, health, mental problems, professional interventions with them, etc.), the adopters and the placement process (motivation to adopt and expectations, mutual adaptation process, attachment and behavioral issues, family life, etc.), and the professional interventions (adoption preparation, suitability assessment, matching, postplacement support and interventions). Definitions and examples were provided regarding variables such as reasons for removal, mental health problems, attachment and behavioral problems, unrealistic expectations, and family life problems.
All the adoption caseworkers in Andalusia were sent the data collection instrument with instructions for completion. Their task was to identify all breakdown cases of which they had evidence in the period 2003–2012 and complete the required information. When the adoption professionals were unable to complete the data collection, child protection authorities approved members of the research team to extract the information. Researchers found that the files contained lots of bureaucratic information and were very often more limited in terms of substantive details. It was typical, for instance, that the information in the files referred to “attachment difficulties,” with a description of the problems but, typically, with no use of any diagnostic tool other than observations or interviews. The same was true for behavioral problems, detected by adoption professionals in their home visits or through interviews, usually with no further use of standardized methods. The implication is that we could only consider attachment difficulties or behavioral problems in terms of yes/no, not in terms of scores or more precise diagnostic labels, and always according to the caseworkers’ reports.
In general, the files contained more information on domestic than intercountry adoptions. All domestic adoptions involve a preadoption period during which there is a more intense scrutiny by the caseworkers (social workers and psychologists), who in all cases work for the public adoption agency. Once this period ends with the court decision formalizing the adoption, the contact with the family is much more limited. In most of the intercountry adoptions, the usually quite short preadoption period happens in the country of origin. Reports in the files tend to be much shorter and less detailed than in domestic adoptions. For these children, once in their new families, the majority of the postadoption follow-up is in the hands of professionals working for the accredited agency involved in the international adoption. Very often, their follow-up reports, both in terms of frequency and content, are mainly geared toward completing the limited information requested from the country of origin. In consequence, the information for CA tends to be less rich than in the case of PA, both for domestic and intercountry adoptions.
Our survival analysis is limited to the breakdown cases. The analysis will therefore not include censoring, as all the cases have experienced adoption breakdown. For our survival and regression analysis, with time since placement to breakdown as the dependent variable, we do not have a nondisruption group, as studying the narrative reports for the more than 7,000 adoption cases occurring during the decade under consideration was not logistically possible, and we only were authorized to study the files of breakdown cases. For the breakdown cases, Cox regressions will be used to compare the dependent variable “time to disruption” (i.e., speed of breakdown) in the two groups of interest (PA and CA), as well as the role of different independent variables over the duration of the placement, the first goal of this article.
For our second goal, the analysis of breakdown by age at placement, information regarding the child’s age at the time of breakdown was available through the administrative data for all cases. Once the breakdown cases had been identified, disrupted and intact cases during the observation period can be compared. In this comparison, using χ2 and rate ratio techniques, we will analyze the increment of breakdown risk associated with increasing age at placement.
The age of the child was analyzed by age groups, in accordance with how the Spanish adoption agencies categorize children: adopted as infants (0–2 years; 0–24 months), in their preschool years (2–6; 25–72 months), during the elementary school years (6–10; 73–120 months), or later on (+10 years; from 121 months).
The study was conducted with full respect of the adoption agency ethical guidelines for anonymity and confidentiality. The University ethics committee approved the study.
Data Analysis
Statistical analyses were undertaken with IBM SPSS Statistics 23.0 program. A descriptive analysis of the disruption cases was followed by a bivariate analysis using the Mann–Witney U test examining differences between PA and CA cases. Kaplan–Meier and Cox regression methods were used to examine the survival function and the variables associated with the duration of the placement. In these analyses, the outcome variable was the presence of breakdown and the time variable was the time in months from placement to disruption. Two criteria were used to include variables in the Cox regression: theoretical relevance and presence of information about the specific variable in at least 85% of the cases. The final model was developed through four initial models that focused on variables corresponding to birth family, adoptees, adopters and family life, and professional intervention. Statistically significant variables in each of the four models were included in the final model. All variables were categorical. The proportional hazards assumption was met by estimating the log-rank in the statistically significant variables in the final model. For the analysis of the risk associated with increasing age at placement, our second goal, χ2 and rate ratio analyses were used.
Results
Birth Family
Little information was available on the birth family, largely due to scant information in the case of intercountry adoptions. More information was available regarding the reason why the child was removed. Statistically significant differences were found only for sexual abuse (p < .05, ϕ = .28), with 16% of the breakdown cases in PA cases and none in CA ones (see Table 1). Neglect was the principal reason for removal in both PA (84%) and CA (93%) disruptions, followed by voluntary relinquishment (55% vs. 43%, respectively) and maltreatment (59% vs. 55%), but with nonsignificant differences (p values not reported for nonsignificant differences henceforth). It was very common for a child to be removed due to more than one of the previous reasons (Table 1).
Descriptive Analysis of the Disruption Cases Related to Birth Family, Adopted Children, and Family Life After Placement in PA and CA Breakdown Cases.
Note. PA = prelegalized adoptions; CA = court confirmed adoptions.
a Cohen’s d for mean age at placement and mean age at breakdown, Cramer’s V for timing of intervention and type of professional, and Cramer’s ϕ for all other variables.
Adopted Children
In the case of PA placements ending in breakdown, 53% were boys and 47% girls (Table 1). All of them were domestic adoptions. Prior to being placed for adoption, 97% of the children had lived in a residential unit and 25% in family foster care (a child could have lived in both). For the whole PA disruption group (all domestic adoptions), the average age at placement was 8 years and 11 months. At the time of the breakdown, the average age was 12 years and 5 months.
In CA cases with breakdown, 44% were boys and 56% girls. Domestic adoptees (69.5%) were overrepresented compared to intercountry adoptees (30.5%; p < .05, ϕ = .41). Prior to being placed for adoption, almost all the children had lived in a residential unit and 25% had also spent some time in a foster family before being placed in the new family for adoption. For the whole CA breakdown group, the mean age at placement was 6 years and 6 months (SD = 3 years and 5 months), 6 years and 4 months (SD = 3 years) in domestic adoptions, and 6 years and 9 months (SD = 4 years and 4 months) in intercountry adoptions, a nonstatistically significant difference. The mean age at adoption breakdown for the whole CA breakdown group was 14 years and 1 month (SD = 2 years and 9 months), 14 years and 11 months (SD = 1 year and 9 months) in domestic adoptions, and 12 years and 2 months (SD = 3 years and 9 months) in intercountry adoptions (p < .01; d = 1.14).
Family Life After Placement
In all the adoption breakdown cases, several problems in family life were detected since the beginning of the placement. In the adoptees, behavior problems were the most frequently reported (77% in PA and 82% in CA), followed by emotional problems (35% vs. 32%, respectively) and sexualized behavior (24% vs. 15%, respectively). Furthermore, according to the caseworkers’ judgment, some type of psychological disorder had been identified (32% vs. 44%), as well as attachment difficulties (58% vs. 62%). For the adoptive parents, statistically significant differences were found in the presence of unrealistic expectations, more frequent in PA (43%) than in CA (16%) disruptions (p < .05, ϕ = .31). Domestic violence was observed more often in CA cases (71%) than in PA ones (41%; p < .05, ϕ = .30). Lastly, as could be expected according to the adoption stage, significant differences were found in the timing of problems arising; this happened in the first months of the life together in 61% of the PA cases and in 40% of the CA ones (p < .01, ϕ = .21); in this latter group, it was more frequent that the problems manifested later on (Table 1).
Professional Intervention
A clear contrast was observed between the abundant problems in the adoptive family life and the lack of professional support, with no significant differences between our two comparison groups (the average values for both groups are reported in Table 1). No professional intervention after the placement was recorded in 31% of the cases. In 23% of the remaining families, the interventions consisted of routine follow-ups as required by the public adoption agency or the country of origin. In the other 46%, the interventions were a response to the manifestation of difficulties, with standardized assessment only in 27% of the cases. For these interventions, private mental health professionals (53%) were frequently involved, while only 18% of the families went to the public, specialized and free of charge postadoption service; in the remaining 29%, the interventions were performed by the adoption caseworkers or by professionals at the residential units where some children had been sent during the family crisis.
Placement Duration: Survival Analysis
In all the cases, the placement starts when the child moves to live with a family that is expected to become the adoptive family and ends when the child leaves the family due to breakdown. The Kaplan–Meier survival curves in Figure 1 show statistically significant differences when comparing the duration of placement for PA and CA breakdown groups, χ2(1) = 15.990, p < .001, ϕ = .48. The median for the time elapsed from placement until breakdown was 26 months in PA (95% CI [3.143, 48.857]) and 85 months in CA (95% CI [64.428, 105.572]). All disruptions together, the median was 57 months between placement and breakdown (95% CI [37.807, 76.193]).
Kaplan–Meier’s curves for prelegalized adoptions, court confirmed adoptions, and average.
Cox regressions were used to determine the variables that predicted the speed of breakdown (duration of placement) in the PA and CA groups. Initial Cox models were run first for each group of variables (characteristics of birth parents, adoptees, adoptive parents and family life, and professional interventions) and thereafter a final model was run with the significant variables from the initial models.
PA Disruptions
The initial models were significant for the variables related to the child, χ2(3) = 10.442, p = .015, V = .39, the adopters and family life, χ2(2) = 8.334, p < .05, V = .54, and the professional intervention, χ2(1) = 5.573, p < .05, ϕ = .40, but not to the birth family characteristics, χ2(4) = 1.210, p = .876, V = .11.
With the significant variables from each prior Cox model, a new regression was performed to test the final model. The contribution of each variable was significant, as was the overall model, χ2(5) = 22.765, p ≤ .001, V = .44. The inclusion of the significant variables from the initial models did not significantly improve the final one, χ2(1) = 2.750, p = .097, ϕ = .31. Table 2 shows the results for the final Cox model, with the main predictors of the duration of the placement. Age at placement was a significant risk factor, p < .05. None of the PA placed in the 0- to 2-year-old group broke down, so the reference group is 2–6 years. Compared to children placed at a younger age, those placed between the ages of 6 and 10 ended their placement 3 times faster, HR = 2.927, 95% CI [0.589, 14.543], p = .189, although the difference with the reference group 2–6 years does not reach statistical significance. Being adopted older than 10 years of age multiplied the speed of breakdown by 10, HR = 10.235, 95% CI [1.632, 64.188], p ≤ .001. Other risk factors for the duration of placement were unrealistic expectations in the adopters, HR = 5.516, 95% CI [1.932, 15.751], p ≤ .001, and attachment difficulties, HR = 3.172, 95% CI [1.277, 7.883], p < .05. Therapeutic treatment in the early postplacement period was a protective factor, HR = 0.215, 95% CI [0.082, 0.561], p < .01.
Cox Regression Final Model for Breakdown Speed in PA Cases.
Note. PA = prelegalized adoptions.
a R.C. = reference category.
CA Disruptions
Cox regression models were calculated for each of the variable groups. The models were significant for the variables related to the birth family, χ2(1) = 4.411, p < .05, ϕ = .39, the adoptee, χ2(4) = 39.320, p ≤ .001, V = .62, and the adopters and family processes, χ2(1) = 7.731, p < .01, ϕ = .48, while the variable related to professional intervention was not significant, χ2(1) = 1.596, p = .206, ϕ = .219. Variables with significant p values included age at placement, HR = 0.212, 95% CI [0.049, 0.921], domestic or intercountry adoption, HR = 0.179, 95% CI [0.060, 0.533], mental health problems in adopters (e.g., depression) during the life together, HR = 0.146, 95% CI [0.027, 0.790], and attachment difficulties, HR = 3.037, 95% CI [1.037, 8.895].
For this CA group, Table 3 shows the information from the final Cox model, where the contribution of each variable was significant, as was the overall model, χ2(4) = 39.320, p ≤ .001, V = .62. The inclusion of variables referring to the initial models did not significantly improve the final model, χ2(1) = 0.145, p = .703, ϕ = .07. The results indicate that there were only two main predictors of the placement duration: type of adoption (domestic vs. intercountry) and age at placement.
Cox Regression Final Model for Breakdown Speed in CA Cases.
Note. CA = court confirmed adoptions.
a R.C. = reference category is domestic adoption.
The speed of breakdown multiplied by 7.4 in intercountry compared to domestic adoptions, HR = 7.412, 95% CI [2.212, 24.838], p ≤ .001, indicating that domestic adoptions ending in breakdown lasted longer than the equivalent in intercountry adoptions. The Kaplan–Meier survival curves comparing these two groups showed statistically significant differences, χ2 (1) = 4.703, p < .005, ϕ = .37. All CA disruptions together, the median between placement and breakdown was 85 months (95% CI [64.428, 105.572]), but domestic placements lasted 108 months (95% CI [85.275, 130.725]), while intercountry placements lasted 54 months (95% CI [29.208, 78.792]).
In terms of age at placement, compared with placements in infancy, placements at an older age were significantly shorter, as attested by the HR values in Table 3: for 2–6 years placements, HR = 27.913, 95% CI [2.784, 279.865], p < .01; for 6–10 years placements, HR = 83.676, 95% CI [7.729, 905.958], p ≤ .001; for placements above 10 years of age, HR = 238.632, 95% CI [19.556, 2,911.923], p ≤ .001. The speed of the breakdown increased linearly with increased age at placement.
Breakdown Risk by Age at Placement
If the previous analysis indicated the factors associated with the duration of the placement, the following one shows the role of age at placement in the breakdown incidence (number of new cases over the 10-year observation period). Using the administrative data files from the adoption authority in the region, to which we added the identification of the breakdown cases, these were analyzed per age at placement group in the context of all the existing adoptions within the same age-group. Broken into four different age grouping sets, Table 4 shows, per age-group, the total number of children adopted in Andalusia during the period 2003–2012, followed by the number of breakdown cases for each group. There were no breakdown cases for PA children in the group 0–2 years at placement and the following analysis will be carried out for the whole breakdown group, PA and CA included. The distribution of total adoptions and breakdown cases per age-group (first three columns in the Table 4) is statistically significant, χ2(3) = 323.900, p < .001; V = .226.
Andalusian Adoptions and Breakdown Cases (2003–2012) According to Age at Placement.
Breakdown rate per 1,000 adoptions in the third column shows the proportion of all children adopted who experienced a breakdown in each age-group, with values indicating a linear increase in breakdown incidence with each increase in placement age. The rate ratio compares the incidence rate of any age group against the incidence rate of the reference group (0–2 years at placement) and informs about the risk of breakdown, with values greater than 1 indicating an increased risk. The rate ratio is obtained by dividing the incidence rate of each age-group (e.g., those placed at 2–6 years) by the incidence rate of the reference group (0–2 years). The rate ratio values in the last column of Table 4 show that, compared against the incidence of breakdown in the group placed in infancy, the risk of breakdown increased linearly with placement age.
Discussion
The two main goals for this article were the analysis of the factors associated with the duration of adoptive placements ending in breakdown and the analysis of the role of age at placement in the incidence of such breakdown. All the breakdown cases identified during a decade in a Spanish region were considered, including both those occurring before (PA) and after (CA) legal formalization at the court, as well as both domestic and intercountry adoptions. The results have supported the hypothesis according to which the child’s age at placement is a very relevant factor for both the duration of the placements ending in breakdown and the incidence of breakdown. While the second finding is commonplace in adoption breakdown research (Palacios et al., this issue), the first one adds a new piece to the complex puzzle of the breakdown experience.
Concerning the placement duration for PA breakdown cases, an older age at placement (child-related factor), unrealistic parental expectations, and difficulties in the attachment relationships between the child and the parents (parents- and family life–related factors) were found to be significantly associated with a shorter placement duration. Therapeutic intervention soon after the child’s placement (professional intervention–related factor) was found to be significantly associated with a longer duration of the placements. Confirming our first hypothesis, these results extend to placement duration previous findings about adoption breakdown incidence as reviewed by Palacios et al. (this issue): more than with any single factor, the speed of disruption relates to an accumulation of risks in the child, the adoptive parents and family life, and the professional intervention.
Information available in the adoption files is more limited for the CA cases, once the preadoption scrutiny in domestic PA ends with the formalization of the adoption at the court, and also due to more limited information for the intercountry adoptions. In the case of CA, the only two factors associated with the speed of the breakdown were the type of adoption (domestic/intercountry) and the age of the child at placement. It is important to consider that for Cox regressions to identify a variable as significantly associated to the outcome, the variable needs to be present with a certain amount of variability in the sample under consideration. This means that if the vast majority of the cases share a certain characteristic (e.g., behavioral problems), this will not be identified as making a difference among the studied cases. In this case, a matched sample of intact and disrupted cases (not possible in our case due to lack of relevant information for the intact group) would have probably provided a better approach. In our case, the variables domestic/intercountry and age at placement capture a variability that perhaps does not exist in some other aspects such as the already mentioned widespread behavioral problems, present in 80% of the cases in the period following the placement and in the remaining 20% cases during the transition to adolescence (Palacios et al., 2015; Paniagua et al., 2016).
In our results, the duration of the CA placements before breakdown was double in domestic (108 months) compared to intercountry adoptions (54 months). Even if the duration of the preadoption stage (normally, between 1 and 2 years in Spain) is considered, the duration of the placement after court formalization is considerably longer in domestic compared to intercountry adoptions.
Since parents in domestic adoptions had lived with the child for a significant period of time before the court decision, we interpret the longer duration of the placement to mean a stronger commitment with the child and the adoption. Intercountry adopters were adopting an unknown child, while domestic adopters were adopting a child who already was part of the family in the preadoption period. This does not imply that parents of intercountry adoptees were not committed to their children. In fact, in their case, the placement lasted an average of 54 months (more than 4 years), which means that they did not give up easily. But with placements lasting twice as much before breakdown, and with no less problems, parents of domestic adoptees showed a strengthened commitment.
The longer duration of domestic placements involving a child already in the family during a long preadoption period parallels the literature reported finding of a higher breakdown incidence in the case of children unknown to the adopters (Palacios et al., this issue). The implication is that, although intercountry adoptions break down less often, the duration of intercountry placements is significantly shorter compared to domestic placements.
Our results also support our second hypothesis, as there is a linear increase in breakdown incidence with increasing age at placement. As discussed by Palacios et al. (this issue), age at adoption may be considered a proxy for the accumulation of problems and adversities, so that the longer a child suffers maltreatment, toxic stress, and institutionalization, the more complex and difficult his or her psychological and behavioral profile will be, as shown in research reviewed in the introductory article such as Heim and Nemeroff (2001) and Turecki, Ota, Balangero, Jackowski, and Kaufman (2014).
While our results clearly show that placements involving older children last a shorter time and break down more frequently, it is important to emphasize that they are far from being condemned to failure. In fact, in the study by Palacios et al. (2015), 86% of PA and 98% of CA cases involving children placed at 6 years or older remained intact, and the same was true for 83% of PA and 96% of CA cases placed at 10 years or older. The main implication is that children should achieve adoption permanency as young as possible, not that the placement of older children should be avoided. They need to be better supported, so that the risk factors associated with an older age at placement are compensated with protective factors, as illustrated by the therapeutic interventions in the PA cases of our study.
This study shares some of the limitations typical in adoption breakdown research such as a problematic case identification and the difficulties in gaining access to sound information regarding breakdowns and the associated factors. These limitations were increased in our study by two undesirable, but unavoidable, facts: first, with all probability, our 1.3% incidence underrepresents the magnitude of the problem under study. Alongside the usual problems in identifying breakdown cases as discussed in Palacios et al. (this issue), the long period involved (10 years) in our study and the fact that breakdown is not contained in the existing adoption files add to the difficulty. Second, we only had access to the narrative case recording, and it was not possible to verify the accuracy or completeness. To give just an example, we coded attachment difficulties in a particular case not because we assessed the children and parents involved, but because those difficulties had been mentioned in the adoption files by a professional, usually without the use of specific diagnostic tools other than personal observation. In our conclusions to the adoption agency that commissioned the study, we urged them to be more systematic in keeping a formal record of the breakdown cases and to be more accurate in the description of children and families, if possible with the use of reliable assessment procedures.
The main practice implication of the findings reported in this article has been already mentioned: Any avoidable postponement of an adoptive placement places the child at higher risk for serious difficulties and is against the child’s best interest. In working with adoption agencies and prospective adopters, it is of critical importance that reliable assessment tools are used, and that a more solid adoption preparation is provided, so that, for instance, the expectations of the prospective adopters are better adjusted to the children. Even of greater importance is the quality and duration of the support offered to both children and parents once their journey together starts, including the postformalization period. In the process of analyzing the files of the cases in this study, the acute contrast between the many and serious problems experienced by children and parents and the few and feeble resources offered to them was quite disheartening. Particularly in the case of placements involving older children, both adopters and adoptees need (and deserve) much more, much earlier, and much better.
Footnotes
Authors’ Note
Carmen Paniagua worked with a predoctoral contract within the University of Seville Fifth Research Program (CONV-741, USE-12682-D).
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study reported in this article was commissioned and supported by the Department of Children and Families, Andalusian Government, under the contract #197-0446.
