Chapter 3

The major findings of the analyses conducted on States’ 1999 context and outcome data are presented and discussed in this chapter. The primary objective of the analyses was to enhance the understanding of the Nation’s child welfare systems with respect to State child welfare agency populations and performances on the outcome measures. The analyses were not designed as evaluative assessments of State child welfare agencies. When States are identified by name, as they are for the maximums and minimums of the ranges for particular variables, the purpose is to demonstrate that the States in these positions cannot be consistently characterized by any particular features, such as population size or population characteristics. That is, States with extremely large populations of children are no more likely to be at the maximum (or minimum) point on a given variable or outcome measure than States with relatively small child populations.

The initial analyses focused on identifying similarities and differences among States, with a finding of extensive State variation for almost all context variables and outcome measures. Potential explanations for State variation were explored by assessing the correlations between relevant context variables and between context variables and outcome measures. ¹ Although the results of these assessments do not provide definitive explanations for State variation, they identify areas of interest for further investigation by both Federal and State agencies concerned with understanding child welfare systems and improving the outcomes experienced by children who come into contact with these systems.

This chapter provides the results of the analyses as they pertain to the goals of child safety and permanency. The next chapter presents the data for individual States.

Key Findings Regarding Attaining the Goal of Child Safety
Children’s safety is the paramount concern of all State child welfare systems. The Child Welfare Outcomes Annual Reports address the issue of children’s safety through two outcome measures. One measure assesses the incidence of recurrence of child maltreatment among children substantiated by the child welfare system as victims of maltreatment; the other examines the incidence of maltreatment of children in foster care by their caretakers.

To better understand State efforts to attain the goal of child safety, the Child Welfare Outcomes Annual Reports also provide State-level information pertaining to child maltreatment, such as the number of child maltreatment victims, the rate of child maltreatment victims per 1,000 children in the population, the age and race/ethnicity of child maltreatment victims, and the type of maltreatment experienced by victims.

The questions addressed in the analyses of child maltreatment-related context variables and outcome measures are presented below. All analyses were conducted on data from the National Child Abuse and Neglect Data System (NCANDS) Detailed Case Data Component (DCDC).² The number of States included in the analyses ranges from 20 to 28, depending upon the availability of data. Additional information on child maltreatment from the NCANDS data can be found in Child Maltreatment 1999. Highlights from this report are provided in Appendix B of this report.

What was the variation among States in their rates of child maltreatment victims (children who are the subject of substantiated or indicated abuse or neglect reports)’

The rate of child maltreatment victims per 1,000 children in the population exhibited considerable variation among the 28 States included in this analysis. State variation in child victim rates was found to be unrelated to the percentages of children living in poverty in the States, the percentages of children living in metropolitan areas in the States, and the percentages of children of color in the States’ child populations.

The term “child victims ” is defined in this report as children who are the subject of substantiated or indicated maltreatment reports. The data represent the number of unique children for whom a maltreatment report or reports were substantiated or indicated during the reporting year. It is important to note that the term does not refer to all children in the State who may have experienced maltreatment, but only to those who were the subject of maltreatment reports and for whom the maltreatment was substantiated or indicated.

Twenty-eight States provided sufficient data to calculate the rate of child maltreatment victims per 1,000 children in the population. This rate ranged from 1.7 in Pennsylvania to 18.3 in Florida, with a median of 9.7.

The low rate of child victims in Pennsylvania can be attributed, at least in part, to the fact that, in accordance with State policy, many reports of child neglect in that State receive a general protective services investigation rather than a child protective services investigation and are not classified as child maltreatment. However, child neglect that meets the definition of “serious ” is classified as child maltreatment by the State. It is important to note that the fact that some reports of neglect are not classified as child maltreatment in Pennsylvania and not investigated by child protective services does not mean that the children and families involved do not receive services. It does mean, however, that information concerning the children involved in these reports is not provided to the NCANDS DCDC.

One question raised by Pennsylvania’s policy regarding reports of child neglect was whether State variation in the rate of child victims may be related to State differences in statutory requirements, policies, or definitions pertaining to substantiating a child maltreatment report. This possibility has been supported by findings reported in Child Maltreatment 1996. Although the question could not be addressed with data available for the Child Welfare Outcomes Annual Reports, it may become particularly important to consider as States begin to implement differential response systems that involve alternatives to child maltreatment investigations for child maltreatment reports judged to be “non-severe. ”

The variation among States in the rate of child victims (children who are the subject of substantiated or indicated maltreatment reports) per 1,000 children in the population was not correlated with State variations in (1) the percentages of children living in poverty (Spearman’s rho = 0.121), (2) the percentages of children residing in metropolitan areas (Spearman’s rho = 0.222), or (3) the percentages of children of color in the child population (Spearman’s rho = -0.018).

What was the variation among States in the ages of child maltreatment victims (children who were the subject of substantiated or indicated maltreatment reports)’

Little variation among States was found in the analysis of the distribution of child maltreatment victims among the age groups of (1) younger than 4 years, (2) 4 to 7 years, (3) 8 to 11 years, (4) 12 to 15 years, and (5) 16 to 17 years. A key finding, however, was that for 15 of the 27 States included in the analysis, at least one-fourth of all child maltreatment victims were 12 years of age or older.

Twenty-seven States were included in the analysis of State variation in the distribution of child maltreatment victims across age groups. States were highly similar with respect to the percentages of child victims who were aged 4 to 7 years and 8 to 11 years, but exhibited some differences in the percentages of children who were younger than 4 years old. However, the extent of the variation for this age group was less than 18 percentage points.

Although the majority of child maltreatment victims in all States were younger than 12 years old, in 15 States, 25 percent or more of the child maltreatment victims were 12 years of age or older, and in 4 States, at least one-third of child victims were in this age group. Only 3 States had less than one-fifth of child maltreatment victims in the 12-to-17 year-old age group. These findings demonstrate that child maltreatment is not a problem that is exclusive to young children. State child welfare agencies are responsible for ensuring the safety of adolescents as well as young children, and, in many States, a substantial percentage of the children requiring protection are adolescents.

What was the variation among States in the types of maltreatment experienced by child maltreatment victims (children who are the subject of a substantiated or indicated maltreatment report)’

States varied with respect to the percentages of child victims experiencing each type of maltreatment (physical abuse, neglect, medical neglect, sexual abuse, and psychological maltreatment/emotional abuse). In all but 3 of the 26 States included in the analysis, the largest percentage of child victims experienced neglect.

The analysis of variation in types of maltreatment included 26 States. Pennsylvania was excluded from the analysis because its policy pertaining to reports of non-severe neglect affects its comparability with other States with respect to types of maltreatment.

Neglect was the most frequent type of maltreatment experienced by children who were the subject of substantiated or indicated child maltreatment reports (child victims) in all States except Florida, Hawaii, and Utah. In 15 of the 26 States included in the analyses, more than 50 percent of child victims experienced neglect. The correlation between the percentages of child victims experiencing neglect in a State and the percentages of children living in poverty in a State was not significant (Pearson’s r = 0.228).

The percentages of child victims experiencing child sexual abuse ranged from 3 in Connecticut to 39.5 in Vermont, with a median of 7.8. In 18 of the 26 States, the incidence of child victims experiencing sexual abuse was less than 10 percent. Vermont ’s high percentage of sexual abuse victims reflects the fact that for this State, the “sexual abuse ” category for type of maltreatment includes abuse by nonfamilial/noncustodial perpetrators.

State variation in the incidence of physical abuse was less than the variation for sexual abuse or neglect, ranging from 3.6 percent in North Carolina to 28.5 percent in New Mexico, with a median of 21.7 percent. In no State was physical abuse the dominant category.

The incidence of psychological maltreatment accounted for less than 10 percent of child victims in all but 4 States. However, in Utah, it was the most frequently experienced type of maltreatment (37.9 percent), and in Connecticut, psychological maltreatment accounted for 29.8 percent of child victims. The median for this type of maltreatment was 3.2 percent.

What was the variation among States in their success in keeping child maltreatment victims safe from future maltreatment’

Success in keeping children safe from future maltreatment was assessed through (1) the incidence of recurrence of a substantiated or indicated child maltreatment within 6 months of a prior substantiated or indicated maltreatment, and (2) the incidence of child maltreatment by foster parents or facility staff member. Variation among States in the incidence of child maltreatment recurrence was found to be correlated with variation in the rate of child victims per 1,000 children in the State population. The incidence of maltreatment from foster parents or facility staff exhibited little State variation and was an infrequent event in all States.

Outcome measure 1.1 is: For all children who were victims of substantiated or indicated child abuse and/or neglect during the first 6 months of the reporting period, what percentage had another substantiated or indicated report within a 6-month period’ Of the 28 States providing data to calculate this outcome measure, the recurrence of substantiated or indicated maltreatment ranged from 2.1 percent in Delaware to 13.5 percent in New York, with a median of 6.7 percent.

State variation in the incidence of maltreatment recurrence was not correlated with variations in the percentages of child victims (children who were the subject of a substantiated or indicated maltreatment report) experiencing neglect (Pearson’s r = 0.273). However, State variation in the incidence of maltreatment recurrence was significantly correlated with the rate of child victims per 1,000 children in the State population (Spearman ’s rho = 0.451, p = 0.016 [two-tailed]). States with higher rates of child maltreatment victims tended to have correspondingly higher incidences of child maltreatment recurrence and vice versa. For example, Pennsylvania had the lowest child maltreatment victim rate (1.7 per 1,000 children in the population) and the second lowest incidence of recurrence of maltreatment within 6 months (2.5 percent). The two States with the highest incidences of recurrence (New York at 13.5 percent and Connecticut at 13.0 percent) also had relatively high rates of child victims per 1,000 children in the population (New York at 13.8 and Connecticut at 15.8). Although this pattern represents a general tendency, it was not observed for all States. In Florida, for example, the rate of child victims was high (18 per 1,000 children in the population), but the incidence of recurrence (6.2 percent) was below the median.

The finding of a significant correlation between child maltreatment recurrence and the rate of child victims per 1,000 children in the population cannot be interpreted without additional information, particularly about States’ substantiation policies and maltreatment definitions and what happens to children after a maltreatment report has been substantiated. This correlation raises the question of whether State policies or definitions regarding child maltreatment substantiation account for State variation in both child victim and recurrence rates, since both variables require a finding of a substantiated or indicated child maltreatment. Also, it may be that States that tend to substantiate only the most serious cases of maltreatment have lower maltreatment recurrence rates within 6 months of a prior substantiated maltreatment because the maltreatment victims are in foster care during that time.

Outcome measure 2.1 is: Of all children who were in foster care during the reporting period, what percentage was the subject of substantiated or indicated maltreatment by a foster parent or facility staff member’ For the 20 States with sufficient data to calculate this measure, the percentages ranged from 0.1 percent in Arizona, Delaware, and Wyoming, to 2.3 percent in Rhode Island, with a median of 0.5 percent. It is important to note that not all States report the same information to NCANDS pertaining to this outcome. For example, 6 of the 20 State child welfare agencies included in this analysis do not receive information regarding children in institutions who have been maltreated. Because investigations of these incidents are conducted by other agencies, the child welfare agency does not report the numbers to the NCANDS DCDC.

Key Findings Regarding Attaining the Goal of Permanency
The passage of the Adoption and Safe Families Act (ASFA) strengthened the focus on permanency as a goal for the Nation ’s child welfare systems. The Department considers “permanent homes ” as reunifications with birth families, finalized adoptions, and legal guardianships. Although some States do not currently view legal guardianship as a permanency option, current research on assisted guardianship and its impact on permanency for children eventually may result in an increase in the number of States for which guardianship is a permanency option.

In addition to emphasizing the goal of permanency for children, ASFA specified that permanency must be achieved in a timely manner, and identified specific practices and procedures for achieving this objective. ASFA’s focus on achieving permanency in a timely manner is reflected in the child welfare outcome measures that assess the length of time children are in foster care before they are reunified with birth families or have finalized adoptions.

To better understand the context in which State child welfare agencies operate with respect to attaining permanency for children, the Child Welfare Outcomes Annual Reports present information in the State data pages on such variables as the number of children in foster care on the first and last days of the fiscal year; the number of children entering and exiting the system during a given fiscal year; the number of children served by the foster care system during the year; the number of children waiting for adoption; the number of children adopted; and the age and race/ethnicity of children in foster care, waiting children, and adopted children.

The analyses of data pertaining to the context variables and outcome measures relevant to attaining permanent homes for children focused on the differences and similarities among States and examined potential explanations for differences. The data pertaining to State foster care populations (including data on waiting children), and to outcomes 3 through 7, come from the Adoption and Foster Care Analysis Reporting System (AFCARS) Annual Foster Care Database; the data on the adoption populations that pertains to the context variables are from the AFCARS Annual Adoption Database.

The total number of jurisdictions that could be included in an analysis was 52 all 50 States, the District of Columbia, and Puerto Rico. However, because Alaska, Michigan, and Nevada did not have sufficient data to calculate either the foster care context variables or the permanency-related outcome measures, these States were included only in the analyses of adoption data reported to the AFCARS Annual Adoption Database. Other jurisdictions were excluded from a given analysis either because more than 30 percent of the data for a given outcome measure was coded as “missing, ” or because there was a problem with the data submitted to AFCARS for a particular outcome or measure. One State, for example, used an incorrect code for the data on race/ethnicity, and therefore was excluded from all analyses of State variation pertaining to children’s race/ethnicity.

What was the variation among States in the rate of children entering the foster care system in FY 1999’

States varied with respect to the rate of entry into foster care per 1,000 children in the State population. This variation was found to be unrelated to State variations in the rate of child victims, the percentages of children living in poverty, and the percentages of children living in metropolitan areas. The rate of entry into foster care for children of color per 1,000 children of color in the population was higher in all States than the rate of entry for Caucasian children. This difference was primarily accounted for by the higher rate of entry into foster care of African-American children than for Caucasian children.

Forty-seven States, the District of Columbia, and Puerto Rico provided data to calculate the rate of entry into foster care per 1,000 children in the State population. For these jurisdictions, the rate of entry ranged from 1.7 in Texas and Virginia to 13.4 in the District of Columbia, with a median of 4.6.

State variation in the rate of entry into foster care was found to be unrelated to State variations in the rate of child victims (children who were the subject of a substantiated or indicated maltreatment report) per 1,000 children in the population (Spearman ’s rho = 0.302), the percentages of children in the State residing in metropolitan areas (Spearman’s rho = -0.148), and the percentages of children in the State living in poverty (Spearman’s rho = -0.200). (The District of Columbia was excluded from these analyses.) With respect to poverty, however, many States that had relatively low percentages of children living in poverty had high rates of entry into foster care and vice versa. Minnesota, for example, which had the highest rate of entry per 1,000 children into foster care, had one of the lowest percentages of children living in poverty (7.9), while Texas, which had the lowest rate of entry into foster care (1.7), had one of the highest percentages of children living in poverty (22.1).

In all States, more children entered than exited foster care in FY 1999. In 16 States, the difference between entries and exits exceeded 20 percent, while in 17 States the difference was 10 percent or less.

The rate of entry into foster care for children of color per 1,000 children of color in the population ranged from 1.9 in Texas to 24.7 in South Dakota, with a median of 6.8. This rate was higher in all States than the rate of entry into foster care for Caucasian children. The difference between children of color and Caucasian children was primarily accounted for by rates of entry into foster care for African-American children, which ranged from 2.9 in Mississippi to 37.8 in Minnesota, with a median of 12.0. For all States, the rate of entry of African-American children was higher than the rate of entry for Caucasian children and, in 30 States, the rate of entry of African-American children was more than 3 times higher than the rate of entry of Caucasian children.

On the last day of FY 1999, the percentage of African-American children in foster care exceeded the percentage of African-American children in the State population in all States. For 37 States, the percentage of African-American children in foster care on September 30, 1999 was more than twice the percentage of African-American children in the population, and for 14 States the percentage of African-American children in foster care on that date was more than 4 times the percentage of African-American children in the population.

What was the variation among States in the ages of children entering the foster care system in FY 1999’

Differences among States with respect to the percentages of children entering foster care occurred primarily in the 16-to-18 year-old age group, with a range of 3.4 to 30. This finding may be due to the fact that some State child welfare systems provide foster care services to children who enter the system as a result of behavioral problems or through the juvenile justice system, while other State child welfare systems do not. For most States, the highest percentages of entries into foster care occurred in the 11-to-15 year-old age group.

Forty-seven States and the District of Columbia provided data on the ages of children entering care across five age groups (younger than 1, 1 to 5 years, 6 to 10 years, 11 to 15 years, and 16 to 18 years). For 31 States, the highest percentage of entries into foster care was in the 11-to-15 year-old age group; for 13 States, at least one-third of the entries into foster care were in this age group. Also, in all but 3 States, more 11-to-15 year-olds entered foster care than exited. These findings draw attention to the fact that many of the adolescents in foster care have not been in foster care since they were young children. For many States, a substantial percentage of children who enter foster care enter as preadolescents or adolescents.

The percentages of children entering foster care who were 16 to 18 years old ranged from 3.4 percent to 30 percent. In 7 States, at least 20 percent of the entries into foster care were in that age group, while in 19 States, fewer than 10 percent of entries were in that age group. This variation among States in the percentage of entries into foster care of older children may reflect the fact that some State child welfare systems include children in their foster care population who enter the system through the juvenile justice system, or who enter as a result of behavior problems, while other States do not. Information from the Child and Family Services (CFS) reviews may enhance our understanding of this issue.

What was the variation among States in the median length of stay of children in foster care on the last day of FY 1999’

Median length of stay in foster care ranged widely across States. In general the median length of stay was longer for children in care on the last day of FY 1999 than it was for children who exited foster care in FY 1999.

The median length of stay of children in foster care on the last day of FY 1999 ranged from 9.9 months in Florida to 40.9 months in Illinois, with a median of 16.5 months. Of the 47 States, the District of Columbia, and Puerto Rico that reported information to calculate this measure, 6 States and Puerto Rico had median lengths of stay that exceeded 2 years, while 8 States had median lengths of stay of less than 1 year. Children exiting foster care during FY 1999 tended to have shorter lengths of stay than children in care on the last day of the fiscal year, with a range of 2.1 months in Idaho to 47.3 months in Illinois. The median length of stay for children exiting care in FY 1999 was 10.3 months.

State variation in median length of stay is particularly difficult to interpret because of the many factors that influence this statistic. For example, children in kinship care tend to stay in care for longer periods of time than children in non-kinship care, and the placement is generally viewed as a permanent one. Therefore, States (such as Illinois) in which a large percentage of children are in kinship foster care may demonstrate longer median lengths of stay than other States. The establishment of assisted guardianship in several States may result in a gradual reduction of median length of stay as children in kinship care move into guardianship arrangements. Information obtained from the CFS reviews may enhance the understanding of variations among States regarding median length of stay in foster care.

What was the variation among States in the percentages of children in foster care who had finalized adoptions in FY 1999 and in the age and race/ethnicity of these children’

The percentage of all children in foster care on the first day of FY 1999 who had finalized adoptions in FY 1999 varied across States. In most States, the largest percentage of adoptions occurred in the 1-to-5 year-old age group. The percentage of adoptions that involved children of color also varied among States. However, the percentage of all adoptions that involved children of color was similar to the percentage of all children in the foster care population who were children of color.

The percentage of all children in care on the first day of FY 1999 who had finalized adoptions in FY 1999 ranged from 2.7 in Puerto Rico to 16.6 in Utah, with a median of 9.0. In 42 States, the largest percentages of finalized adoptions occurred in the 1-to-5 year-old age group, with a median of 44.7. In 9 States, more than 50 percent of adoptions occurred in this age group. The variation among States in the percentages of finalized adoptions in the 1-to-5 year-old age group was found to be significantly correlated with State variation in the percentages of 1-to-5 year-olds in the foster care population (Pearson’s r = 0.32; p = 0.02). That is, States with larger percentages of 1-to-5 year-olds in their foster care population also tended to have larger percentages of 1-to-5 year-olds in their finalized adoptions population and vice versa.

The 6-to-10 year-old age group accounted for at least one-third of finalized adoptions in 36 States. Finalized adoptions of children younger than 1 year (median = 13 percent) or 16 to 18 years old (median = 1.6 percent) were infrequent, while finalized adoptions of children in the 11-to-15 year-old group ranged from 0 percent in Delaware to 24.9 per cent in New York, with a median of 14.4 percent. There were only 3 States (Kansas, Arkansas, and New York) in which 20 percent or more of the finalized adoptions involved children 11 to 15 years old.

In 46 States, the percentages of all adoptions that involved children of color ranged from 4.3 in Vermont to 86.5 in Hawaii (the District of Columbia and Puerto Rico were excluded from these analyses), with a median of 42.5. State variation in the percentages of adoptions that involved children of color was found to be correlated with State variation in the percentage of children in foster care (on the first day of FY 1999) who were children of color (Pearson’s r = 0.906, p < 0.0001 [two-tailed]). States with higher percentages of children of color in foster care tended to have higher percentages of adoptions of children of color, and vice versa.

What was the variation among State child welfare systems with respect to their success in achieving permanency for children’

Although States demonstrated some differences in the percentages of children exiting to permanent homes through reunification with their families, adoption, or guardianship, the vast majority of children who left the foster care system in FY 1999 exited to permanent homes. However, children who were older than 12 years at the time of entry into foster care were less likely than other children to exit to permanent homes and more likely to exit to emancipation.

Outcome 3 is “Increase permanency for children in foster care. ” Five measures were developed to assess this outcome:

• Measure 3.1: For all children who exited foster care, what percentage left either to reunification, adoption, or legal guardianship’
• Measure 3.2: For children who exited foster care and were identified as having a diagnosed disability, what percentage left either to reunification, adoption, or legal guardianship’
• Measure 3.3: For children who exited foster care and were older than age 12 at the time of their most recent entry into care, what percentage left either to reunification, adoption, or legal guardianship’
• Measure 3.4: Of all children exiting foster care to emancipation, what percentage was age 12 or younger at the time of entry into care’
• Measure 3.5: For all children who exited foster care, what percentage by racial/ethnic category left either to reunification, adoption, or legal guardianship’

For these measures, adoption, guardianship, and reunification represent exits to permanent homes, while the category of “other ” represents “discharge reasons ” that are not considered permanency, such as transfers to another system, emancipations from foster care, running away, and death. The analyses did not include an assessment of exits to guardianship because many States do not yet include this arrangement as a permanency option.

It is important to note that the findings of the analyses of outcome measures 3.1 through 3.5 may not accurately reflect State variation in exits to adoption (or to “other ”) because, for many States, the number of exits to finalized adoptions is under-reported. In 37 States, the number of adoptions reported to the AFCARS Adoption Database (as provided in Context Section E) was greater than the number of exits to adoption reported to the AFCARS Foster Care Database (used to calculate outcome measure 3.1). While a slight difference would be expected because the AFCARS Adoption Database includes a small number of private adoptions in which the public agency is involved, the number of private adoptions reported to the AFCARS Adoption Database accounted for only a small percentage of the discrepancies, which were extremely large in several States. For example, in eight States the number of adoptions reported to the AFCARS Adoption Data Base was at least twice as many as the number of exits to adoption reported to the AFCARS Foster Care Data Base. In 17 States, the number reported to the Adoption Data Base exceeded the number of exits to adoption by 25 percent or more. For one State, the number of children with finalized adoptions reported to the AFCARS Adoption Data Base was 2,052 in FY 1999, while the number of exits to adoption reported to the AFCARS Foster Care Data Base for FY 1999 was 75. This State’s data for outcome measure 3.1 were not included in the analyses because of the extent of this discrepancy.

The adoptions that were not reported as exits to adoption may have been reported as “missing data ” or exits to “other. ” Some States indicated in their comment letters that the discrepancy in their adoption numbers is due to the fact that the State uses adoption-related discharge categories that are not compliant with the AFCARS categories and, therefore, are classified as “missing data ” in AFCARS. Other States attributed the discrepancy to the possibility that their child welfare workers are reluctant to provide a discharge reason of adoption because of confidentiality issues, and therefore record discharges to adoption as “other. ” Most States exhibiting this inconsistency indicated that the numbers reported to the AFCARS Adoption Database accurately reflect the actual number of finalized adoptions for Fiscal Years 1998 and 1999, and that the exits to adoption shown in outcome measures 3.1 and 5.1 are not correct. States commenting on this discrepancy indicated that they have taken measures to resolve it.

Thirty-nine States and the District of Columbia were included in the analysis of outcome measure 3.1, which assesses exits to permanency. The percentages of children exiting foster care to reunification with their families ranged from 30.8 in Illinois to 84.7 in Idaho, with a median of 66.3. In 7 States, at least 75 percent of exits from foster care were reunifications. Exits to adoption ranged from 4.6 percent in Wyoming to 45.6 percent in Illinois, with a median of 12.5 percent. Only one State reported a greater percentage of adoptions than reunifications. The percentages of children who did not exit to adoption, reunification, or guardianship (i.e., the percentages included in the category “other ”) ranged from 4.3 in Maine to 22.8 in Wisconsin, with a median of 10.1.

Thirty-nine States and the District of Columbia also provided analyzable data for outcome measure 3.2, which examines exits to permanency for children with disabilities. The percentages of children with disabilities who exited foster care to situations other than reunification, adoption, or guardianship ranged from 4.5 in North Dakota to 41.2 in Vermont, with a median of 17.1. For four States (Maine, Illinois, Hawaii and Montana), the percentages of children with disabilities exiting to adoption were greater than the percentages exiting to reunification.

Thirty-eight States and the District of Columbia reported sufficient data for outcome measure 3.3, which assesses exits to permanency for children who were older than 12 years at the time of entry into care. The percentages of these children who did not exit to a permanent arrangement ranged from 12.7 in Alabama to 47.3 in Illinois. In all States, less than 2.1 percent of these children exited to adoption. However, in the District of Columbia, 21.9 percent of children who were older than 12 at the time of entry into foster care exited foster care to a finalized adoption.

Forty-five States and the District of Columbia reported sufficient data to calculate measure 3.4, which examines whether children exiting to emancipation (i.e., “aging out ” of the foster care system) entered foster care when they were 12 years or younger or when they were older than 12 years. Overall, the percentages of all children exiting foster care that exited to emancipation ranged from 1.7 in Oregon to 14.2 in Tennessee. In 19 States, at least 75 percent of the children exiting to emancipation entered when they were older than 12 years.

The findings of the analyses for outcome measures 3.2, 3.3, and 3.4 suggest that attaining permanency for children who are disabled or older than 12 at the time of entry into foster care is a challenge for many States. The finding with regard to age is particularly critical because in many States, more than 50 percent of the children entering the foster care system in FY 1999 were aged 11 years or older.

Thirty-seven States and the District of Columbia provided sufficient data to be included in the analysis of measure 3.5, which pertains to the race/ethnicity of children exiting foster care. This analysis revealed no clear patterns of variation among the States. In general, children of color exited to permanency at the same rate as Caucasian children. Overall, the median percentage for African-American children not exiting to permanent arrangements (the category of “other ”) was 9.3, while the median percentage for Caucasian children not exiting to permanent arrangements was 9.8.

What was the variation among States in reunifying children in a timely manner’

Although most States appear to achieve reunifications in a timely manner, a significant positive correlation was found between the percentage of reunifications in FY 1999 occurring within 12 months of entry into foster care and the percentage of entries into foster care in FY 1999 that were re-entries within 12 months of a prior foster care episode. This suggests that assessments of a State’s success with respect to timely reunifications must take into account the percentages of re-entries into foster care within 12 months of a prior episode.

Child welfare outcome 4 is “Reduce time to reunification without increasing re-entries into foster care. ” This outcome is assessed through the following measures:
• Measure 4.1: Of all children who were reunified with their parents or caretakers at the time of discharge from foster care, what percentage was reunified in the following time periods: (1) fewer than 12 months, (2) at least 12 months but fewer than 24 months, (3) at least 24 months, but fewer than 36 months, (4) at least 36 months but fewer than 48 months, and (5) 48 or more months’
• Measure 4.2: Of all children who entered foster care during the reporting period, what percentage re-entered care within 12 months of a prior foster care episode’

For the 47 States, the District of Columbia, and Puerto Rico providing sufficient data to calculate these measures, the percentages of exits to reunifications that occurred within 12 months of entry into foster care ranged from 33.8 in Illinois to 87.3 in New Mexico, with a median of 67.3. In 44 States, more than 50 percent of the exits to reunification were achieved within 12 months, and in 45 States fewer than 10 percent of the exits to reunification involved children who had been in foster care for 48 months or more.

Although these data suggest that most States are reunifying children in a timely manner, when the incidence of re-entries into care is considered, a different picture emerges. The percentages of entries into foster care in FY 1999 that were re-entries within 12 months of a prior episode ranged from 0 in West Virginia to 26.7 in Iowa, with a median of 10.8. For 9 States, 20 percent or more of entries into foster care were re-entries within 12 months.

The correlation between State variation in the percentages of reunifications within 12 months of entry into foster care and State variation in the percentages of children entering foster care who were re-entering within 12 months of a prior episode was significant (Pearson’s r = 0.407, p = 0.004 [two-tailed]). States with relatively high percentages of reunifications occurring within 12 months tended to have relatively high percentages of entries into foster care in FY 1999 that were re-entries within 12 months of a prior foster care episode. States with low percentages of reunifications occurring within 12 months tended to have correspondingly low percentages of re-entries within 12 months. Although this finding does not establish a causal relationship between these factors, it does support the position that time to reunification must be considered in conjunction with re-entry rates in assessing State performance. Expedited reunification should not be achieved at the risk of children re-entering the system.

What was the variation among States in their achieving finalized adoptions within 24 months of a child’s entry into foster care’

There was considerable State variation in the percentage of exits to adoption that occurred in FY 1999 within 24 months of entry into foster care. It is difficult to interpret the findings for this measure, however, because time to adoption is based on the number of exits to adoption reported to the AFCARS Foster Care Database. As noted previously, in 37 States the number of exits to adoption reported to the AFCARS Foster Care Database was less than the number of finalized adoptions reported to the AFCARS Adoption Database. Consequently, for many States, the data on time to adoption do not include all of the adoptions that took place.

Child welfare outcome 5 is “Reduce time in foster care to adoption. ” The outcome is assessed by measure 5.1 Of all children who exited foster care to a finalized adoption, what percentage exited care in the following time periods: (1) Fewer than 12 months from the time of latest removal, (2) at least 12 months but fewer than 24 months, (3) at least 24 months but fewer than 36 months, (4) at least 36 months but fewer than 48 months, and (5) 48 or more months’

For the 45 States and the District of Columbia providing data to calculate this measure, the percentages of children experiencing a finalized adoption within 24 months of entry into foster care ranged from 3 in New York to 81.6 in Connecticut, with the median being 25. In 6 States, at least 50 percent of the adoptions occurred within 24 months of the time of entry into care.

The findings of this analysis are difficult to interpret. First, as noted in the discussion of outcome measure 3.1, in 37 States, the number of children exiting to a finalized adoption reported to the AFCARS Foster Care Database was less than (sometimes half as many as) the number of finalized adoptions reported to the AFCARS Adoption Database. For these States, the time to adoption reported for measure 5.1 does not represent all adoptions that took place in FY 1999. Because some of these States indicated in their comment letters that they were taking measures to resolve this discrepancy, future Child Welfare Outcomes Annual Reports are expected to have more complete data on the number of exits from foster care to adoption.

Second, many States have recently begun intensive efforts to find adoptive homes for children who have been in the foster care system for long periods of time. A State’s success in achieving finalized adoptions for this “backlog ” of children in foster care will result in data for outcome measure 5.1 showing a relatively high percentage of children adopted as having been in care for long periods of time, which results in a potentially misleading impression that these States are performing poorly with respect to the percentage of adoptions occurring within 24 months. For this reason, an analysis of an individual State’s performance on measure 5.1 should include an assessment of changes in the number of adoptions occurring within 24 months of entry into care, as well as changes in the percentage of adoptions occurring during that time period.

What was the variation among States with respect to providing stable placements for children in foster care’

In most States, the majority of children who are in foster care for fewer than 12 months experience no more than 2 placements while in care. However, children who are in care for more than 12 months often experience 3 or more placement settings.

There is general agreement in the child welfare field that multiple placements of children while in foster care do not promote positive growth and development. This perspective is reflected in child welfare outcome 6: Increase placement stability. Attainment of outcome 6 is assessed by measure 6.1 Of all children served who had been in foster care for the time periods listed below, what percentage had no more than two placement settings during that time period: (1) Less than 12 months from the time of latest removal from home, (2) At least 12 months, but less than 24 months, (3) at least 24 months, but less than 36 months, (4) at least 36 months, but less than 48 months, (5) 48 or more months’

Of the 46 States reporting data for this measure, the percentages of children in foster care for fewer than 12 months who experienced no more than 2 placements ranged from 58.9 in Utah to 99.8 in West Virginia, with a median of 80.8. Despite these generally positive findings, the number of placements experienced by children increased with increasing time in foster care, with substantial percentages of children experiencing three or more placements while in foster care. For children who were in foster care for 48 or more months, the percentages experiencing no more than 2 placement settings ranged from 12.5 in Arizona to 90.6 in Puerto Rico, with a median of 37.0.

What was the variation among States with respect to ensuring that young children are not placed in group homes or institutions’

Only a few States exhibited high percentages of placements of young children in group homes and/or institutions. Comment letters from these States indicated that the high number of placements in these settings reflects the States’ use of emergency shelter facilities for short periods of time after receiving a child maltreatment report if there are questions about the child’s safety.

There is concern in the child welfare field that placements of young children in group homes or institutions may be detrimental to optimum growth and development. Child welfare outcome 7� Reduce placements of young children in group homes or institutions � is intended to address this concern. This outcome is assessed by measure 7.1: For all children who entered foster care during the reporting period and were age 12 or younger at the time of their most recent placement, what percentage was placed in a group home or institution’

For the 47 States, District of Columbia, and Puerto Rico providing data for this measure, the percentages of children who were aged 12 years old or younger when they were placed in group homes ranged from 0.2 in Illinois to 25 in Utah, with a median of 3.8. The percentages of children who were 12 years old or younger when placed in institutions ranged from 0.1 in Florida to 24 in Minnesota, with a median of 4.9.

States exhibiting high percentages of placements of younger children in either group homes or institutions attributed these data to their practice of placing children in emergency shelters for short periods of time when there are concerns regarding the ability to maintain the child safely in the home. These States noted that because AFCARS defines group homes and institutions in terms of the number of beds, emergency shelter placements become coded as institutional or group home placements (depending on the size of the placement facility). Although the use of shelters rather than family homes for emergency short-term placements may be viewed by some child welfare professionals as an undesirable practice, emergency shelter placements are not equivalent to group homes or institutions, and some procedure for distinguishing emergency shelter placements would enhance interpretation of the data for outcome measure 7.1.

Summary of Findings
What have we learned from the analyses of the data presented in Child Welfare Outcomes 1999 regarding the outcomes experienced by children who come into contact with the Nation’s foster care systems’

The analyses presented in this chapter focused on the differences and similarities among State child welfare systems with respect to the child welfare context variables and outcome measures provided in the Child Welfare Outcomes Annual Reports. A key finding is that State child welfare systems differ extensively with respect to child welfare populations and performances on the outcome measures. Furthermore, this variation is not characterized by any systematic pattern. States that are similar with respect to a given context variable or outcome measure are quite different with regard to other variables or outcome measures. The considerable differences among State child welfare systems regarding the context variables provide strong support for the Department’s decision to assess State performance on the outcome measures by examining changes in performance over time within a State, rather than by comparing performances across States.

Highlights of the findings regarding State performance on the seven child welfare outcomes are as follows:
• Outcome 1: Reduce recurrence of child abuse and/or neglect.
The incidence of recurrence of a substantiated or indicated maltreatment within 6 months of a prior substantiated or indicated maltreatment may not be sufficient information to assess a State ’s performance with regard to ensuring the safety of children who come into contact with the child welfare system. In some instances, a low incidence of maltreatment recurrence may reflect State policies or definitions pertaining to child maltreatment substantiation, rather than a State’s success in keeping children safe. Additional information about State child welfare systems’ policies pertaining to substantiation is necessary for a meaningful interpretation of the incidence of maltreatment recurrence. The CFS reviews will take a similar approach to understanding the incidence of maltreatment recurrence through reviews of case records and interviews with various stakeholders.

• Outcome 2: Reduce the incidence of child abuse and/or neglect in foster care.
Maltreatment of children in foster care by their caretakers is an infrequent occurrence, and State differences with respect to this statistic are small. However, there is general agreement that any incidence of child maltreatment by a foster parent or facility staff person is unacceptable, and States must continue their efforts to reduce the occurrence of these events.

• Outcome 3: Increase permanency for children in foster care.
Greater efforts are needed on the part of child welfare agencies to find permanent homes for children who are older than 12 years when they enter foster care. These children were less likely than younger children to exit foster care in FY 1999 to permanent homes and more likely to exit to emancipation. This is an important concern because in many States, adolescents constituted a substantial percentage of the entries into foster care in FY 1999. Two resource centers within the Department’s training and technical assistance network the National Resource Center for Foster Care and Permanency Planning and the National Resource Center on Youth Development are available to assist State child welfare agencies in addressing this challenge.

• Outcome 4: Reduce time in foster care to reunification without increasing re-entry.
The significant positive correlation found between the percentages of reunifications occurring within 12 months of entry into foster care and entries into foster care in FY 1999 that were re-entries within 12 months, suggests that these events must be assessed conjointly. Although a causal relationship between the variables cannot be established based on the analysis performed, the size of the correlation suggests a substantial linkage between these measures.

• Outcome 5: Reduce time in foster care to adoption.
In assessing a State’s performance with respect to the timeliness of adoptions, it is necessary to examine trends in the numbers of adoptions occurring during each time period as well as trends in the percentages of adoptions. As States continue to experience success in their efforts to find adoptive families for children who have been in care for long periods of time, there is a risk that the distribution of percentages across the selected “time periods to adoption ” will be skewed toward the higher time periods, resulting in a potentially misleading impression that the State is not performing well in achieving adoptions quickly. Once the children who have been in foster care for long periods of time are moved to permanent homes, there will be less risk of “misinterpreting ” the data presented in percentages.

• Outcome 6: Increase placement stability
Most States appear to be doing a good job of limiting the number of placement settings experienced by children who are in foster care for fewer than 12 months. However, as time in foster care increases, the percentage of children experiencing multiple placements also increases. For many States, greater efforts are needed to minimize the number of placement settings experienced by children, regardless of their time in foster care.

• Outcome 7: Reduce placements of young children in group homes or institutions.
For most States, the incidence of placement of young children in group homes or institutions is low. However, even in situations in which the incidence is fairly high, it is difficult to interpret. Several States reported using emergency shelters as short-term placements for children who are the subjects of child maltreatment reports when there is concern about their immediate safety. Based on the AFCARS definitions for placements, these shelters are coded as group homes or institutions in the AFCARS database.

A final lesson from the data analyses is that more information than is captured in the data presented in the Child Welfare Outcomes Annual Reports may be necessary for a meaningful interpretation of a State’s performance on the child welfare outcome measures. At present, the Child Welfare Outcomes Annual Reports include only context information pertaining to numbers of children and some of their characteristics. The Department anticipates that the CFS reviews will provide additional information regarding the broader context in which child welfare systems operate and thus will enhance the understanding and interpretation of States’ performances on the outcome measures.

State Context and Outcome Data for 1998 and 1999.
The Child Welfare Outcomes Annual Reports present each State ’s data separately to enhance an understanding of changes in the State context and State performance on outcome measures over time. The following chapter provides the data for each State.

Child welfare context data are presented first, followed by data for the outcome measures. The last page is reserved for the State ’s comments regarding its data. As noted in chapter 2, initial copies of each State ’s data were sent to the State child welfare agency directors and data managers for their review and comment. Thirty-eight States provided comments ranging from explanations of their data to concerns about their performance on particular measures. Clarifications and definitions of the terms used and data issues relevant to the State data pages are provided in Appendix K.

1) A probability level of at least 0.05 was established as necessary for a finding to be considered statistically significant. Back

2) The child maltreatment context data that are presented in section B of the State data pages are from the NCANDS Summary Data Component (SDC), which collects aggregate data from States on key child maltreatment-related variables. Because States do not aggregate the data they report to the NCANDS SDC in the same manner (e.g., some States may provide the number of child victims while other States may provide the number of substantiated reports), using SDC data to examine State variation is not appropriate. Consequently, the analyses of child maltreatment-related context variables presented in this chapter use data from the NCANDS DCDC. Although fewer States participate in the NCANDS DCDC than in the SDC, the DCDC data are more amenable to examinations of State differences and similarities. Back