Assessing the Context of Permanency and Reunification in the Foster Care System. 5. Adoption from Foster Care

12/01/2001

Logistic regression models of the likelihood of adoption from foster care are presented in Tables IIF.I (pooled data) and IIF.2 (individual states). The universe is composed of all spells that began in 1990-93 in the nine study states. Adoptive exits are observed for up to four years following entry.

Table IIF.1 
Logistic Regression Model of Adoption From Foster Care Within 4 Years of Entry. Nine states, 1990-1993 Admissions.

Predictor Variable

Category Standardized parameter Odds Ratio

Age at Entry

0 - 2 months 0.51 15.54
3 - 11 months 0.26 5.69
1 - 2 years 0.23 3.36
3 - 5 years 0.16 2.28
6 - 8 years 0.08 1.61
9 - 11 years * - - 1.00
12 - 14 years -0.23 0.32
15 - 17 years -0.41 0.11

Gender

Male -0.01 0.97
Female * - - 1.00

Race/Ethnic

African American -0.18 0.51
Hispanic -0.07 0.72
White, Other * - - 1.00

Region

Primary Urban County -0.26 0.38
Remainder of State - - 1.00

Sequence

First Spell in Care * - - 1.00
Reentry Spell 0.07 1.51

Care Type

Congregate Care -0.15 0.47
Foster Care, Kin Care * - - 1.00

Spell Duration

Log(months in spell) 0.44 1.74

State

Alabama -0.16 0.15
California -0.46 0.18
Illinois -0.26 0.24
Maryland -0.05 0.60
Michigan * - - 1.00
Missouri -0.09 0.49
New Mexico -0.06 0.40
New York -0.38 0.20
Wisconsin -0.17 0.28

Year of Entry

1990 -0.02 0.91
1991 -0.02 0.93
1992 * - - 1.00
1993 0.02 1.10
  Intercept -3.30  
exp(intercept) 0.04  
Concordance 0.86  
N spells observed 313,359  
N adoptions in 4 years 22,141  
Percent adopted in 4 years 7.1%  

Note: Predictors noted with "*" are not contained in model, but are the excluded category for their variable. The odds ratios of 1.00 associated with these predictors are predetermined. The odds ratios for other categories of the variable express the effect in relation to that of the excluded category


Table IIF.2
Logistic Regression Model of Adoption Within Four Years of Foster Care Entry.Nine States, All Foster Care Admissions From 1990-1993.
  State 9 states (no kin term) 9 states (with kin term)
AL CA IL MD MI MO NM NY WI
ratios of relative odds of adoption    

Age at Entry

0 - 2 months 117.71 49.32 16.27 49.77 2.98 21.50 7.01 11.57 13.03 12.31 12.28
3 - 11 months 21.91 18.17 4.95 14.79 1.79 7.72 3.23 4.74 4.88 5.20 5.31
1 - 2 years 16.04 10.83 2.56 8.35 1.52 3.36 2.72 2.61 3.60 3.17 3.32
3 - 5 years 7.43 5.38 1.94 5.87 1.47 2.48 1.59 1.82 2.35 2.18 2.29
6 - 8 years 3.47 2.49 1.57 2.74 1.40 1.43 1.19 1.43 1.90 1.59 1.65
9 - 11 years * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
12 - 14 years - - 0.46 0.23 0.32 0.35 0.45 0.25 0.20 0.25 0.33 0.32
15 - 17 years - - 0.26 0.06 0.11 0.07 0.15 0.16 0.08 0.20 0.11 0.11

Gender

                     
Male 1.15 0.93 0.95 0.89 0.97 1.03 0.84 0.99 1.05 0.98 0.97
Female * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Race/Ethnic

African American 0.34 0.33 0.51 0.54 0.76 0.86 0.56 0.58 0.43 0.54 0.58
Hispanic - - 0.73 0.72 0.48 0.66 2.00 0.79 0.56 0.74 0.53 0.62
White, Other * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Region
Primary Urban County 0.83 0.25 0.61 0.30 0.49 0.65 0.93 0.60 0.35 0.38 0.44
Remainder of State * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Spell Sequence

First Spell in Care * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Reentry Spell 1.52 1.49 1.14 1.57 1.16 1.19 1.70 1.81 1.41 1.47 1.35

Care Type

Kinship Care - - 0.31 0.37 0.38 - - 0.51 0.93 0.27 - - - - 0.26
Congregate Care 0.22 0.99 0.06 0.35 0.07 0.23 0.71 0.55 0.11 0.47 0.36
Foster Care * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Spell Duration

Log(months) 2.56 1.79 1.24 2.03 1.94 1.95 3.32 1.82 2.08 1.60 1.69

Year Admitted

1990 1.47 1.02 1.04 0.81 0.86 1.23 1.10 0.67 0.87 0.90 0.89
1991 0.96 1.01 0.87 0.83 0.93 1.09 0.88 0.85 0.86 0.95 0.94
1992 * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
1993 1.13 1.08 0.95 1.02 1.23 1.02 1.26 1.15 1.24 1.09 1.10

Intercept

-8.09 -5.64 -3.25 -4.90 -3.09 -4.80 -6.04 -4.80 -5.05 -4.24 -4.26

exp(intercept)

0.00 0.00 0.04 0.01 0.05 0.01 0.00 0.01 0.01 0.01 0.01

Concordance

0.95 0.89 0.84 0.89 0.80 0.88 0.90 0.87 0.89 0.84 0.86

N spells observed

7,511 110,889 40,981 10,639 24,470 17,609 4,735 76,849 19,676 313,359 313,359

N adoptions in 4 years

289 7,080 2,586 1,090 4,170 1,613 333 4,186 794 22,141 22,141

Percent adopted in 4 years

3.8% 6.4% 6.3% 10.2% 17.0% 9.2% 7.0% 5.4% 4.0% 7.1% 7.1%

Note: Predictors noted with "*" are not contained in model, but are the "excluded" category for their variable. Predetermined odds ratios of 1.00 are assigned to these categories. Odds ratios for other categories of the same variable express effects in relation to that of the excluded category.


The pooled analysis highlights the dramatic relationship between age at spell entry and the likelihood of exit by adoption. Children who entered foster care as newborns (0-2 months) were 49 times more likely [15.54/0.32] to have an adoptive exit from care than were children who entered care at the ages of 12-14 years, and over 15 times more likely [ 15.54/1.00] to have an adoptive exit than children who entered care between the ages of 9-11. Newborns were even 2.73 times more likely [ 15.54/5.69] to be adopted from care than were children who entered care between the ages of 3 months and 1 year.

Regarding the other variables:

  • African American (.51) and Hispanic (.72) children were less likely to be adopted than were white children.
  • Children in congregate care placements (.47) were less likely to be adopted than were children in family foster placements.
  • Children in primary urban counties (.38) were less likely to be adopted than were children from the other regions of their state.
  • Children in reentry spells (1.51) were more likely to be adopted than were children in their first spell in care.
  • The likelihood of adoption increased steadily between the 1990 and 1993 entry cohorts.
  • Michigan was by far the most successful state at achieving adoptive exits from foster care, with Maryland, Missouri, and New Mexico showing adoptive likelihoods higher than those of the remaining states.

Adoption models for each state are shown separately in Table IIF.2. Here, we can see that Michigan differed from the other eight states in several ways. Most importantly, the adoption activity in Michigan was less limited to the population of children who entered foster care as newborns. The reason that the odds ratio for newborns in Michigan (2.98) is so much smaller than those for the other states (which range from 7 to 118) is not because Michigan has fewer infant adoptions. It is because Michigan has many more adoptions of children who entered at 9-11 years (the "excluded" category which forms the basis of the odds ratios). The odds ratio is the relative likelihood of adoption for children from the two categories. Michigan is also one of three states (Illinois and Missouri are the others) that move almost as high a share of children in their first spells to adoption as reentry spells. It is also one of two states (Missouri is the other) where adoption of African American children from foster care occurs with at least a 75 percent likelihood as the adoption of white children. Thus, the markedly higher adoption rates in Michigan seem to result from an expansion of the domain of potentially adoptable foster children in comparison to the other states observed.12

The likelihood of adoptive outcomes from kinship spells in care is very low in the six states where kinship care can be tracked.

Several individual states show unusual patterns in the likelihood of adoption from foster care. California had extremely low adoption levels in Los Angeles, relative to other parts of the state; and unusually high levels of adoption from congregate care. In New York, an unusually high proportion of foster care adoptions is from reentry spells, and the likelihood of adoption almost doubled between 1990 and 1993.

The time trend for all states pooled is an increase in the likelihood of adoption for successive cohorts. At least five of the individual states reflect this general trend, and only Missouri showed a substantial decrease in adoptions during this period.

Like reunification, adoptions are generally regarded as positive exits from foster care, in that they define new life situations that are intended to be permanent, nonrestrictive, and nurturing. Adoptions are a preferred resolution when family-centered options seem unfeasible. Because adoption and reunification are mutually exclusive events, in that they cannot both be discharge destinations for the same spell in care, we might expect that they would be negatively related to one another at the aggregate level. But again the statewide findings do not support such a conclusion.

However, there is a clear relation between reunification and adoption in child welfare practice. This was made explicit in the Adoption and Safe Families Act, which included language that directed states to initiate pre-adoptive activity for most foster care cases where reunification had not been accomplished within 15 months of entry. This law does not rule out future reunification activity, in fact concurrent planning activity is encouraged. It does, however, explicitly target an increase in adoptions from foster care and a reduction in the time that elapses before adoption.


12.  It is also worth noting that some states pay providers more according to the degree of difficulty to adopt.