Purpose This paper reports findings from two research studies that set out to calculate the rate and predictors of post-order adoption disruption in England and Wales. Methods All available national level administrative data on adopted children in England and Wales were analysed, supplemented by national surveys adoption managers. Complete national datasets were available 12 years in England and for 11 years in Wales. Results Of the 36,749 and 2317 adoptions considered, 565 in England and 35 in Wales had disrupted over the follow up period. Kaplan-Meier analyses indicate that cumulative post-order adoption disruption rates were 3.2% and 2.6% respectively for England and Wales. Cox regression models indicate that being older than four years adoptive placement, adoptive parents taking longer than a year to legalise the adoption, being a teenager and previous multiple placement in care were risk factors for post-order adoption disruption. Conclusion The post order adoption disruption rate is low. Implication for policy and practice are discussed.

Figures - uploaded by Dinithi Wijedasa

Author content

All figure content in this area was uploaded by Dinithi Wijedasa

Content may be subject to copyright.

ResearchGate Logo

Discover the world's research

  • 20+ million members
  • 135+ million publications
  • 700k+ research projects

Join for free

Wijedasa, D., & Selwyn, J. (2017). Examining rates and risk factors for post-

order adoption disruption in England and Wales through survival analyses.

Children and Youth Services Review, 83, 179-189. DOI:

https://doi.org/10.1016/j.childyouth.2017.10.005

License (if available):

CC BY-NC-ND

Link to published version (if available):

https://doi.org/10.1016/j.childyouth.2017.10.005

Link to publication record in Explore Bristol Research

PDF-document

University of Bristol - Explore Bristol Research

General rights

This document is made available in accordance with publisher policies. Please cite only the published

version using the reference above. Full terms of use are available:

http://www.bristol.ac.uk/pure/about/ebr-terms.html

Examining rates and risk factors for post-order adoption disruption

in England and Wales through survival analyses

Dinithi Wijedasa and Julie Selwyn, University of Bristol, UK

Reference:

Wijedasa, D. and Selwyn, J. (2017) 'Examining rates and risk factors for post-order adoption disruption in

England and Wales through survival analyses', Children and Youth Services Review, 83(Supplement C), pp.

179-189.

1. Abstract

Purpose: This paper report findings from two research studies that set out to calculate the rate and

predictors of post-order adoption disruption in England and Wales.

Methods: All available national level administrative data on adopted children in England and Wales

were analysed, supplemented by national surveys adoption managers. Complete national datasets

were available for 12 years in England and for 11 years in Wales.

Results: Of the 36,749 and 2,317 adoptions considered, 565 in England and 35 in Wales had disrupted

over the follow up period. Kaplan Meier analyses indicate that cumulative post-order adoption

disruption rates were 3.2% and 2.6% respectively for England and Wales. Cox regression models

indicate that being older than four years adoptive placement, adoptive parents taking longer than a

year to legalise the adoption, being a teenager and previous multiple placements in care were risk

factors for post-order adoption disruption.

Conclusion: The post order adoption disruption rate is low. Implications for policy and practice are

discussed.

2. Introduction

At any given time, some 70,000 children in England and around 6,000 children in Wales are looked

after by local authorities (Department for Education, 2016; Welsh Government, 2016a). The reasons

for entry into care are varied, but more than half the children (54% in England and 64% in Wales) are

taken into state care due to abuse or neglect (Department for Education, 2016; Welsh Government,

2016b). Other reasons for entering care include family dysfunction, acute stress in families and absent

parenting, with only a small number entering care due to disability and illness of children or parents.

Children who are returned to their families after a period of being looked after constitute the largest

proportion of exits from care in England and Wales. The most recent statistics indicate that 34% of

the 31,710 children in England and 37% of the 2,020 children in Wales who ceased to be looked after

were reunified with their families (Department for Education, 2016; Welsh Government, 2016c).

However, previous research (Farmer et al., 2011; Wade et al., 2011; Farmer and Lutman, 2012;

NSPCC, 2012) show that a large proportion of reunifications are unsuccessful and that around 40%

of children re-enter care. Looked after children who are unable to return home safely require another

permanent arrangement to be made within the state care system, and these children may be placed

with relatives, long-term foster carers or adoptive parents. These placements can be legally secured

through special guardianship orders, child arrangement orders (previously referred to as residence

orders) or adoption orders.

Around 6% of looked after children in England and Wales are adopted each year. Unlike in the US,

where the majority (52%) of children are adopted by their foster carers (U.S. Department of Health

and Human Services, 2016), the majority (85%) of children in England and Wales are adopted by

'stranger' adopters (Department for Education, 2016 ; Welsh Government, 2016c). Around 70% of

adoptees in England and Wales are taken into care due to adverse experiences such as abuse and

neglect (Department for Education, 2016; Welsh Government, 2016c). Although experiences of

maltreatment such as abuse or neglect can pose a risk to children's wellbeing and emotional and

behavioural outcomes (Nanni et al., 2012; Smith, 2013), research indicate that previously maltreated

children are able to make significant developmental gains in growth, attachment and cognitive

capacities once adopted (Triseliotis, 2002; Van IJzendoorn and Juffer, 2006; Lloyd and Barth, 2011;

Fisher, 2015; Sonuga-Barke et al., 2017). However, for some children, the effects of maltreatment

can be lifelong (Shonkoff et al., 2012; Hughes et al., 2016) and despite the documented gains,

research show that adoption per se is not able to attenuate the impact of all previous adverse

experiences (Smith, 2013). For example, compared with children growing up in traditional family

settings, adoptees report higher rates of psychopathology and lower self-esteem (Burns et al., 2004;

Gagnon-Oosterwaal et al., 2012; Fisher, 2015; Sánchez-Sandoval, 2015), which may increase their

vulnerability to adoption disruption.

1.1. Adoption disruption in England and Wales

Little is known about adoption disruption in England and Wales. Legally, an adoption cannot be

dissolved or revoked in England and Wales except under very exceptional circumstances (see for

example, PK v Mr & Mrs K [2015] EWHC 2316 (Fam)), and can only be reversed through making

of another adoption order. A mapping review (Grant and Booth, 2009) revealed a few studies in the

UK which had estimated the post-order adoption disruption rate to be between 3-7% using survey

methodologies (Lowe et al., 1999; Selwyn et al., 2006; Ra ndall, 2013) (see Table 1). However, no

national-level studies have been conducted in the UK to calculate the rate of post-order adoption

disruption. This is mainly because: local authorities terminate collection of new data on adopted

children as children cease to be looked after the adoption order and; in the instances where children

come back into care after an adoption order, they are given new identification numbers, which are

not linked to their previous identities and care histories. Gathering accurate data is further made

difficult because adopted children are usually given a new surname, a new National Health Service

(NHS) number and a new education pupil number (UPN), therefore severing the links between pre-

adoption care files and post order disruption records. There are also no mechanisms to track children

as they move with their adoptive families around the country or emigrate. If children re-enter care in

a different local authority area, social workers are dependent on adoptive parents informing them that

the child is adopted.

From 1997, the English government began a reform programme to increase the number of children

adopted from care and to speed up decision-making by introducing timescales and national standards.

These reforms were critiqued by many academics (Kirton, 2013) concerned at the politicisation of

adoption practice and the failure to consider the impact of austerity policies and reduced services for

the poorest families (Featherstone and Bywaters, 2014). Concerns were also raised that these reforms

would lead to higher adoption disruption rates. Although similar reforms in the US promoting the use

of adoption and speeding up the process did not produce higher disruption rat es (Smith et al., 2006;

Rolock and White, 2016), there were no national figures on post-order adoption disruptions available

for England and Wales. In this article, we report on findings from studies that set out to provide the

evidence on post-order adoption rates for England and Wales and to explore the risk factors that

contributed to adoption disruption.

Table 1. UK studies of adoption disruption rates (1990-2015)*

* Adapted from authors' own publication (Selwyn et al., 2014)

No differentiation made between pre/post order adoption disruptions

Pre/post Order

disruption rate

Post-Order

disruption rate

1,165 children with special needs. Adoptions made by 24

Voluntary Adoption Agencies between 1980-1984.

129 children with a permanence plan in one local authority

between 1986-1990

61 families with 130 late placed children (placed when 5-9

years)

72% of adoption agencies in 1994. 138 disruptions reported

210 Ethnic minority children placed for adoption by a VAA

72 families with 133 children

97 older children (4-12years) placed for adoption between

1991-1996 from one Local Authority

(Rushton and Dance,

2006)

99 children 5-11 years old at placement

7.6 years since

entry to care

328 children placed by one Voluntary Adoption Agency

between 2001-2011

22 children adopted by non- relatives

3. Methodology

A retrospective longitudinal research design was used to consider the rates of and reasons for post-

order adoption disruptions in England and Wales. The studies were commissioned by the Department

for Education in England (Selwyn et al., 2014) and by the Welsh Government (Wijedasa and Selwyn,

2014) and used similar methodologies. Ethical clearances for the studies were obtained from the

research ethics committee at School for Policy Studies, University of Bristol. Adoptees were defined

as previously looked after children adopted from care. Inter-country adoptions and step-parent

adoptions were excluded. Adoption disruption was defined as when a previously looked after child,

with an adoption order, ceased to live with the adoptive parents before the age of 18 years.

2.1. Sample

All children who were adopted in England between 1st April 2000 and 31st March 2011 and between

1st April 2002 and 1st April 2012 in Wales were included in the analyses. Earlier data were not used

due to data being unavailable on the total population of children in care.

2.2 Data

2.2.1 National level administrative data on adopted and looked after children

Once a year, local authorities in England are required to submit data on all children looked after and

those adopted to the Department for Education (DfE) and in Wales to the Welsh Government. The

information collected includes care histories, information on care leavers and children adopted from

care. From the national data, two longitudinal databases were provided by the Department for

Education and the Welsh Government. In each database, every child has a unique identifying number,

which enables record linkage.

I. Longitudinal database on all children adopted in England and Wales: These databases

contained the details of all the children in care who had been the subject of an adoption order

between April 2000 and March 2011 in England (n= 35,355) and between April 2002 and

March 2012 in Wales (n=2,352). Each row in the database contained data on an individual child

and included variables such as the name of the local authority, child's gender; whether adopted

by foster carers, date of the adoption decision, date of match with adopters, date of placement

with adopters; and date of adoption order.

II. The longitudinal database on all children looked after in England and Wales: These

databases contained the details of all children who were in care between April 2003 and March

2011 in England and between April 2002 and March 2012 in Wales. Asylum seeking children

and children on a series of agreed short breaks in the care system were excluded from our

analyses. The database enabled the longitudinal tracking of the care histories of adopted

children and had variables such as date of entry to care; reason for entry to care; type of legal

status; type of placement; dates of placement changes and; dates of legal status changes.

It was not possible to identify sibling groups within the administrative datasets.

Using these datasets, it was possible to identify and establish some of the characteristics of all the

adopted children. It was also possible to explore the children's care careers from the time of entry to

care by linking the two files. Although no data on the adoptees were available on the administrative

datasets after the legal order, it is likely that all families who have experienced an adoption disruption

contacted the local authority that placed the children. Therefore, national surveys of adoption team

managers and surveys of adoptive parents were undertaken to identify children who had experienced

an adoption disruption.

2.2.2. National surveys of adoption managers in England and Wales

National postal/telephone surveys of all local authority adoption managers were conducted in England

(n=148) and with managers in Wales (n=22) achieving a response rate of 86% in England and 92% in

Wales. Every manager was asked to provide data on previously looked after children who had been

adopted within the study time frames and who had subsequently experienced an adoption disruption.

The following information was requested: the unique local authority ID number of the child before the

adoption order, date of birth, gender, date of placement, date of adoption order and date of adoption

disruption. Data from the surveys were then merged with the administrative data on adopted children.

These surveys provided information on 505 children in England and 31 children in Wales who had

experienced an adoption disruption.

2.2.3. Supplementary surveys

Some of the adoption managers we spoke to were concerned that the rate of adoption disruption might

be underestimated, as some children were placed outside the boundary of their local authority and there

was no requirement for the receiving adoption teams to notify the placing local authority of adoption

disruptions. These concerns were addressed by undertaking (1) a survey of all voluntary adoption

agencies (VAA) in England (n=22) and Wales (n=2), with a 55% and 100% response rates

respectively, (2) a survey of 14 local authority adoption teams who did not place many children for

adoption, but had many children placed within their local authorities by other adoption teams (3) a

survey of adoptive parents (n=620) who had legally adopted 880 children from 13 local authorities in

England between April 1st 2002 and March 31st 2004 and (4) the survey was also published on

AdoptionUK, which is an online forum for adoptive parents, which resulted in 180 families who had

legally adopted 310 children responding to the survey.

The survey of voluntary agencies and the 14 local authorities provided details of an additional 20

children in England who had not been identified through the national adoption managers' survey. The

survey of adoptive parents in 13 local authorities and the responses from the AdoptionUK forum

members did not lead to any new information on disrupted adoptions.

2.2.4. Additional data on disruptions from the administrative data files

Although local authorities are required legally to provide adopted children with a new identification

number (ID) if they re-enter care, an exploration of the new merged datasets revealed that this

procedure had not been correctly followed as some adopted children's ID had remained the same.

Forty children in England and four in Wales were identified in this manner as being adoption

disruptions.

Amalgamation of data from all the above sources indicated that there had been in total 565 adoption

disruptions in England between 2000-2011 and 35 adoption disruptions in Wales between 2002-2012.

2.3. Data analyses

The aims of the statistical analyses were threefold, (1) to establish whether adoptions disrupted after

the order (the rate of disruption), (2) to establish when children were most at risk of disruption (timing

of disruption), and (3) to establish who was at a greater risk of experiencing an adoption disruption

(predictors of disruption).

As can be seen in Figure 1, the children were at different stages of their adoption journeys: (1) children

who had experienced a disruption (2) children who had reached the age of 18 before the end of the

study and had not experienced a disruption and (3) those who were not yet 18 and had not experienced

a disruption. Data from the third group of children were treated as incomplete or censored, as although

they had not experienced a disruption by the end of the follow-up period, it is possible that some may

have gone on to experience a disruption at some later point. Therefore, data on whether the disruption

had occurred or not, was not available for all children.

Figure 1 The timelines of adopted children at the beginning and end of data collection

Adapted from (Yamaguchi, 1991)

It was not possible to meet the aims of the analyses with proportions, risk/odds ratios or logistic

regressions as these statistical methods assume complete data for all children (whether they had a

disruption or not) and ignores time to event (when a disruption was most likely to occur). Ordinary

multiple regression, on the other hand, would have taken the time of disruption into account as an

outcome variable, but not whether the event occurred or not.

2.3.1. Survival analyses

When compared with the methods mentioned in the preceding section, statistical methods known as

survival or event-history analyses (Singer and Willet, 2003) take into consideration both the event

(whether the event occurred) and the time to event ( when the event occurred). Survival analyses has

the benefits of also using all data, including that from the censored observations. Therefore, in this

study, survival analyses methods known as Kaplan-Meir and Cox Proportional Hazards were used to

calculate the post-order adoption disruption rate, the times of higher risk and to establish the predictors

of adoption disruption.

Kaplan-Meier analyses were conducted to estimate the overall post-order adoption disruption rates in

each country over the follow-up periods and to estimate the risk of adoption disruption over time.

Although a Kaplan-Meier analyses allows exploration of variation in disruption rates for different

levels of a categorical variable such as gender, they cannot adjust for several predictor variables

simultaneously. Therefore survival analyses methods known as Cox Proportional Hazards models,

were utilized to investigate the risk factors for adoption disruption (Yamaguchi, 1991; Singer and

Willet, 2003; Machin et al., 2006).

Univariate analyses were conducted, with a Bonferroni correction ( p= 0.05/15= .003) for multiple

comparisons, to explore differences between the intact and disrupted groups. Then, all variables to be

included in the model were tested to establish whether the proportionality assumption was met.

Variables, which showed a skewed relationship with disruption were recoded as categorical variables

before entering in the final Cox regression models (age since adoption order/ age at adoptive

placement/ and time between adoptive placement and adoption order as seen in Table 6 and Table 7).

Most often, variables included in regression models are fixed in value. For example, age at placement

would remain 'fixed' irrespective of the follow-up period. Cox regression models on the other hand,

allow inclusion of 'time varying' covariates as well as ' fixed ' covariates. In this analysis, we wanted

to explore whether children's age per se, such as being a teenager, had any influence on adoption

disruptions. Age 'now' was included as a ' time varying ' covariate in the Cox regression models.

4. Results

All adoptions from the care system in England between 1st April 2000 and 31st March 2011 and in

Wales between 1st April 2002 and 31st March 2012 were considered in the analyses. Start of the follow-

up period was defined as the date of the adoption order when a child ceased to be looked after. The

end date of the follow-up period was either the date of disruption or August 2011 for the England

adoptees and August 2012 for the Welsh adoptees. By the end of the follow-up period, 565 out of

37,335 adoptions in England and 35 out of 2,352 adoptions in Wales had disrupted.

The average age at disruption was 13 years in England (M= 12.7, SD=3.2, CI=10.3-12.0) and 12 years

in Wales (M= 11.6, SD=3.9, CI=12.4-13.0). The time to disruption from the date of the adoption order

was on average 5 years in England (M= 5.4, SD=2.9, CI=5.2-5.7) and Wales (M= 4.6, SD=2.4, CI=3.8-

5.4). Univariate analyses were conducted to explore the variables in the datasets and the derived

variables to test for significant statistical differences between the children who were in intact adoptive

placements at the end of the follow-up period and those children who had experienced a disruption.

3.1. Differences between children in intact adoptive placements and children who

had experienced post-order adoption disruptions in England.

As can be seen in Table 2, compared with children who were in intact adoptive placements, children

in England who had experienced an adoption disruption were significantly more likely to be older at

every stage of the adoption process, including at entry to care, at the time of the adoption decision,

match with adopters, adoptive placement and at the time of the adoption order. Children who

experienced disrupted adoptions also had significantly longer waiting times between each of the

adoption milestones and were more likely to take longer to go through the adoption process. On

average, time from entry to care to the adoption order was 2.9 years for children who were in intact

placements, whilst the children who were in the disrupted group waited 4.2 years in the care system

from the time of entry until the adoption order.

Although the univariate analyses found that foster care adoptions were not a protective factor and that

children who were adopted by their foster carers were more likely to experience an adoption disruption,

the strength of this association reduced to non-significance when considered alongside other variables

in a Cox regression model (See Table 6).

There were no statistically significant differences between the disrupted and intact groups in terms of

gender, ethnicity and the reason for entry to care.

Table 2 Characteristics of the children adopted from care in England

Children in

intact

adoptions

(N=36,749)

Children who

experienced

an adoption

disruption

(N=565)

Difference

between the intact

and disrupted

groups* and effect

size

t (576.4) = -21.02,

p <.001, r = .66

t (37312) = -24.46,

p <.001, r = .13

Age at match with adopters

t (37312) = -24.46,

p <.001, r = .13

Age at adoptive placement

t (37221) = -24.77,

p <.001, r = .13

t (37312) = -26.22,

p <.001, r = .13

Time from entry to care to adoption decision

t (577.5) = -10.88,

p <.001, r = .41

Time from adoption decision to match with

adopters

t (576.0) = -5.98,

p <.001, r = .24

Time from match with adopters to adoptive

placement

t (567.5) = -3.15,

p <.001, r = .13

Time from adoptive placement to adoption

order

t (570.4) = -9.63,

p <.001, r = .37

Total time from entry to care to adoption

order

t (579.3) = -17.3,

p <.001, r = .58

Ethnicity

White

Mixed

Black

Asian

Other ethnicity

Main reason for entry to care

Abuse or neglect

Family dysfunction

Family in acute stress

Absent parenting

Parental illness or disability

Child's disability

Socially unacceptable behaviour

Low income

Child adopted by foster carer

Yes

No

2 (1)=13.22,

p<.001, OR= 1.2

ᶧ Data missing for 21 of the 37,770 intact adoptions.

*p value adjusted with Bonferroni correction at p<.003 for multiple comparisons.

∞ The percentages might not add up to 100% due to rounding

† Chi-Square tests not statistically significant at p< .003 level.

Note. Although pertinent, the number of moves a child had had in the care system could not be explored in depth, as this

data were available for only 50% of children in the disrupted group and 72% in the intact group. Adopter characteristics

such as their gender and marital status could also not be explored, as t he data were unavailable for many children due

to the data being collected only from 2006.

3.2. Differences between children in intact adoptive placements and children who

had experienced post-order adoption disruptions in Wales

As seen with the adopted in England, when compared with those in intact adoptive placements, the

children who experienced a disruption were significantly more likely to be older at entry to care and

at every point through the adoption process (See Table 3). The total waiting time between entry to care

and adoption order was 2.7 years for the intact group and 3.8 years for the disrupted group, similar to

the results from the England. There were no statistically significant results between the intact and

disrupted groups in terms of the children's gender, ethnicity, reason for entry to care or whether they

were adopted by their previous foster carers.

Table 3 Characteristics of the children adopted from care in Wales

Children in

intact

adoptions

(N=2,317)

Children who

experienced

an adoption

disruption

(N=35)

Difference

between the intact

and disrupted

groups* and effect

size

t (34.6)= -4.56,

p <.001, r = .61

t (2330)= -4.34,

p <.001, r = .09

Age at match with adopters

t (2332)= -5.40,

p <.001, r = .11

Age at adoptive placement

t (2349)= -5.34,

p <.001, r = .11

t (2350)= -6.42,

p <.001, r = .13

Time from entry to care to adoption decision

Time from adoption decision to match with

adopters

Time from match with adopters to adoptive

placement

Time from adoptive placement to adoption

order

t (33.7)= -3.73,

p <.001, r = .54

Total time from entry to care to adoption order

t (2350)= -4.68,

p <.001, r = .10

Number of moves in care before adoptive

placement

t (2170)= -3.70,

p <.001, r = .08

Ethnicity

White

Mixed

Other ethnicity

Black

Asian

Main reason for entry to care

Abuse or neglect

Family in acute stress

Family dysfunction

Absent parenting

Parental illness or disability

Child's disability

Socially unacceptable behaviour

Other reason

Child adopted by foster carer

Yes

No

*p value adjusted with Bonferroni correction at p<.003 for multiple comparisons.

The percentages might not add up to 100% due to rounding.

† Chi-Square tests not statistically significant at p< .003 level.

Note. Adopter characteristics such as their gender and marital status could not be explored as the data were not available

for many children due this data being collected only since 2006.

3.3. Post-order adoption disruption rates in England and Wales

The rates of post-order adoption disruptions were calculated with Kaplan-Meier survival curves.

Follow-up data were available for a maximum of 12.3 years for England and 11.3 years in Wales. As

can be seen in Figure 2 and Figure 3, the cumulative proportion of adoption disruptions increased over

the years since the making of an adoption order.

The post-order adoption disruption rates were similar across England and Wales. The cumulative

adoption disruption rate in England was 3.2% over a period of 12.3 years, whilst in Wales over a period

of 11.3 years, post-order disruption rate was 2.6%. The estimated risk disruption for each year after

the adoption order are given in Table 4 and Table 5.

Figure 2. Kaplan-Meier plot of the predicted proportion of adoption disruptions for all children adopted in England between

2000 and 2011 (1-Survival curve plotted), (N= 36,749)

Figure 3. Kaplan-Meier plot of the predicted proportion of adoption disruptions for all children adopted in Wales between 2002

and 2012 (1-Survival curve plotted), (N=2,352)

Table 4. Estimated proportions of post-order adoption disruptions in England

Time in years

since adoption

order

Kaplan-Meier estimates of

cumulative proportion of

adoption disruptions over time

(%)

Estimated proportion of

adoption disruptions

within the year

(%)

Risk of disruption

per 1,000 children

within the year

Table 5. Estimated proportions of post-order adoption disruptions in Wales*

Time in years

since adoption

order*

Kaplan-Meier estimates of

cumulative proportion of

adoption disruptions over time

(%)

Estimated proportion of

adoption disruptions

within the year

(%)

Risk of disruption

per 1,000 children

within the year

*Estimates are not provided for the period between 8 and 11.3 years as the Kaplan-Meier estimates plateaued at 2.6%

3.4. Risk factors for adoption disruption in England and Wales

Before conducting Cox Proportional Hazard analyses with the statistically significant variables in the

univariate analyses (as seen in Table 2 and Table 3), checks for multicollinearity between the variables

indicated that all variables pertaining to age and most 'time' variables were highly correlated, therefore

only some of the 'age' and 'time' variables were included in the models. Age was included as a time

varying covariate, which considered how children grew older over time. Although we could include

the number of moves in the care system in the analyses of the Welsh data, the same variable was not

included in the England data analyses due to the data not being available for most of the children. The

variables included and the resulting parameters of the Cox regression model for predicting adoption

disruption are given in Table 6 and Table 7.

3.4.1. Risk factors for adoption disruption in England

As can be seen in table 6, controlling for all other variables, children's age since order, ag e at adoptive

placement and time between adoptive placement and order were all significant predictors of post order

adoption disruption in England.

Table 6. Hazard ratios for post-order adoption disruption for children adopted in England between 2000-2011#

Age at adoptive placement

Time between adoptive placement

and order

Total time between entry to care

and adoptive placement

Whether the child was adopted by

the foster carer

# Other variables, which were significant in the univariate analyses were not included due to multicollinearity.

* Reference category.

** Hazard ratio represents the incremental increase in risk of adoption disruption in one category, relative to the

reference category.

The categories do not overlap.

As can be seen by the Wald statistics, the most influential contributor to the model was the child's age

since the order. Teenage years (11-18 years) posed the highest period of risk, with the risk of disruption

being 10 times more than for a child who was under 4 years of age. Children who were older than 16

years were also nearly 5 times more likely to experience a disruption compared to children who were

younger than 4 years of age.

The second most influential contributor to the model was the child's age at placement for adoption.

Compared with those who were placed for adoption as infants (0-1 years of age), children who were

older than 4 years at placement were 13 times more likely to experience an adoption disruption, whilst

children who were placed between 2-4years of age were 6 times more likely to experience a disruption

after the order.

The length of time between the adoptive placement and the legal order was also a significant predictor.

Children who had to wait more than two years for the legal order were one and a half times more likely

to experience a disruption compared with children whose adoptive placement was converted to a legal

order within a year of placement with adoptive parents.

Controlling for all other variables, the influences of whether the child was adopted by a former foster

carer and the total time taken in the care system from entry to care to the time of the adoptive placement

were not statistically significant.

3.4.2. Risk factors for adoption disruption in Wales

The risk factors for adoption disruption in Wales are given in Table 7. Age at adoptive placement,

period between the adoptive placement and legal order and the number of moves a child has had in the

care system before being placed for adoption were all significant predictors of post-order adoption

disruptions in Wales. As seen by the Wald statistics, all three variables contributed to the model in a

similar manner.

Table 7. Hazard ratios for post-order adoption disruption for children adopted in Wales between 2002-2012#

Age at adoptive placement

Time from adoptive placement to

order

Number of moves before adoptive

placement

Total time between entry to care

and adoptive placement

# Other variables, which were significant in the univariate analyses were not included due to multicollinearity.

* Reference category.

** Hazard ratio represents the incremental increase in risk of adoption disruption in one category, relative to the

reference category.

The categories do not overlap.

Children in Wales who were older than four years of age at the time of being placed with adoptive

parents were around three times more likely to leave their adoptive home after the legal order,

compared with their contemporaries who were placed when they were younger than 4 years. The risk

of disruption also increased two-fold when the legal order was made more than a year after the child

was placed for adoption. Furthermore, each move a child had in care before being placed for adoption,

increased the likelihood of an adoption disruption by nearly threefold.

5. Discussion

The aim of this paper was to present national rates and the risk factors for post-order adoption

disruption in England and Wales. Having access to more than a decade of national data and the very

high response rates to the national adoption manager surveys gives validity and greater reliability to

the findings. A further strength of the analyses was the use of survival analyses methods, a rarely used

statistical method in UK social work research. More commonly used statistical methods in UK social

work research such as regression or ANOVA are able to explore differences between groups and

'whether' an outcome of interest has occurred (disruption/no disruption) or 'when' it is most likely to

occur (time to disruption), they are not able to answer the 'whether' and 'when' questions

simultaneously (Singer and Willet, 2003). As social work researchers, we are most often interested in

whether an event occurred as well as when it is most likely to occur. Here, survival analyses methods

were used to establish whether and when adopted children were most likely to experience an adoption

disruption and explore what child and family characteristics predicted post-order adoption disruptions.

Utilisation of survival analyses provided information not only on the national levels of post-order

adoption disruption rates, but also added new knowledge on the risky periods and the characteristics

of children most at risk of disruption.

The Kaplan-Meier survival analyses results indicated that the incidence of post-order adoption

disruption was low. It could be argued that disruption rates might be an under-estimation due to some

adoption managers being unaware of all disruptions. However, the very high response rate to the

adoption managers' survey, disruption rates being similar across both England and Wales and previous

research showing similar post-order disruption rates in England (3.7%) (Randall, 2013) and in the USA

(3%) (McDonald et al., 2001; Festinger, 2002) increase confidence in the estimates .

The research evidence is consistent on factors that are associated with disruptions. These include the

child's age at placement; a history of previous breakdowns; maltreatment; continuing negative

influence of the birth parents; and children's behaviour difficulties (Rushton, 2003; Evan B Donaldson

Adoption Institute, 2004; Coakley and Berrick, 2008; Faulkner et al., 2016; White, 2016). It should be

noted that several of these variables, which have been shown to be linked with adoption disruption

could not be included in the Cox proportional hazards models due to the data not being collected at a

national level in England and Wales (e.g. emotional behavioural difficulties of child, presence of birth

or other children in the adoptive family home (Rushton, 2004; Coakley and Berrick, 2008). Data were

also not available on agency variables such as the quality of the matching between the adopters and

the child (McGinnis et al., 2009; Dance et al., 2010; Quinton, 2012) ; whether adopters were provided

with adequate and accurate information about the child's past and whether the adopters received

support, all of which have been linked with stability of adoptions (Barth and Miller, 2000; Rycus et

al., 2006).

However, results from the Cox regressions provided new factors associated with adoption disruption:

a) being a teenager and b) when there was a time lag of more than a year between placement and the

adoption order. Both factors need greater attention in social work practice. Administrative data does

not provide answers to why there was delay between the placement of the child and the legalisation,

but it is possible that this delay reflects parents' uncertainties about their relationship with the child.

Social workers need to be aware that delays between placement and order might be a sign that

relationships are in difficulty and that families who delay applying for the Adoption Order are likely

to need more intervention.

The child being a teenager was the greatest risk factor. Previous studies of maltreated children who

were adopted indicate that about a third are described as troubled and unhappy during early

adolescence (Rushton, 2004; Selwyn et al., 2006) and that more stable adoptions result when

appropriate and sufficient adoption support is provided to the families (Smith, 2014). As most children

are younger than four years at the time of adoption in the UK, the adolescent years are mostly ignored

in the development of adoption support services, with most of the effort going into supporting adoptive

placements in the first few years. Therefore, considering the group of children who experienced

disruptions in this study, it is likely that the families or the teens had few services available to them.

However, this lack of services is beginning to change with the introduction of the government funded

Adoption Support Fund. The fund, which has made available since 2015, enables adoptive parents to

access therapeutic services for their child up to and including the age of 21 (or 25 with a Statement of

Special Educational Needs or Education Health & Care Plan)

(http://www.adoptionsupportfund.co.uk/). In its first year, of the 3,765 children helped by the fund,

42% were teenagers, which indicates that the services are becoming more responsive to the needs of

adopted teenagers. Although a recent evaluation shows that access to the adoption support fund has

been positively received by adoptive families, evidence from parents suggest that most families still

reach a crisis point before seeking help and support (King et al., 2017).

Another significant predictor of post-order disruption was children's age at placement for adoption.

The analyses indicated that children would have a better chance of adoption stability if they had been

placed with their adopters before four years of age. Although the children who were older than 4 years

at placement were not further subdivided in the Cox regression due to small numbers of older (11+

years) children being placed (n=221), further exploration of the older group revealed that the risk of

disruption increases between 4 and 11 years of age and then decreased between 11 and 16 years of

age. This latter group had a similar level of disruptions as those who were placed when they were 0-4

years. This maybe because older children might have been more involved in the adoption decision

making process, whereas it is unlikely that the younger children would have been included in the

decision-making process. For children who are unable to return home, it is important that the

permanence decisions are made quickly. It is also important to ensure that children's right to be heard

is respected. Speedy decision making would also minimise multiple placement in care, which was

another strong predictor of adoption disruptions. The 26-week limit on care proceedings introduced in

April 2014 and the ongoing adoption reforms that have encouraged speedy decision making on

permanence for children in care may be having an impact. The latest statistics show that the percentage

of children who were younger than 4 years at adoption has been increasing steadily from 73% in 2011

to 79 % in 2016 (Department for Education, 2016).

6. Conclusion

The results indicate that adoptions are successful for most children. The majority of the adoptees

considered in the two retrospective longitudinal studies of adoptees in England and Wales did not

experience an adoption disruption. The results of the Cox proportional hazards models indicate that

there are areas that could be improved such as: avoidance of delay and facilitation of early permanence

decisions; acknowledgement of the needs of families with adoptive teens and; ensuring that adoptive

placements are monitored for delay in legalisation, which could be indicative of parents' underlying

concerns about the adoption and a need for more intervention and support.

Funding

The study on adoption disruptions in England was funded by the Department for Education and the

study in Wales by the Welsh Government.

Acknowledgements

The authors would like to acknowledge the support of all local authority and voluntary agency adoption

managers and the adoptive parents who provided data for the research. Authors also thank Paul Basset

from Stat Consultancy for statistical advice.

References

Barth, R.P. and Miller, J.M. (2000) 'Building Effective Post-Adoption Services: What is the

Empirical Foundation?', Family Relations, 49(4), pp. 447-455.

Beckett, C., Pinchen, I. and McKeigue, B. (2013) 'Permanence and "Permanence": Outcomes of

Family Placements', British Journal of Social Work, 44(5), pp. 11621179.

Biehal, N., Ellison, S., Baker, C. and Sinclair, I. (2010) Belonging and Permanence. Outcomes in

long-term foster care and adoption, London, BAAF.

Burns, B.J., Phillips, S.D., Wagner, H.R., Barth, R.P., Kolko, D.J., Campbell, Y. and Landsverk, J.

(2004) 'Mental Health Need and Access to Mental Health Services by Youths Involved With

Child Welfare: A National Survey', Journal of the American Academy of Child & Adolescent

Psychiatry, 43(8), pp. 960-970.

Coakley, J.F. and Berrick, J.D. (2008) 'Research Review: In a rush to permanency: preventing

adoption disruption', Child & Family Social Work , 13(1), pp. 101-112.

Dance, C., Ouwejan, D., Beecham, J. and Farmer, E. (2010) Linking and Matching: A survey of

Adoption Agency Practice in England and Wales London, BAAF.

Department for Education (2016) 'Statistical First Release: Children looked after in England

(including adoption) year ending 31 March 2016', London, Department for Education.

Evan B Donaldson Adoption Institute (2004) 'What's Working for Children: A Policy Study of

Adoption Stability and Termination', New York Evan B Donaldson Adoption Institute.

Farmer, E. and Lutman, E. (2012) Effective Working with Neglected Children and their Families,

Jessica Kingsley Publishers.

Farmer, E., Sturgess, W., O'Neill, T. and Wijedasa, D. (2011) Achieving Successful Returns from

Care: What makes reunification work , London, BAAF.

Faulkner, M., Adkins, T., Fong, R. and Rolock, N. (2016) 'Promoting permanency in adoption: A

review of the literature', Southfield,MI, Quality Improvement Center for Adoption &

Guardianship Support and Preservation and Spaulding for Children.

Featherstone, B. and Bywaters, P. (2014) 'An ideological approach to adoption figures means we are

missing important trends', Community Care, Community Care.

Festinger, T. (2002) 'After Adoption: Dissolution or Permanence?', Child Welfare, 81(3), pp. 515-

533.

Fisher, P.A. (2015) 'Review: Adoption, fostering, and the needs of looked-after and adopted

children', Child and Adolescent Mental Health, 20(1), pp. 5-12.

Fratter, J., Rowe, J., Sapsford, D. and Thoburn, J. (1991) Permanent Family Placement: A Decade of

Experience, London, British Agencies for Adoption & Fostering.

Gagnon-Oosterwaal, N., Cossette, L., Smolla, N., Pomerleau, A., Malcuit, G., Chicoine, J.-F., Jéliu,

G., Belhumeur, C. and Berthiaume, C. (2012) 'Pre-Adoption Adversity and Self-Reported

Behavior Problems in 7 Year-Old International Adoptees', Child Psychiatry & Human

Development, 43(4), pp. 648-660.

Grant, M.J. and Booth, A. (2009) 'A typology of reviews: an analysis of 14 review types and

associated methodologies', Health Information & Libraries Journal, 26(2), pp. 91-108.

Holloway, J.S. (1997) 'Outcome in placements for adoption or long term fostering', Archives of

Disease in Childhood, 76(3), pp. 227-230.

Hughes, K., Lowey, H., Quigg, Z. and Bellis, M.A. (2016) 'Relationships between adverse childhood

experiences and adult mental well-being: results from an English national household survey',

BMC Public Health, 16(1), p. 222.

King, S., Gieve, M., Iacopini, G., Hahne, A. and Stradling, H. (2017) 'The Evaluation of The

Adoption Support Fund', London, Tavistock Institute of Human Relations.

Kirton, D. (2013) ''Kinship by design' in England: reconfiguring adoption from Blair to the

coalition', Child & Family Social Work, 18(1), pp. 97-106.

Lloyd, E.C. and Barth, R.P. (2011) 'Developmental outcomes after five years for foster children

returned home, remaining in care, or adopted', Children and Youth Services Review, 33(8), pp.

1383-1391.

Lowe, N., Murch, M., Borkowski, M., Weaver, A., Beckford, V. and Thomas, C. (1999) Supporting

adoption : reframing the approach, London, BAAF.

Machin, D., Cheung, Y.B. and Parmar, M. (2006) Survival Analysis: A Practical Approach,

Chichester, Wiley.

McDonald, T.P., Propp, J.R. and Murphy, K.C. (2001) 'The postadoption experience: Child, parent,

and family predictors of family adjustment to adoption', Child Welfare, 80(1), pp. 71-94.

McGinnis, H., Smith, S.L., Ryan, S. and Howard, J. (2009) 'Beyond Culture Camp: Promoting

Healthy Identity Formation in Adoption', New York Evan B. Donaldson Adoption Institute.

Nanni, V., Uher, R. and Danese, A. (2012) 'Childhood Maltreatment Predicts Unfavorable Course of

Illness and Treatment Outcome in Depression: A Meta-Analysis', The American Journal of

Psychiatry, 169(2), pp. 141 - 151.

NSPCC (2012) 'Returning home from care: What's best for children?', London, NSPCC.

Quinton, D. (2012) Rethinking matching in adoptions from care: A conceptual and research review,

London, BAAF.

Quinton, D., Rushton, A., Dance, C. and Mayes, D. (1998 ) Joining New Families: A Study of

Adoption and Fostering in Middle Childhood, London, Wiley.

Randall, J. (2013) 'Failing to settle: a decade of disruptions in a voluntary adoption agency in

placements made between 2001 and 2011', Adoption & Fostering, 37(2), pp. 188199.

Rolock, N. and White, K.R. (2016) 'Post-permanency discontinuity: A longitudinal examination of

outcomes for foster youth after adoption or guardianship', Children and Youth Services Review,

70, pp. 419-427.

Rushton, A. (2003) 'Support for adoptive families: a review of current evidence on problems, needs

and effectiveness', Adoption and Fostering, 27(3), pp. 41-50.

Rushton, A. (2004) 'A Scoping and Scanning Review of Research on the Adoption of Children

Placed from Public Care', Clinical Child Psychology and Psychiatry, 9(1), pp. 89-106.

Rushton, A. and Dance, C. (2006) 'The Adoption of Children from Public Care: A Prospective Study

of Outcome in Adolescence', Journal of the American Academy of Child & Adolescent Psychiatry,

45(7), pp. 877-883.

Rushton, A., Dance, C., Quinton, D. and Mayes, D. (2001) Siblings in Late Permanent Placements ,

London, BAAF.

Rycus, J.S., Freundlich, M., Hughes, R.C., Keefer, B. and Oakes, E.J. (2006) 'Confronting barriers to

adoption success', Family Court Review, 44(2), pp. 210-230.

Sánchez-Sandoval, Y. (2015) 'Self-perception, self-esteem and life satisfaction in adopted and non-

adopted children and adolescents / Autopercepción, autoestima y satisfacción vital en niños y

adolescentes adoptados y no adoptados', Infancia y Aprendizaje, 38(1), pp. 144-174.

Selwyn, J., Sturgess, W., Quinton, D. and Baxter, C. (2006) Costs and outcomes of non-infant

adoptions, London, BAAF.

Selwyn, J., Wijedasa, D. and Meakings, S. (2014) 'Beyond the adoption order:

challenges,interventions and adoption disruption ', London, Department for Education.

Shonkoff, J.P., Garner, A.S., Siegel, B.S., Dobbins, M.I., Earls, M.F., Garner, A.S., McGuinn, L.,

Pascoe, J. and Wood, D.L. (2012) 'The Lifelong Effects of Early Childhood Adversity and Toxic

Stress', Pediatrics, 129(1), pp. e232-e246.

Singer, J.D. and Willet, J.B. (2003) Applied longitudinal data analysis, New York, Oxford

University Press.

Smith, S.L. (2013) 'A family for life: The vital need to achieve permanency for children in care',

New York, Donaldson Adoption Institute.

Smith, S.L. (2014) Facilitating adoption from care : A compendium of effective and promising

practices, London, CoramBAAF.

Smith, S.L., Howard, J.A., Garnier, P.C. and Ryan, S.D. (2006) 'Where Are We Now?', Adoption

Quarterly, 9(4), pp. 19-44.

Sonuga-Barke, E.J.S., Kennedy, M., Kumsta, R., Knights, N., Golm, D., Rutter, M., Maughan, B.,

Schlotz, W. and Kreppner, J. (2017) 'Child-to-adult neurodevelopmental and mental health

trajectories after early life deprivation: the young adult follow-up of the longitudinal English and

Romanian Adoptees study', The Lancet, 389(10078), pp. 1539-1548.

Thoburn, J., Norton, L. and Rashid, S.P. (2000) Permanent family placement for children of minority

ethnic origin, Jessica Kingsley.

Triseliotis, J. (2002) 'Long-term foster care or adoption? The evidence examined', Child & Family

Social Work, 7(1), pp. 23-33.

U.S. Department of Health and Human Services (2016) 'The AFCARS Report: Preliminary

Estimates for FY 2015 as of June 2016 ', Children's Bureau.

Van IJzendoorn, M.H. and Juffer, F. (2006) 'The Emanuel Miller Memorial Lecture 2006: Adoption

as intervention. Meta-analytic evidence for massive catch-up and plasticity in physical, socio-

emotional, and cognitive development', Journal of Child Psychology and Psychiatry, 47(12), pp.

1228-1245.

Wade, J., Biehal, N., Farrelly, N. and Sinclair, I. (2011) Caring for Abused and Neglected Children:

Making the Right Decisions for Reunification or Long-Term Care, London, Jessica Kingsley.

Welsh Government (2016a) 'Children looked after at 31 March by local authority, gender and age',

Welsh Government.

Welsh Government (2016b) 'Children starting to be looked after during year to 31 March by local

authority and need for care', Welsh Government.

Welsh Government (2016c) 'Episodes finishing for children looked after during year to 31 March by

local authority and reason for finishing', Welsh Government.

White, K.R. (2016) 'Placement Discontinuity for Older Children and Adolescents Who Exit Foster

Care Through Adoption Or Guardianship: A Systematic Review', Child and Adolescent Social

Work Journal, 33(4), pp. 377-394.

Wijedasa, D. and Selwyn, J. (2014) 'Beyond the Adoption Order: An Investigation of Adoption

Disruption in Wales', Cardiff.

Yamaguchi, K. (1991) Event History Analysis, Newbury Park, Sage.

... From a wide range of studies in other countries, it is generally agreed that adoption is better for children than foster care (see, for example, Christoffersen, 2012;Christoffersen et al, 2008;Hjern et al, 2019;. Breakdown rates for children placed for adoption from care are low, less than 4 per cent, as displayed in two recent longitudinal studies of large sample in the UK by Wijedasa and Selwyn (2017) and . Of specific relevance for Norway is the Vinnerljung and Hjern (2011) administrative data-based study of the outcomes for three groups of Swedish children: 900 adopted previous foster children; 3,062 children in long-term foster care; and 900,000 children from the majority population. ...

... Adoption should always be 'in the child's best interests' (Art 21 CRC), yet the child seems absent in much of the process. For instance, the beneficence of adoption is typically framed in terms of outcomes compared to children remaining in foster care (Vinnerljung and Hjern, 2011), and by reference to low adoption breakdown rates in countries such as England and Wales (Wijedasa and Selwyn, 2017) as a sign of success in providing stability. 1 This emphasis on 'hard' empirical data, however, only expresses one aspect of adoption, albeit an important one. ...

... From a wide range of studies in other countries, it is generally agreed that adoption is better for children than foster care (see, for example, Christoffersen, 2012;Christoffersen et al, 2008;Hjern et al, 2019;. Breakdown rates for children placed for adoption from care are low, less than 4 per cent, as displayed in two recent longitudinal studies of large sample in the UK by Wijedasa and Selwyn (2017) and . Of specific relevance for Norway is the Vinnerljung and Hjern (2011) administrative data-based study of the outcomes for three groups of Swedish children: 900 adopted previous foster children; 3,062 children in long-term foster care; and 900,000 children from the majority population. ...

... Adoption should always be 'in the child's best interests' (Art 21 CRC), yet the child seems absent in much of the process. For instance, the beneficence of adoption is typically framed in terms of outcomes compared to children remaining in foster care (Vinnerljung and Hjern, 2011), and by reference to low adoption breakdown rates in countries such as England and Wales (Wijedasa and Selwyn, 2017) as a sign of success in providing stability. 1 This emphasis on 'hard' empirical data, however, only expresses one aspect of adoption, albeit an important one. ...

... From a wide range of studies in other countries, it is generally agreed that adoption is better for children than foster care (see, for example, Christoffersen, 2012;Christoffersen et al, 2008;Hjern et al, 2019;. Breakdown rates for children placed for adoption from care are low, less than 4 per cent, as displayed in two recent longitudinal studies of large sample in the UK by Wijedasa and Selwyn (2017) and . Of specific relevance for Norway is the Vinnerljung and Hjern (2011) administrative data-based study of the outcomes for three groups of Swedish children: 900 adopted previous foster children; 3,062 children in long-term foster care; and 900,000 children from the majority population. ...

... Adoption should always be 'in the child's best interests' (Art 21 CRC), yet the child seems absent in much of the process. For instance, the beneficence of adoption is typically framed in terms of outcomes compared to children remaining in foster care (Vinnerljung and Hjern, 2011), and by reference to low adoption breakdown rates in countries such as England and Wales (Wijedasa and Selwyn, 2017) as a sign of success in providing stability. 1 This emphasis on 'hard' empirical data, however, only expresses one aspect of adoption, albeit an important one. ...

... From a wide range of studies in other countries, it is generally agreed that adoption is better for children than foster care (see, for example, Christoffersen, 2012;Christoffersen et al, 2008;Hjern et al, 2019;. Breakdown rates for children placed for adoption from care are low, less than 4 per cent, as displayed in two recent longitudinal studies of large sample in the UK by Wijedasa and Selwyn (2017) and . Of specific relevance for Norway is the Vinnerljung and Hjern (2011) administrative data-based study of the outcomes for three groups of Swedish children: 900 adopted previous foster children; 3,062 children in long-term foster care; and 900,000 children from the majority population. ...

... Adoption should always be 'in the child's best interests' (Art 21 CRC), yet the child seems absent in much of the process. For instance, the beneficence of adoption is typically framed in terms of outcomes compared to children remaining in foster care (Vinnerljung and Hjern, 2011), and by reference to low adoption breakdown rates in countries such as England and Wales (Wijedasa and Selwyn, 2017) as a sign of success in providing stability. 1 This emphasis on 'hard' empirical data, however, only expresses one aspect of adoption, albeit an important one. ...

... From a wide range of studies in other countries, it is generally agreed that adoption is better for children than foster care (see, for example, Christoffersen, 2012;Christoffersen et al, 2008;Hjern et al, 2019;. Breakdown rates for children placed for adoption from care are low, less than 4 per cent, as displayed in two recent longitudinal studies of large sample in the UK by Wijedasa and Selwyn (2017) and . Of specific relevance for Norway is the Vinnerljung and Hjern (2011) administrative data-based study of the outcomes for three groups of Swedish children: 900 adopted previous foster children; 3,062 children in long-term foster care; and 900,000 children from the majority population. ...

... Adoption should always be 'in the child's best interests' (Art 21 CRC), yet the child seems absent in much of the process. For instance, the beneficence of adoption is typically framed in terms of outcomes compared to children remaining in foster care (Vinnerljung and Hjern, 2011), and by reference to low adoption breakdown rates in countries such as England and Wales (Wijedasa and Selwyn, 2017) as a sign of success in providing stability. 1 This emphasis on 'hard' empirical data, however, only expresses one aspect of adoption, albeit an important one. ...

Full text available here: https://library.oapen.org/handle/20.500.12657/47833

... From a wide range of studies in other countries, it is generally agreed that adoption is better for children than foster care (see, for example, Christoffersen, 2012;Christoffersen et al, 2008;Hjern et al, 2019;. Breakdown rates for children placed for adoption from care are low, less than 4 per cent, as displayed in two recent longitudinal studies of large sample in the UK by Wijedasa and Selwyn (2017) and . Of specific relevance for Norway is the Vinnerljung and Hjern (2011) administrative data-based study of the outcomes for three groups of Swedish children: 900 adopted previous foster children; 3,062 children in long-term foster care; and 900,000 children from the majority population. ...

... Adoption should always be 'in the child's best interests' (Art 21 CRC), yet the child seems absent in much of the process. For instance, the beneficence of adoption is typically framed in terms of outcomes compared to children remaining in foster care (Vinnerljung and Hjern, 2011), and by reference to low adoption breakdown rates in countries such as England and Wales (Wijedasa and Selwyn, 2017) as a sign of success in providing stability. 1 This emphasis on 'hard' empirical data, however, only expresses one aspect of adoption, albeit an important one. ...

  • Jenny Krutzinna Jenny Krutzinna

Full text available here: https://library.oapen.org/handle/20.500.12657/47833

... Similarly, children's emotional and behavioural difficulties can have a significant deleterious impact on adoptive and foster families and placement stability (e.g. Wijedasa and Selwyn, 2017;Palacios et al. 2019). These issues are highlighted in studies about the mental health needs of care-experienced children and young people (e.g. ...

  • Suzanne Mooney
  • Carolyn Blair
  • Tom Teggart Tom Teggart

Literature review and exploration of stakeholder views regarding the need for Allied Health Professional support and intervention for Looked After Children.

... Internasjonalt har forskningen om adopsjon i barnevernet i hovedsak dreid seg om betydningen av adopsjon som plasseringstiltak i forhold til bruk av fosterhjem eller institusjonsplassering. Denne forskningen viser tydelig at adopsjon gir bedre resultat for barnet når barnets livssituasjon vurderes som tenåring og ung voksen (se eksempelvis Bohman & Sigvaldsson, 1985;Christoffersen, 2012;Quinton & Selwyn, 2009;Selwyn & Quinton, 2004;Triseliotis, 2002;Triseliotis & Hill, 1990;Vinnerljung & Hjern, 2011;Wijedasa & Selwyn, 2017; se også Gypen, Vanderfaeillie, De Maeyer, Belenger & Van Holen, 2017). Bedre resultater dreier seg om barnas fysiske og kognitive utvikling, selvverd, atferdsmessige og følelsesmessige problemer. ...

Sammendrag Barnevernet kan gjennomføre adopsjoner av fosterbarn uten biologiske foreldres samtykke. Dette representerer en betydelig utøvelse av statlig makt, og norsk rett krever særlig tungtveiende grunner for å tillate adopsjon. Det er også sjelden slike adopsjoner gjøres i Norge. Artikkelen undersøker hvordan befolkningen vurderer adopsjon som barneverntiltak, og om de er på linje med politikerne og retten i dette spørsmålet. Vi har gjennomført en surveyundersøkelse til et representativt utvalg (n = 1000) av den norske befolkningen om synet på barnevernsadopsjon. Resultatene fra studien viser at befolkningen generelt synes å være positive til adopsjon som tiltak i barnevernet, og at politikerne synes å være på linje med befolkningen. Synet på verdien av biologiske bånd i befolkningen varierer med politisk orientering, alder, kjønn og inntekt, men majoriteten er av den oppfatning at fordelene med adopsjon trumfer verdien av biologiske bånd. Når det kommer til bruk av tvang, er befolkningen delt i sitt syn. Vi tror våre funn skyldes at et barneperspektiv har fått større plass i det norske samfunnet.

... This represents a fall of 28% from the peak figure of 5360 in 2015-2016(Department for Education, 2018, demonstrating the decrease in adoption placements as a result of the change in legal ruling. In addition, it is estimated that approximately 3% to 4% of children adopted in the UK are returned to care after an adoption order is granted (Triseliotis, 2002;Wijedasa and Selwyn, 2017). However, estimates range from 10% to 50% when additional factors (including age of the child at placement, specific learning or developmental difficulties, specific challenging behaviours) are considered (see also Selwyn, Wijedasa and Meakings, 2014). ...

Understanding the interplay between genetic factors and family environmental processes (e.g. interparental relationship quality, positive vs negative parenting practices) and children's mental health (e.g. anxiety, depression, conduct problems, ADHD) in the contexts of adoption and foster care research and practice is critical for effective prevention and intervention programme development. While evidence highlights the importance of family relationship processes (e.g. interparental relationship quality, parent‒child relationship quality) for the mental health and well-being of children in adoption and foster care, there is relatively limited evidence of effective interventions specifically for these families. Additionally, family-based interventions not specific to the context of adoption and foster care typically show small to medium effects, and even where interventions are efficacious, not all children benefit. One explanation for why interventions may not work well for some is that responses to an intervention may be influenced by an individual's genetic make-up. Alternatively, the targets of family relationship level interventions (e.g. parenting processes) may not always affect the specific environment 'trigger' deemed salient to specific child/adolescent outcomes. This article summarises how genetically informed research designs can help disentangle genetic from environmental processes underlying psychopathology outcomes for children, and how this evidence can provide improved insights into the development of more effective preventive intervention targets for adoptive and foster families. We discuss current difficulties in translating behavioural genetics research to prevention science and provide recommendations to bridge the gap between behavioural genetics research and prevention science, with lessons for adoption and foster care research and practice.

Purpose This mixed-methods study reports findings from 337 LGBTQ adults in the United States who reported delays or disruptions in the adoption or foster care process. Methods An online survey was distributed by Clark University and the Human Rights Campaign (HRC), a large LGBTQ organization, with the goal of understanding LGBTQ individuals' experiences with adoption and foster care. Results Respondents highlighted LGBTQ specific and general barriers, at multiple levels (legal, adoption agency, birth/foster family, child) that interfered with (a) timely progression through the adoption/foster care process and (b) permanency planning. Conclusion Adoption practitioners need training in the explicit and implicit ways that LGBTQ prospective adopters may be exposed to marginalization and stress in the adoption/foster care process, and legislation preventing the discrimination of LGBTQ prospective parents is needed.

Background: Time-limited, early-life exposures to institutional deprivation are associated with disorders in childhood, but it is unknown whether effects persist into adulthood. We used data from the English and Romanian Adoptees study to assess whether deprivation-associated adverse neurodevelopmental and mental health outcomes persist into young adulthood. Methods: The English and Romanian Adoptees study is a longitudinal, natural experiment investigation into the long-term outcomes of individuals who spent from soon after birth to up to 43 months in severe deprivation in Romanian institutions before being adopted into the UK. We used developmentally appropriate standard questionnaires, interviews completed by parents and adoptees, and direct measures of IQ to measure symptoms of autism spectrum disorder, inattention and overactivity, disinhibited social engagement, conduct or emotional problems, and cognitive impairment (IQ score <80) during childhood (ages 6, 11, and 15 years) and in young adulthood (22-25 years). For analysis, Romanian adoptees were split into those who spent less than 6 months in an institution and those who spent more than 6 months in an institution. We used a comparison group of UK adoptees who did not experience deprivation. We used mixed-effects regression models for ordered-categorical outcome variables to compare symptom levels and trends between groups. Findings: Romanian adoptees who experienced less than 6 months in an institution (n=67 at ages 6 years; n=50 at young adulthood) and UK controls (n=52 at age 6 years; n=39 at young adulthood) had similarly low levels of symptoms across most ages and outcomes. By contrast, Romanian adoptees exposed to more than 6 months in an institution (n=98 at ages 6 years; n=72 at young adulthood) had persistently higher rates than UK controls of symptoms of autism spectrum disorder, disinhibited social engagement, and inattention and overactivity through to young adulthood (pooled p<0·0001 for all). Cognitive impairment in the group who spent more than 6 months in an institution remitted from markedly higher rates at ages 6 years (p=0·0001) and 11 years (p=0·0016) compared with UK controls, to normal rates at young adulthood (p=0·76). By contrast, self-rated emotional symptoms showed a late-onset pattern with minimal differences versus UK controls at ages 11 years (p=0·0449) and 15 years (p=0·17), and then marked increases by young adulthood (p=0·0005), with similar effects seen for parent ratings. The high deprivation group also had a higher proportion of people with low educational achievement (p=0·0195), unemployment (p=0·0124), and mental health service use (p=0·0120, p=0·0032, and p=0·0003 for use when aged <11 years, 11-14 years, and 15-23 years, respectively) than the UK control group. A fifth (n=15) of individuals who spent more than 6 months in an institution were problem-free at all assessments. Interpretation: Notwithstanding the resilience shown by some adoptees and the adult remission of cognitive impairment, extended early deprivation was associated with long-term deleterious effects on wellbeing that seem insusceptible to years of nurturance and support in adoptive families. Funding: Economic and Social Research Council, Medical Research Council, Department of Health, Jacobs Foundation, Nuffield Foundation.

Background Individuals' childhood experiences can strongly influence their future health and well-being. Adverse childhood experiences (ACEs) such as abuse and dysfunctional home environments show strong cumulative relationships with physical and mental illness yet less is known about their effects on mental well-being in the general population. Methods A nationally representative household survey of English adults (n = 3,885) measuring current mental well-being (Short Edinburgh-Warwick Mental Well-being Scale SWEMWBS) and life satisfaction and retrospective exposure to nine ACEs. Results Almost half of participants (46.4 %) had suffered at least one ACE and 8.3 % had suffered four or more. Adjusted odds ratios (AORs) for low life satisfaction and low mental well-being increased with the number of ACEs. AORs for low ratings of all individual SWEMWBS components also increased with ACE count, particularly never or rarely feeling close to others. Of individual ACEs, growing up in a household affected by mental illness and suffering sexual abuse had the most relationships with markers of mental well-being. Conclusions Childhood adversity has a strong cumulative relationship with adult mental well-being. Comprehensive mental health strategies should incorporate interventions to prevent ACEs and moderate their impacts from the very earliest stages of life.

  • Kevin R. White

For over two decades, practitioners, advocates, and scholars involved with the U.S. child welfare system have engaged in coordinated efforts to increase the number of foster youth who find stable, permanent homes through adoption or guardianship, and these efforts have been shaped and guided by federal policies and directives. As a result, the number of children adopted or placed into guardianship out of foster care has increased significantly. This trend has significant implications for child welfare research, policy, and practice. However, the risk and protective factors for post-permanency discontinuity, or placement changes that occur after legal finalization of an adoption or guardianship, have received little attention in the literature. Also, many previous studies that investigated post-permanency adjustment for former foster youth have been limited by serious design and/or conceptual flaws. The purpose of this study is to investigate the peer-reviewed literature that examines risk or protective factors for discontinuity, or outcomes proximal to discontinuity, for older foster youth. A systematic search located 18 quantitative, quasi-experimental studies published in peer-reviewed journals that implemented multivariable methods. This review finds that the quality of the research evidence is generally weak, but previous studies do suggest several risk and protective factors for post-permanency discontinuity, including child, family, and service characteristics.

  • Yolanda Sánchez-Sandoval Yolanda Sánchez-Sandoval

This paper analyses whether the specific challenges of adolescence and adoption may affect the self-assessment of adopted adolescents. A total of 618 children and adolescents from Andalusia (southern Spain) between 8 and 16 years of age (105 adopted and 513 non-adopted) participated in this study. For comparison purposes, different groups of non-adopted children (current classmates, children in residential care and children who have grown up with their biological families in socially disadvantaged areas) also participated. The Self-Perception Profile for Adolescents, the Rosenberg Self-Esteem Scale (RSE) and the Students' Life Satisfaction Scale (SLSS) were utilized. The results reflect positive self-assessments on average at the beginning and during adolescence. There is a decline in self-perception, self-esteem and life satisfaction at around 11 or 12 years, with a slight recovery beginning at the age of 14. This trend occurs in adopted and non-adopted children and adolescents. Adopted children present self-assessments similar to children who live with their biological family. Children in residential care exhibit self-perception, self-esteem and life satisfaction that is well below the rest.

  • Nancy Rolock Nancy Rolock
  • Kevin R. White

For over two decades, federal policies and case practices in child welfare have shifted to prioritize legal permanence for children in foster care, and increasing numbers of children have been placed in permanent adoptive or guardianship homes. Despite this change, little research has examined the long-term stability of legally permanent adoptive and guardianship homes for former foster youth. This study used child welfare administrative records to track a population of 51,576 children in Illinois who exited foster care through adoption or guardianship for ten years or until the age of majority. Univariate and bivariate analyses were conducted to describe the population, and a multivariable Cox proportional hazards model was estimated to examine the relationship between child age and discontinuity, controlling for several pre-placement characteristics. Results indicated that the vast majority (87%) of children did not experience post-permanency discontinuity. In addition, African American children and children who had more moves in foster care had a higher hazard of discontinuity, while children placed with siblings and children who spent three or more years in foster care had a lower hazard of discontinuity. Study findings also indicated that prevention efforts should be targeted at families with adolescents. This study contributes significantly to the scant literature on long-term outcomes for adoptive and guardianship families, and suggests several areas for future research.

This article describes a follow-up of a complete cohort of fifty-nine children involved in care proceedings (legal proceedings to determine whether a local authority should be given parental responsibility for a child) in the year 2004–05 in an English local authority. The study took place between four and six years after the conclusion of proceedings. It looked, first, at whether the plan agreed at the final hearing had been implemented and (where relevant) what had happened in the interim. Second, it looked at whether placements had been successful, factors associated with success or failure, and the nature and outcome of new plans when these had had to be made. At the time of the study, the care plans of 97 per cent of the children had been implemented, either immediately or within eighteen months. Sixty-six per cent of the children were still in their designated permanent placements, 15 per cent had experienced moves either within or back to their immediate or extended family and 15 per cent had experienced a placement breakdown. The reasons for changes and breakdowns are discussed and the possibility raised that a more flexible approach to permanency might, in some cases, be helpful.