comparison of public and private foster care in colorado ... · was created by merging colorado...

39
Comparison of Public and Private Foster Care in Colorado | Social Work Research Center i October 2012 Comparison of Public and Private Foster Care in Colorado Final Report

Upload: others

Post on 21-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center i

    October 2012

    Comparison of Public

    and Private Foster

    Care in Colorado

    Final Report

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center ii

    ACKNOWLEDGEMENTS

    Report Authors

    Marc Winokur Graig Crawford Keri Batchelder

    Applied Research in Child Welfare (ARCH) Project Participants

    Adams County

    Arapahoe County Boulder County

    Broomfield County Denver County Douglas County El Paso County

    Jefferson County Larimer County Pueblo County

    Colorado Department of Human Services Colorado Administrative Review Division

    Touchstone Health Partners

    Workgroup Contributors

    Lee Oesterle Barb Weinstein

    Research Associates

    Valerie Ebanks-Thompson

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center iii

    TABLE OF CONTENTS

    Executive Summary .......................................................................................................................... v

    1. Introduction ................................................................................................................................ 1

    1.1 Study Rationale ..................................................................................................................... 1

    1.2 Study Context ........................................................................................................................ 1

    1.3 Research Questions ............................................................................................................... 2

    2. Literature Review ........................................................................................................................ 2

    2.1 Child Welfare Privatization .................................................................................................... 2

    2.2 Foster Care Privatization ....................................................................................................... 3

    2.3 Private and Public Foster Care Outcomes ............................................................................. 3

    3. Methods ...................................................................................................................................... 4

    3.1 Data Collection ...................................................................................................................... 5

    3.2 Study Parameters .................................................................................................................. 5

    3.3 Variable Definitions ............................................................................................................... 6

    3.4 Outcome Measures ............................................................................................................... 7

    3.5 Cost Measures ....................................................................................................................... 8

    3.6 Sample Selection ................................................................................................................... 8

    3.7 Data Analysis ......................................................................................................................... 9

    4. Demographic Comparison .......................................................................................................... 9

    4.1 0-5 Age Group ..................................................................................................................... 10

    4.2 6-12 Age Group ................................................................................................................... 11

    4.3 13-18 Age Group ................................................................................................................. 12

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center iv

    5. Outcome Comparison ............................................................................................................... 13

    5.1 Demographic Characteristics .............................................................................................. 14

    5.2 0-5 Age Group ..................................................................................................................... 14

    5.3 6-12 Age Group ................................................................................................................... 16

    5.4 13-18 Age Group ................................................................................................................. 17

    6. Cost Comparison ....................................................................................................................... 19

    6.1 0-5 Age Group ..................................................................................................................... 19

    6.2 6-12 Age Group ................................................................................................................... 21

    6.3 13-18 Age Group ................................................................................................................. 23

    7. Discussion .................................................................................................................................. 25

    7.1 Conclusions .......................................................................................................................... 25

    7.2 Limitations ........................................................................................................................... 26

    7.3 Implications ......................................................................................................................... 27

    8. References ................................................................................................................................ 29

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center v

    Comparison of Public and Private Foster Care in Colorado

    Executive Summary

    The Applied Research in Child Welfare (ARCH) Project is a partnership between Colorado

    State University (CSU), the Colorado Department of Human Services (CDHS), and the

    Departments of Human/Social Services in Adams, Arapahoe, Boulder, Broomfield, Denver,

    Douglas, El Paso, Jefferson, Larimer, and Pueblo counties. The purpose of the ARCH Project is to

    conduct applied research on child welfare interventions that informs social work practice and

    policy in Colorado. This report presents results from an outcome study of foster care in

    Colorado, which was conducted by the Social Work Research Center in the School of Social

    Work at CSU with funding from the ARCH Project.

    Introduction

    Facing unprecedented budgetary constraints and accountability requirements, child

    welfare agencies are hungry for research that analyzes the outcomes and costs of children

    placed in foster care. The steady growth in the privatization of foster care services has

    magnified this need, as policymakers also are looking to the research for guidance on how

    private and public foster care compare on outcomes and costs. However, there are few studies

    that have used experimental designs because randomly assigning children who require out-of-

    home placement to either private or public foster care has ethical implications for the safety

    and well-being of the children. As a result, non-experimental comparative and associational

    research designs are more commonly used in foster care research. Overall, there is little

    independent, empirical evidence on this topic, and the research that has been conducted

    shows mixed findings in both outcome and cost differences between private foster care and

    public foster care.

    According to child welfare practitioners, outcomes for children in foster care are

    multidimensional as they are based on the intersection of provider performance, caregiver

    involvement, county practice, and child characteristics. While they are not responsible for the

    initial difficulties that precipitate an out-of-home (OOH) placement, child placement agencies

    (CPA) and foster parents are expected to provide a safe and nurturing environment that helps

    children to mitigate some of these deficits. Therefore, this study is not designed to determine

    which placement type is the most appropriate, as that depends on the specific child, family,

    caseworker, and provider. The study is intended to supply caseworkers and providers with

    information on foster care outcomes and costs to assist them in making data-driven decisions

    about how to best serve children. The following three research questions guided the study:

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center vi

    1. Are there statistically significant differences between children in CPA and county foster

    care on demographic characteristics for the 0-5, 6-12, and 13-18 age groups?

    2. Are there statistically significant differences between children in CPA and county foster

    care on permanency, safety, and subsequent involvement outcomes for the 0-5, 6-12,

    and 13-18 age groups?

    3. Are there statistically significant differences between children in CPA and county foster

    care on average daily rates, administration rates, provider rates, and costs per episode

    of care for the 0-5, 6-12, and 13-18 age groups?

    Methods

    This study employed a between-county comparative research design to analyze the

    outcomes and costs for children with a closed foster care placement in the 10 ARCH counties

    during the 2008-2010 fiscal years. The outcome data were aggregated across the three fiscal

    years while the cost data were collected from fiscal year 2008 and fiscal year 2009. The

    demographic, placement, outcome, and cost data for this study were collected from individual

    case records entered into Colorado Trails, which is Colorado’s Statewide Automated Child

    Welfare Information System (SACWIS). The main inclusion criteria for the outcome study were

    children with a closed service authorization at any time between 7/1/07 and 6/30/10 from

    these placement types: CPA Foster Care (Private Foster Care) and County Foster Care (Public

    Foster Care). If a foster care authorization had a governing body at any time during the

    placement, it was considered to be CPA foster rather than county foster. The unit of analysis for

    the outcome study was “episode of care” rather than service authorization. An episode of care

    was created by merging Colorado Trails service authorizations into a continuous span of care

    for a child with a given placement provider.

    The study employed a non-probability sampling design that was both purposive and

    convenient. The selected sample included children ages 0-18 that were in the child welfare

    system under delinquency and/or dependency and neglect court actions in the ARCH counties.

    Children residing within the Division of Youth Corrections (DYC), children from out-of-state, and

    children with privately paid placements were not eligible for selection. The overall sample size

    was 11,661 with 6,740 episodes of care for CPA foster and 4,921 for county foster. The

    outcomes for this study were as follows: permanency at discharge, reentry, institutional abuse,

    subsequent placements of any type, subsequent community placements, subsequent

    residential placements, subsequent placements days of any type, subsequent community

    placement days, subsequent residential placement days, subsequent DYC involvement days,

    and subsequent DYC commitment days. The cost measures for this study were average daily

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center vii

    rate, average provider rate, average administration rate, and average cost per episode. The

    foster care groups were separated into three age groups (0-5, 6-12, and 13-18) because of the

    diverse needs and services for children in the different age ranges. The comparison of

    demographic characteristics was intended to examine the equivalence between the two foster

    care groups. The demographics were chosen based on their accessibility and demonstrated

    relationships with case outcomes.

    Findings

    Similar to the studies cited in the literature, the results for the outcome and cost

    comparisons for public and private foster care were mixed and varied by age group. Overall, the

    outcomes for county foster care were greater than or comparable to the outcomes for CPA

    foster care, but CPA foster care had lower costs per episode of care than did county foster care.

    Furthermore, there were differences in client acuity but no definitive sense of which placement

    type was serving higher risk cases. For the 0-5 age group, there were statistically significant

    differences between the foster care groups on permanency at discharge, institutional abuse,

    subsequent any placements, and subsequent any placement days, with county foster care

    having an advantage for these outcomes. However, there was no difference between the foster

    care groups on reentry. The cost per episode of care was more expensive for county foster care,

    with county foster care having higher administration costs and CPA foster care having higher

    provider costs. The higher provider costs suggest that children in the 0-5 age range have higher

    needs as measured by the Needs Based Care (NBC) instrument which sets CPA foster parent

    rates. Furthermore, the CPA foster group was comprised of a higher percentage of non-

    Caucasian children. However, the county foster group was comprised of a higher percentage of

    children in Program Area 6 (special needs are barrier to adoption). As a result, the difference in

    client acuity between the foster care groups for this age range is unclear.

    For the 6-12 age group, there were statistically significant differences between the

    foster care groups on subsequent any placements, subsequent any placement days, subsequent

    community placements, and subsequent community placement days, with county foster care

    having an advantage for these outcomes. There were no differences between the groups on

    permanency at discharge, reentry, institutional abuse, subsequent residential placements and

    placement days, and subsequent DYC involvement days and DYC commitment days. The cost

    per episode of care was more expensive for county foster care, with county foster care having

    higher administration costs and CPA foster care having higher provider rates. The two foster

    care groups appeared to be equivalent for the 6-12 age group, as there were no demographic

    differences.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center viii

    For the 13-18 age group, there were statistically significant differences between the

    foster care groups on permanency at discharge, subsequent any placement days, and

    subsequent community placements, with county foster care having an advantage for these

    outcomes. There also was a statistically significant difference for institutional abuse with CPA

    foster care having an advantage for this outcome. There were no differences between the

    groups on reentry, subsequent any placements, subsequent community placement days,

    subsequent residential placements and residential placement days, and subsequent DYC

    involvement days and DYC commitment days. The cost per episode of care was more expensive

    for county foster care, with CPA foster care having higher administration costs and county

    foster care having higher provider costs. However, CPA foster care appeared to serve higher

    risk cases based on ethnicity and prior placement.

    Limitations

    The outcome study faced numerous methodological challenges common to most

    research on child welfare interventions using comparative research designs and secondary

    data. The most notable limitation is the differences in populations served by the private and

    public foster care agencies due to client acuity. However, these differences were not consistent

    within and across age groups. Nonetheless, these observed differences in child characteristics

    along with other unobserved differences (e.g., behavioral functioning) likely explain some of

    the differences in outcomes between the foster care groups.

    The integration of cost data into a child welfare outcome study brings its share of

    complications. It is important to note the context behind county foster care cost data, as

    counties have different practices that lead to different cost accounting. The major issue is the

    difficulty in capturing the administration costs to provide county foster care. These cost data

    are not available in Colorado Trails and it is up to the counties and the State to calculate and

    report this information. Thus, counties provided cost data from their own data systems for

    administration costs associated with county foster care placements. Furthermore, the cost

    comparisons have limitations because they are so heavily based on per episode length of stay.

    This study should be interpreted in light of the reality that foster care providers are only

    partially responsible for child outcomes, particularly length of stay, permanency at discharge,

    reentry, and subsequent placements. For example, child placement agencies typically do not

    provide services to biological parents who are directly responsible for whether and how quickly

    a child returns and/or stays home. Furthermore, caseworkers and courts often have the most

    influence in regard to where and when a child is placed. The generalizability of the findings to

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center ix

    the larger child welfare population in Colorado is somewhat problematic because of the lack of

    random sampling and the criteria used to select the sample. In addition, the findings are most

    applicable to larger child placement agencies and larger counties, as smaller agencies and

    counties have a lower percentage of the overall sample size. Even with these limitations, this

    study has important implications for child welfare practice, policy, and research in Colorado.

    Implications

    Based on the overall comparability in outcomes and costs between public and private

    foster care, the main implication of this study is that caseworkers can be confident in placing

    children in the most appropriate foster home setting regardless of the placement type. Meezan

    and McBeath (2008) believe this suggests the need for child welfare decisions to be driven by

    the needs of children and families, not by financial decisions. However, the variability in

    outcomes and costs by age group should also be considered in out-of-home placement

    decisions.

    This study added to the research on foster care by comparing administration costs for

    county and child placement agencies. Critics of private foster care claim that there is an

    increased administration cost due to greater demand for government oversight and increased

    labor cost per case in private agencies (Kamerman & Kahn 1998; Poole, 1999; Meezan &

    McBeath, 2008). The results from the cost comparison did not support this claim, as the daily

    administration rate was actually lower for private foster care placement in all age groups.

    Furthermore, the average cost per episode of care was lower for CPA foster care in all age

    groups. There is a need for more sophisticated outcome and cost analyses, so that comparisons

    among foster care placements can be made more precise.

    Another recommendation is for child placement agencies to have a more collaborative

    role in conducting child welfare research in Colorado. Specifically, CPAs should provide input on

    the measures used and the outcomes evaluated while providing data from their own systems to

    complement the administrative data collected in Colorado Trails. Lastly, child placement

    agencies are encouraged to conduct their own outcome and cost studies to contribute to the

    research-informed discussion about how best to design and implement a child welfare system

    that is responsive to the needs of children and families.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 1

    1. INTRODUCTION

    The Applied Research in Child Welfare (ARCH) Project is a partnership between Colorado

    State University (CSU), the Colorado Department of Human Services (CDHS), and the

    Departments of Human/Social Services in Adams, Arapahoe, Boulder, Broomfield, Denver,

    Douglas, El Paso, Jefferson, Larimer, and Pueblo counties. The purpose of the ARCH Project is to

    conduct applied research on child welfare interventions that informs social work practice and

    policy in Colorado. This report presents results from an outcome study of foster care in

    Colorado, which was conducted by the Social Work Research Center in the School of Social

    Work at CSU with funding from the ARCH Project.

    1.1. Study Rationale

    Facing unprecedented budgetary constraints and accountability requirements, child

    welfare agencies are hungry for research that analyzes the outcomes and costs of children

    placed in foster care. The steady growth in the privatization of foster care services has

    magnified this need, as policymakers also are looking to the research for guidance on how

    private and public foster care compare on outcomes and costs. However, there are few studies

    that have used experimental designs because randomly assigning children who require out-of-

    home placement to either private or public foster care has ethical implications for the safety

    and well-being of the children. As a result, non-experimental comparative and associational

    research designs are more commonly used in foster care research.

    1.2. Study Context

    According to child welfare practitioners, outcomes for children in foster care are

    multidimensional as they are based on the intersection of provider performance, caregiver

    involvement, county practice, and child characteristics. While they are not responsible for the

    initial difficulties that precipitate an out-of-home (OOH) placement, child placement agencies

    (CPA) and foster parents are expected to provide a safe and nurturing environment that helps

    children to mitigate some of these deficits. Therefore, this study is not designed to determine

    which placement type is the most appropriate, as that depends on the specific child, family,

    caseworker, and provider. The study is intended to supply caseworkers and providers with

    information on foster care outcomes and costs to assist them in making data-driven decisions

    about how to best serve children.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 2

    1.3. Research Questions

    The following three research questions guided the study:

    1. Are there statistically significant differences between children in CPA foster care and

    county foster care on demographic characteristics for the 0-5, 6-12, and 13-18 age

    groups?

    2. Are there statistically significant differences between children in CPA foster care and

    county foster care on permanency, safety, and subsequent involvement outcomes

    for the 0-5, 6-12, and 13-18 age groups?

    3. Are there statistically significant differences between children in CPA foster care and

    county foster care on average daily rates, administration rates, provider rates, and

    costs per episode of care for the 0-5, 6-12, and 13-18 age groups?

    2. Literature Review

    This literature review explores the recent research on privatization in child welfare and

    provides a foundation for the study’s comparison of public and private foster care.

    2.1. Child Welfare Privatization

    In the United States, the rise in child welfare privatization has been a reaction to both

    public outcries regarding the safety and well-being of children in care, and mounting concern

    about governmental agency accountability (Else, Groze, Hornby, Mirr, & Wheelock, 1992). The

    1997 Adoption and Safe Families Act (P.L.105-89) helped to fuel privatization as it mandated

    improvement in child welfare outcomes, with particular focus on improving permanency

    outcomes for children in foster care without compromising their safety and well being. Child

    welfare reform has led to new initiatives including varied types of purchase-of-service

    arrangements and community-based partnerships (Sellick, 2007; Yampolskaya, Paulson,

    Armstrong, Jordan, & Vargo, 2004), which entail diverse contractual arrangements (Kamerman

    & Kahn, 1998). Researchers estimate that at least half of child welfare services are being

    satisfied by private entities in the United States (McBeath & Meezan, 2009). However,

    Rosenthal (2000) points out that the level of dependence on privatization of child welfare

    services varies across the country with less reliance in the Western states.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 3

    2.2. Foster Care Privatization

    Advocates of foster care privatization argue that purchasing foster care services will

    improve outcomes with regard to efficiency, safety, and services including cost savings, more

    timely attainment of permanency outcomes, a reduction in foster care placements, less reentry

    into care, and the offering of a more diverse range of services (Meezan & McBeath, 2008).

    Opposing voices contend that the privatization of foster care services is associated with

    negative consequences and/or inconsequential effects that include cost ineffectiveness, as

    exhibited by exorbitant fees charged to governmental authorities (Sellick & Connolly, 2002).

    Additionally, it is posited that it is often more difficult for states to devise new policies and

    respond to changing needs under the auspices of privatization (Center for Public Policy

    Priorities, 2008). The vacillating debate between the pros and cons of purchasing foster care

    services from non-governmental organizations appears to be an ideological argument more

    than one based on evidence substantiated with empirical data (Kamerman & Kahn, 1998). For

    example, Unruh and Hodgkin (2004) indicate that both public and private foster care have

    realized varied degrees of improvement regarding specific child welfare outcomes.

    2.3. Private and Public Foster Care Outcomes

    There are limited studies comparing the outcomes between children in private and

    public foster care. In one study, Zullo (2002) examined reunification rates across six types of

    foster care placements that included a comparison between private and public foster care

    placements. Similar to an earlier study by George (1990), private foster care placement was

    associated with a reduction in the rate of children moving to permanency. Further analysis

    demonstrated no difference between the groups on rates of reentry to foster care placement

    for those children who transitioned to a permanent placement during the study time frame.

    Using the Smith and Lipsky (1993) typology of private foster care agencies, Zullo (2002) found

    that cases managed by newer and smaller community-based agencies realized permanency

    rates similar to public foster care agencies, while the older agencies were associated with

    significantly lower rates of permanency.

    In another study, Zullo (1998) collected ethnographic and case file data to compare

    private and public foster care outcomes in Milwaukee County, Wisconsin. The findings indicated

    lower permanency rates among children placed in older, private foster care agencies. However,

    children served by newer, private foster care agencies attained similar permanency rates as

    public foster care agencies. Public foster care agencies, as compared to older private foster care

    agencies, performed more consistently with regard to reunification of children and/or

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 4

    placement of children in adoptive homes. Rates of reentry were not different between the two

    types of foster care agencies. For cost, Zullo (1998) found that the direct per case cost was

    greater in private foster care agencies because private contracting demanded greater

    administrative overhead expenses associated with public agency oversight.

    Following the 1996 implementation of a pilot project to privatize child welfare services

    in Florida, statewide privatization was mandated in 1998 after the model produced notable

    results (Freundlich & Gerstenzang, 2003). Florida’s Community-Based Care (CBC) model

    combines the outsourcing of foster care and related services to private agencies utilizing a lead

    agency design (Paulson et al., 2003). Paulson et al. (2003) used a quasi-experimental cohort

    design to conduct a comparative evaluation between CBC counties and the remainder of the

    state. The results indicated that counties with earlier implementation dates performed better

    on permanency outcomes as compared to the remainder of the state. These counties also

    performed better than, or the same as, the remainder of the state on finalized adoption and

    reentry rates. However, Paulson et al. (2003) note that there is insufficient evidence to make

    any definitive conclusions regarding the effectiveness of the CBC model due to a few key factors

    during the study. First, the counties were at differing stages of implementation with regard to

    the privatized model. Second, there was a change in the administrative database system that

    impacted the outcome comparison.

    Overall, there is little independent, empirical evidence on this topic, and the research

    that has been conducted shows mixed findings in both outcome and cost differences between

    private foster care and public foster care.

    3. Methods

    This study employed a between-county comparative research design to analyze the

    outcomes and costs for children with a closed foster care placement in the 10 ARCH counties

    during the 2008-2010 fiscal years. Colorado state fiscal years run from July 1 through June 30 of

    the following year. The outcome data were aggregated across the three fiscal years while the

    cost data were collected from fiscal year 2008 (FY08) and fiscal year 2009 (FY09). The data

    collection techniques, study parameters, sample selection, variable definitions, and data

    analysis procedures are described in the methods section.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 5

    3.1. Data Collection

    The demographic, placement, outcome, and cost data for this study were collected from

    individual case records entered into Colorado Trails, which is Colorado’s Statewide Automated

    Child Welfare Information System (SACWIS). Colorado Trails is an online data management and

    analysis system used for child welfare case management documentation. To obtain the sample,

    the appropriate search terms, logic, and filters were used. A template to record and calculate

    the administration costs for county foster care was designed for this study. To collect the

    county-specific cost data, an Excel file with the template was emailed to each county. However,

    only 7 of the 10 counties were able to provide the required data.

    3.2. Study Parameters

    The main inclusion criteria were children with a closed service authorization at any time

    between 7/1/07 and 6/30/10 from these placement types: CPA Foster Care (Private Foster

    Care) and County Foster Care (Public Foster Care). If a foster care authorization had a governing

    body at any time during the placement, it was considered to be CPA foster care rather than

    county foster care. This timeframe was selected to generate a sufficient number of

    authorizations and to allow outcomes to be collected 12 months out for each episode of care.

    There were secondary exclusion criteria which further narrowed the sample. Service

    authorizations with the following placement leave reasons were excluded from the study: (1)

    opened in error, (2) payee wrong code, and (3) runaways and business office corrections when

    the authorization was opened and closed on the same day. If the placement leave reason was

    (a) same provider/change in service type and the new authorization was still open or closed

    after the study end date; (b) transfer to another county and the new authorization was still

    open or closed after the study end date; or (c) client left county/state and the new service

    authorization was still open or closed after the study end date or was open less than 30 days,

    the authorizations were removed from the sample.

    The unit of analysis for the outcome study was “episode of care” rather than service

    authorization. An episode of care was created by merging Colorado Trails service authorizations

    into a continuous span of care for a child with a given placement provider. Episodes of care

    were used in lieu of service authorizations for two primary reasons: (1) to account for nuances

    in data entry practices that lead to multiple service authorizations for a single, uninterrupted

    span of care; and (2) to account for short-term disruptions in placements spans caused by

    runaways or hospitalizations. Episodes of care were created according to the following criteria:

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 6

    a. Service authorizations with the same provider without a 30 day absence (for running away, hospitalization, or detention) were merged into a single episode of care.

    b. Service authorizations with a leave reason of same provider/same service or same provider/change in service type were merged with the subsequent service authorization into a single episode of care. The begin date on the subsequent service authorization was updated to the begin date of the first service authorization.

    c. Service authorizations with a leave reason of transfer to another county with the same provider were merged into a single episode of care for the county in which the authorization was transferred.

    d. Service authorizations closed and re-opened the same day with the same provider were merged into a single episode of care.

    3.3. Variable Definitions

    The variables included in the outcome study were defined as follows:

    1. Age at Placement – age of child when service authorization was opened.

    2. Age Group at Placement – age group of child when service authorization was opened (0-5, 6-12, or 13-18).

    3. Gender – gender of child with service authorization (female or male).

    4. Ethnicity Group – primary ethnicity of child with service authorization. For this study,

    the ethnicities were grouped into Caucasian or non-Caucasian (Hispanic, African American, Asian, and Native American).

    5. Program Area – current program area as of 6/30/10 for child with service

    authorization including Program Area 4 (juvenile delinquency and beyond control of parents), Program Area 5 (child abuse and neglect), and Program Area 6 (special needs are barrier to adoption). Because historical tracking of program area is not available in Colorado Trails, an authorization could have been opened under one program area but recorded under a different program area for this study.

    6. Placement Type – placement setting for which service authorization was opened

    (CPA foster care or county foster care).

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 7

    7. Prior Any Placement – any community or residential placement prior to episode of care.

    8. Per Episode Length of Stay – per episode length of stay without episodes of care with

    a discharge setting of runaway for each placement type.

    9. Discharge Setting – setting upon exit from placement including return home, APR/guardianship/other relatives, Division of Youth Corrections (DYC) commitment and detention, kinship foster care, kinship care, county foster care, CPA foster care, receiving homes, group home/group center, Therapeutic Residential Child Care Facility (TRCCF), Residential Child Care Facility (RCCF), emancipation, adoption, runaway, hospitalization/psychiatric hospitalization, Psychiatric Residential Treatment Facility (PRTF), and transfer to another county/state.

    3.4. Outcome Measures

    The outcomes for this study were defined as follows:

    1. Permanency at discharge – return home, permanent placement with relatives

    (guardianship), or finalized adoption at discharge from placement.

    2. Reentry – subsequent placement in OOH care within 12 months after returning home or having permanency placement with relatives (guardianship).

    3. Institutional Abuse – founded incident of abuse or neglect during placement by foster care provider. Only abuse or neglect assessments reported during the study timeframe were included.

    4. Number of placements – subsequent placements in any setting (community or

    residential), community setting (CPA foster, county foster, group home, or kinship foster), and residential setting (TRCCF or RCCF) within 12 months after discharge.

    5. Number of placement days – subsequent days in any setting, community setting,

    residential setting, DYC involvement (detention or commitment), and DYC commitment within 12 months after discharge.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 8

    3.5. Cost Measures

    The cost measures for this study were defined as follows:

    1. Average Daily Rate – total gross expenditures for placement type divided by total number of placement days based on cost data from FY08.

    2. Average Cost Per Episode of Care – average daily costs multiplied by per episode

    length of stay.

    3. Average Provider Rate – total gross expenditures paid to CPA providers and county foster care providers divided by total number of placement days based on cost data from FY08.

    4. Average Administration Rate (CPA foster care) – total gross expenditures paid to

    child placement agencies for administration costs divided by total number of placement days based on cost data from FY08.

    5. Average Administration Rate (County foster care) – estimate of administration costs

    associated with placing children in county foster care based on cost data from FY09.

    3.6. Sample Selection

    The study employed a non-probability sampling design that was both purposive and

    convenient. The theoretical population is all children in Colorado removed from the home for

    delinquency or maltreatment and placed in foster care. The accessible sample is children in

    Colorado representing the service authorizations that met the aforementioned criteria. The

    selected sample included children and adolescents ages 0-18 that were in the child welfare

    system under delinquency and/or dependency and neglect court actions in the ARCH counties

    (subsequently referred to as “children”). Children residing within the Division of Youth

    Corrections, children from out-of-state, and children with privately paid placements were not

    eligible for selection. Children with a Children's Habilitation Residential Program (CHRP)

    designation were eligible for selection in this study.

    The actual sample was comprised of duplicated cases, in that children who had multiple

    closed service authorizations during the study timeframe were counted for each placement

    type. Children who were placed in the same OOH setting but with different providers were

    counted multiple times, as were children who were placed with the same provider but at

    different times during the study timeframe. As displayed in Table 1, the overall sample size was

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 9

    11,661 with 6,740 episodes of care for CPA foster and 4,921 for county foster. Denver County

    had the highest percentage of episodes of care for both county foster and CPA foster at 25.3%

    and 28.6%, respectively.

    Table 1

    Episodes of Care from FY08-FY10 by County for Foster Care Placement Types

    County

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Adams 581 11.8 1,041 15.4 Arapahoe 819 16.6 620 9.2 Boulder 427 8.7 31 0.5 Broomfield 45 0.9 18 0.3 Denver 1,247 25.3 1,946 28.9 Douglas 57 1.2 15 0.2 El Paso 41 0.8 1,928 28.6 Jefferson 651 13.2 433 6.4 Larimer 577 11.7 107 1.6 Pueblo 476 9.7 601 8.9 Total 4,921 100.0 6,740 100.0

    3.7. Data Analysis

    The data were transmitted in Excel spreadsheets to the Social Work Research Center

    with the unique child and case identifiers removed. The demographic, placement, outcome,

    and cost data were entered into the IBM Statistical Package for Social Sciences (SPSS), checked

    for missing and incorrect data, and recoded into the appropriate variables. Descriptive and

    inferential statistics were generated from the data to answer the research questions posed in

    the study. Specifically, chi-square tests and t-tests were used to analyze the results for the

    foster care outcome comparisons.

    4. Demographic Characteristics Comparison

    The foster care groups were separated into three age groups (0-5, 6-12, and 13-18) and

    were compared on demographic characteristics including ethnicity, gender, program area, prior

    placement history, age at placement, and per episode length of stay for a duplicated count of

    children. The sample was divided by age group because of the diverse needs and services for

    children in the different age ranges. The comparison of demographic characteristics was

    intended to examine the equivalence between the two foster care groups. The demographics

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 10

    were chosen based on their accessibility and demonstrated relationships with case outcomes

    (see Section 5.1).

    4.1. 0-5 Age Group

    As displayed in Table 2, the CPA foster group had 53.4% of the episodes of care for the

    0-5 sample, while the county foster group had 46.6% of the episodes.

    Table 2

    Placement Type Frequencies between FY08-FY10 for 0-5 Age Group

    Placement Type Sample Size Percent

    CPA Foster 3,002 53.4 County Foster 2,617 46.6 Total 5,619 100.0

    The foster care groups for the 0-5 sample were compared on demographic

    characteristics. As displayed in Table 3, the foster care groups had statistically significant

    differences in ethnicity and current program area. Specifically, the CPA foster group had a

    higher percentage of non-Caucasian children (65.6%) than did the county foster group (61.9%),

    and a higher percentage of PA5 children (54.5%) than did the county foster group (48.1%).

    There were no statistically significant differences between the groups on gender and prior

    placement.

    The groups had a statistically significant difference (p < .05) in age at placement, as

    children in the CPA foster group (2.0 years) were older at placement than were children in the

    county foster group (1.7 years). Although not statistically significant, the county foster group

    (209 days) had a longer per episode length of stay than did the CPA foster group (200 days).

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 11

    Table 3

    Demographic Characteristics by Foster Care Placement Type for 0-5 Age Group

    Characteristic

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Ethnicity* Caucasian 998 38.1 1,003 34.4 Non-Caucasian 1,619 61.9 1,969 65.6

    Gender Female 1,237 47.3 1,400 46.6 Male 1,380 52.7 1,602 53.4

    Program Area* PA5 1,260 48.1 1,637 54.5 PA6 1,357 51.9 1,365 45.5

    Prior Any Placement Yes 1,349 51.5 1,512 50.4 No 1,268 48.5 1,490 49.6

    * Statistically significant difference (p < .05)

    4.2. 6-12 Age Group

    As displayed in Table 4, the CPA foster group had 59.2% of the episodes of care for the

    6-12 sample, while the county foster group had 40.8% of the episodes.

    Table 4

    Foster Care Frequencies between FY08-FY10 for 6-12 Age Group

    Placement Type Sample Size Percent

    CPA Foster 1,988 59.2 County Foster 1,371 40.8 Total 3,359 100.0

    The foster care groups for the 6-12 sample were compared on demographic

    characteristics. As displayed in Table 5, there were no statistically significant differences

    between the groups on ethnicity, gender, program area, and prior placement history.

    The groups had a statistically significant difference (p < .05) in per episode length of

    care, as the county foster group (262 days) had a longer mean per episode length of stay than

    did the CPA foster group (228 days). Although not statistically significant, children in the CPA

    foster group (8.9 years) were older at placement than were children in the county foster group

    (8.7 years).

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 12

    Table 5

    Demographic Characteristics by Foster Care Placement Type for 6-12 Age Group

    Characteristic

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Ethnicity Caucasian 514 37.5 777 39.1 Non-Caucasian 857 62.5 1,211 60.9

    Gender Female 639 46.6 905 45.5 Male 732 53.4 1,083 54.5

    Program Area PA5 716 52.2 1,034 52.0 Non-PA5 655 47.8 954 48.0

    Prior Any Placement Yes 888 64.8 1,311 65.9 No 483 35.2 677 34.1

    4.3. 13-18 Age Group

    As displayed in Table 6, the CPA foster group had 65.2% of the episodes of care for the

    6-12 sample, while the county foster group had 34.8% of the episodes.

    Table 6

    Foster Care Frequencies between FY08-FY10 for 13-18 Age Group

    Service Type Sample Size Percent

    CPA Foster 1,750 65.2 County Foster 933 34.8 Total 2,683 100.0

    The foster care groups for the 13-18 sample were compared on demographic

    characteristics. As displayed in Table 7, the foster care groups had statistically significant

    differences in ethnicity and prior placement history. Specifically, the CPA foster group had a

    higher percentage of non-Caucasian children (60.1%) than did the county foster group (55.3%),

    and a higher percentage of episodes of care with any prior placement (83.0%) than did the

    county foster group (76.8%). There were no statistically significant differences on gender or

    current program area.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 13

    The groups had a statistically significant difference (p < .05) in age at placement, as

    children in the CPA foster group (15.2 years) were older at placement than were children in the

    county foster group (15.0 years). Although not statistically significant, the CPA foster group (239

    days) had a longer per episode length of stay than did the county foster group (230 days).

    Table 7

    Demographic Characteristics by Foster Care Placement Type for 13-18 Age Group

    Characteristic

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Ethnicity* Caucasian 417 44.7 699 39.9 Non-Caucasian 516 55.3 1,051 60.1

    Gender Female 491 52.6 965 55.1 Male 442 47.4 785 44.9

    Program Area PA4 448 48.0 879 50.2 Non-PA4 485 52.0 871 49.8

    Prior Any Placement* Yes 717 76.8 1,452 83.0 No 216 23.2 298 17.0

    * Statistically significant difference (p < .05)

    5. Outcome Comparison

    The demographic groups were compared overall and within each age range on

    outcomes including permanency at discharge, reentry, institutional abuse, subsequent any

    placements, any placement days, community placements, community placement days,

    residential placements, residential placement days, DYC involvement days, and DYC

    commitment days. This outcome comparison was intended to identify demographic

    characteristics associated with these outcomes so the equivalence of the foster care groups

    could be explored. The foster care groups within each age range then were compared on the

    study outcomes.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 14

    5.1. Demographic Characteristics

    For the overall sample, there were statistically significant differences (p < .05) between

    the ethnicity groups (Caucasian and non-Caucasian) on permanency at discharge, subsequent

    any placements, any placement days, community days, residential placements, residential days,

    DYC involvement days, and DYC commitment days. However, these differences were mixed as

    the non-Caucasian group had greater outcomes for the residential and DYC placement

    outcomes. There were statistically significant differences (p < .05) between the prior placement

    groups on permanency at discharge, reentry, subsequent any placements, any placement days,

    community placements, residential placements, DYC involvement days, and DYC commitment

    days. The group with no prior placements had greater outcomes for all of these variables. There

    were statistically significant differences (p < .05) between the gender groups on subsequent any

    placement days, residential placements, DYC involvement days, and DYC commitment days. The

    female group had greater outcomes for all of these variables.

    For the 0-5 sample, there were statistically significant differences (p < .05) between the

    program area groups (PA5 and PA6) on permanency at discharge, reentry, subsequent any

    placements and placement days with the PA5 group having greater outcomes for all of these

    variables. For the 6-12 sample, there were statistically significant differences (p < .05) between

    the program area groups (PA5 and non-PA5) on permanency at discharge, reentry, institutional

    abuse, subsequent any placements, any placement days, community placements, community

    placement days, and residential placements with the PA5 group having greater outcomes for all

    of these variables. For the 13-18 sample, there were statistically significant differences (p < .05)

    between the program area groups (PA4 and non-PA4) on permanency at discharge, reentry,

    subsequent any placements, any placement days, community placements, community

    placement days residential placements, DYC involvement days, and DYC commitment days.

    However, these differences were somewhat mixed as the PA4 group had greater outcomes for

    the any placement days and community placement days.

    5.2. 0-5 Age Group

    As displayed in Table 8, there was a statistically significant difference between the

    groups on permanency at discharge with the county foster group (60.0%) having higher rates of

    permanency at discharge than did CPA foster group (53.3%). There was no statistically

    significant difference between the groups on reentry as the county foster group had a 16.6%

    reentry rate and the CPA foster group had a 19.4% reentry rate. There was a statistically

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 15

    significant difference between the groups on institutional abuse with the county foster group

    (0.7%) having lower institutional abuse rates than did the CPA foster group (1.5%).

    Table 8

    Permanency and Safety Outcomes by Foster Care Placement Type for 0-5 Age Group

    Outcome

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Permanency at Discharge* Permanency 1,571 60.0 1,600 53.3 No Permanency 1,046 40.0 1,402 46.7

    Reentry Reentered 144 16.6 222 19.4 Did not Reenter 721 83.4 920 80.6

    Institutional Abuse* Abused 18 0.7 45 1.5 Not Abused 2,559 99.3 2,957 98.5

    * Statistically significant difference (p < .05)

    There were statistically significant differences between the groups on the placement

    outcomes for any placements and any placement days within 12 months of exit. As displayed in

    Table 9, county foster (.68 placements) had fewer subsequent placements of any type than did

    CPA foster (.84 placements). County foster (120 days) also had fewer subsequent any

    placement days than did CPA foster (144 days). Because there were so few residential

    placements for the 0-5 age group, the foster care groups were not compared on subsequent

    community placements and placement days within 12 months of exit because the results were

    almost identical to the subsequent any placement number and days results.

    Table 9

    Subsequent Placements Outcomes by Foster Care Placement Type for 0-5 Age Group

    Outcome

    County Foster Mean

    CPA Foster Mean

    Any Placements within 12 Months of Exit* .68 .84 Any Placement Days within 12 Months of Exit* 120 144 * Statistically significant difference (p < .05)

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 16

    5.3. 6-12 Age Group

    As displayed in Table 10, there were no statistically significant differences between the

    groups on permanency at discharge, reentry, and institutional abuse. Although not statistically

    significant, the county foster group (45.4%) had a higher permanency at discharge rate than did

    the CPA foster group (43.6%). Although not statistically significant, the county foster group

    (14.1%) had a lower reentry rate than did the CPA foster group (16.7%). Although not

    statistically significant, the CPA foster group (0.8%) had a lower institutional rate than did the

    county foster group (1.4%).

    Table 10

    Permanency and Safety Outcomes by Foster Care Placement Type for 6-12 Age Group

    Outcome

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Permanency at Discharge Permanency 622 45.4 866 43.6 No Permanency 749 54.6 1,122 56.4

    Reentry Reentered 63 14.1 112 16.7 Did not Reenter 383 85.9 558 83.3

    Institutional Abuse Abused 19 1.4 16 0.8 Not Abused 1,352 98.6 1,972 99.2

    There were statistically significant differences between the groups on the placement

    outcomes for any placements, any placement days, community placements, and community

    placement days within 12 months of exit. As displayed in Table 11, the county foster group

    (1.10 placements) had fewer subsequent placements of any type than did the CPA foster group

    (1.19 placements), and fewer subsequent placement days of any type (165 days) than did the

    CPA foster group (185 days). The county foster group (.92 placements) also had fewer

    subsequent community placements than did the CPA foster group (1.02 placements), and fewer

    subsequent community placement days (140 days) than did the CPA foster group (157 days).

    There were no statistically significant differences between the groups on the placement

    outcomes for residential placements, residential placement days, DYC involvement days, and

    DYC commitment days within 12 months of exit. Although not statistically significant, the CPA

    foster group (.17 placements) had fewer subsequent residential placements than did the

    county foster group (.18 placements). Although not statistically significant, the county foster

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 17

    group (25 days) had fewer subsequent residential placement days than did the CPA foster

    group (28 days), while the CPA foster care group (0 days) had fewer subsequent DYC

    commitment days than did the county foster group (1 day).

    Table 11

    Subsequent Placement Outcomes by Foster Care Placement Type for 6-12 Age Group

    Outcome

    County Foster Mean

    CPA Foster Mean

    Any Placements within 12 Months of Exit* 1.10 1.19 Any Placement Days within 12 Months of Exit* 165 185 Community Placements within 12 Months of Exit* .92 1.02 Community Placement Days within 12 Months of Exit* 140 157 Residential Placements within 12 Months of Exit .18 .17 Residential Placement Days within 12 Months of Exit 25 28 DYC Involvement Days within 12 Months of Exit 1 1 DYC Commitment Days within 12 Months of Exit 1 0 * Statistically significant difference (p < .05)

    5.4. 13-18 Age Group

    As displayed in Table 12, there was a statistically significant difference between the

    groups on permanency at discharge with the county foster group (26.7%) having higher rates of

    permanency at discharge than did the CPA foster group (22.9%). Although not statistically

    significant, the county foster group (17.3%) had a lower reentry rate than did the CPA foster

    group (21.5%). There was a statistically significant difference between the groups on

    institutional abuse rates with the CPA foster group (0.7%) having a lower rates of institutional

    abuse than did the county foster group (1.9%).

    There were statistically significant differences between the groups on the placement

    outcomes for any placement days and community placements within 12 months of exit. As

    displayed in Table 13, the county foster group (153 days) had fewer subsequent any placements

    days than did the CPA foster group (166 days), and fewer subsequent community placements

    (.84 placements) than did the CPA foster group (.93 placements). There were no statistically

    significant differences between the groups on the placement outcomes for subsequent any

    placements, community placement days, residential placements, residential placement days,

    DYC involvement days, and DYC commitment days within 12 months of exit. Although not

    statistically significant, the county foster group (1.29 placements) had fewer subsequent any

    placements than did the CPA foster group (1.37 placements), and fewer subsequent community

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 18

    placement days (108 days) than did the CPA foster group (119 days). Although not statistically

    significant, the CPA foster group (.44 placements) had fewer subsequent residential placements

    than did the county foster group (.45 placements), while the county foster care group (45 days)

    had fewer residential placement days than did the CPA foster group (47 days). Although not

    statistically significant, the CPA foster group (15 days) had fewer subsequent DYC involvement

    days than did the county foster group (18 days), and fewer subsequent DYC commitment days

    (12 days) than did the county foster group (15 days).

    Table 12

    Permanency and Safety Outcomes by Foster Care Placement Types for 13-18 Age Group

    Outcome

    County Foster Frequency Percent

    CPA Foster Frequency Percent

    Permanency at Discharge* Permanency 249 26.7 401 22.9 No Permanency 684 73.3 1,349 77.1

    Reentry Reentered 39 17.3 76 21.5 Did not Reenter 186 82.7 278 78.5

    Institutional Abuse* Abused 18 1.9 12 0.7 Not Abused 915 98.1 1,738 99.3

    * Statistically significant difference (p < .05)

    Table 13

    Subsequent Placement Outcomes by Foster Care Placement Type for 13-18 Age Group

    Outcome

    County Foster Mean

    CPA Foster Mean

    Any Placements within 12 Months of Exit 1.29 1.37 Any Placement Days within 12 Months of Exit* 153 166 Community Placements within 12 Months of Exit* .84 .93 Community Placement Days within 12 Months of Exit 108 119 Residential Placements within 12 Months of Exit .45 .44 Residential Placement Days within 12 Months of Exit 45 47 DYC Involvement Days within 12 Months of Exit 18 15 DYC Commitment Days within 12 Months of Exit 15 12 * Statistically significant difference (p < .05)

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 19

    6. Cost Comparison

    The foster care groups within each age range were compared on average daily rate,

    average daily administration rate, average daily provider rate, and average cost per episode of

    care. The average daily rate for CPA foster care is based on an administration rate paid to the

    governing agency and a provider rate paid to the specific provider. The administration costs are

    a combination of administrative services (e.g., case management) and administrative

    maintenance (e.g., salaries/benefits, occupancy, insurance, mileage, etc.). However, counties

    have different practices for how they record these costs, so the administration rate was not

    separated out by service and maintenance rates for CPA foster care cost calculations.

    The average daily rate for county foster care is based on administration costs for county

    agencies to operate the foster care unit and a provider rate paid to the specific provider. The

    administration costs are a combination of administrative services (e.g., personnel salary and

    benefits for case management including administrators, supervisors, and caseworkers) and

    administrative maintenance. The administrative maintenance costs collected for this study

    include contract payments for home studies, operational costs for office supplies, document

    copies, computers, postage, foster parent training incentives, telephone charges, special

    services for foster parents, mileage, training, advertising, and promotional items, and

    interdepartmental direct billing costs for accounting, legal fees, human resources, IT training,

    rent, security and facilities.

    To calculate an average cost per day per child, the total administration costs were

    divided by the total placement days (as derived from FY09 ADP numbers from Colorado Trails).

    However, the county foster care administration costs were not calculated for the different age

    groups. Broomfield, Douglas, and El Paso were not included because administration cost data

    were not available. However, these three counties only accounted for three percent of the total

    ADP for the ARCH counties for the 0-5, 6-12, and 13-18 age groups.

    6.1. O-5 Age Group

    The average cost per episode for CPA foster care and county foster care was compared

    for the 0-5 age group. As displayed in Table 14, the weighted average daily rate for CPA foster

    care placements for the 0-5 age group across the 10 ARCH counties was $40.94 with an average

    daily administration rate of $16.56, an average daily provider rate of $24.38, and an overall ADP

    of 490 children.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 20

    Table 14

    CPA Average Daily Rate, Administration Rate, and Provider Rate by County for 0-5 Age Group

    County Average Daily Rate

    Average Daily Administration

    Rate

    Average Daily Provider Rate

    Average Daily Placement

    Adams 44.71 16.89 27.82 59.6 Arapahoe 41.39 15.04 26.36 50.0 Boulder 56.24 43.97 12.27 0.6 Broomfield 53.00 43.76 9.25 1.0 Denver 43.06 17.05 26.01 166.1 Douglas 59.34 20.44 38.90 0.1 El Paso 36.46 15.79* 20.67 153.0 Jefferson 41.37 15.17 26.21 31.4 Larimer 66.25 48.14 18.12 2.2 Pueblo 40.75 17.49 23.26 25.8 Total 40.94 16.56 24.38 489.8 *The administration rate for El Paso County does not include the $10 daily rate paid for casework services

    As displayed in Table 15, the weighted average daily rate for county foster care

    placements for the 0-5 age group across the seven counties that reported administrative costs

    was $42.13 with an average daily administration rate of $22.26, an average daily provider rate

    of $19.87, and an overall ADP of 408 children.

    Table 15

    County Foster Care Average Daily Rate, Administration Rate, and Provider Rate by County for

    0-5 Age Group

    County Average Daily Rate

    Average Daily Administration

    Rate

    Average Daily Provider Rate

    Average Daily Placement

    Adams 39.51 22.82 16.69 64.5 Arapahoe 38.93 20.86 18.07 67.3 Boulder 59.01 42.91 16.10 43.1 Denver 42.36 18.04 24.32 105.3 Jefferson 39.58 18.21 21.37 59.2 Larimer 36.53 20.76 15.77 40.8 Pueblo 42.56 19.18 23.38 28.1 Total 42.13 22.26 19.87 408.3

    As displayed in Table 16, the average cost per episode of care was $617 lower for CPA

    foster as compared to county foster for the 0-5 age group. Based on a per episode length of

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 21

    stay of 200 days and an average daily rate of $40.94, CPA foster care averaged $8,188 per

    episode of care, as compared with $8,805 per episode of care for county foster care, which was

    based on a per episode length of stay of 209 days and an average daily rate of $42.13.

    Table 16

    Foster Care Comparison for Average Daily Rate, Administration Rate, Provider Rate, and

    Average Cost per Episode of Care for 0-5 Age Group

    Service Type Average Daily Rate

    Average Daily

    Admin. Rate

    Average Daily

    Provider Rate

    Per Episode Length of

    Stay

    Average Cost per

    Episode of Care

    County Foster 42.13 22.26 19.87 209 $8,805 CPA Foster 40.94 16.56 24.38 200 $8,188

    6.2. 6-12 Age Group

    As displayed in Table 17, the weighted average daily rate for CPA foster care placements

    for the 6-12 age group across the 10 ARCH counties was $48.19 with an average daily

    administration rate of $17.59, an average daily provider rate of $30.60, and an overall ADP of

    352 children.

    Table 17

    CPA Average Daily Rate, Administration Rate, and Provider Rate by County for 6-12 Age Group

    County Average Daily Rate

    Average Daily Administration

    Rate

    Average Daily Provider Rate

    Average Daily Placement

    Adams 53.47 17.84 35.62 50.6 Arapahoe 43.92 15.59 28.33 26.6 Boulder 108.20 93.68 14.52 1.7 Broomfield 48.30 23.14 25.16 4.2 Denver 48.32 17.90 30.42 109.5 Douglas 60.00 19.66 40.34 1.0 El Paso 41.90 15.46* 26.43 94.5 Jefferson 54.70 16.98 37.72 23.8 Larimer 57.20 26.31 30.89 6.4 Pueblo 51.20 17.92 33.28 33.7 Total 48.19 17.59 30.60 352 *The administration rate for El Paso County does not include the $10 daily rate paid for casework services

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 22

    As displayed in Table 18, the weighted average daily rate for county foster care

    placements for the 6-12 age group across the seven counties that reported administrative costs

    was $49.54 with an average daily administration rate of $21.16, an average daily provider rate

    of $28.38, and an overall ADP of 230 children.

    Table 18

    County Foster Care Average Daily Rate, Administration Rate, and Provider Rate by County for

    6-12 Age Group

    County Average Daily Rate

    Average Daily Administration

    Rate

    Average Daily Provider Rate

    Average Daily Placement

    Adams 49.65 22.82 26.83 18.1 Arapahoe 53.33 20.86 32.47 31.6 Boulder 60.94 42.91 18.03 17.6 Denver 47.04 18.04 29.00 79.7 Jefferson 47.45 18.21 29.24 36.7 Larimer 46.99 20.76 26.23 27.6 Pueblo 50.81 19.18 31.63 18.5 Total 49.54 21.16 28.38 229.8

    As displayed in Table 19, the average cost per episode of care was $1,992 lower for CPA

    foster as compared to county foster for the 6-12 age group. Based on a per episode length of

    stay of 228 days and an average daily rate of $48.19, CPA foster care averaged $10,987 per

    episode of care, as compared with $12,979 per episode of care for county foster care, which

    was based on a per episode length of stay of 262 days and an average daily rate of $49.54.

    Table 19

    Foster Care Comparison for Average Daily Rate, Administration Rate, Provider Rate, and

    Average Cost per Episode of Care for 6-12 Age Group

    Service Type Average Daily Rate

    Average Daily

    Admin. Rate

    Average Daily

    Provider Rate

    Full per Episode

    Length of Stay

    Average Cost per

    Episode of Care

    County Foster 49.54 21.16 28.38 262 $12,979 CPA Foster 48.19 17.59 30.60 228 $10,987

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 23

    6.3. 13-18 Age Group

    As displayed in Table 20, the weighted average daily rate for CPA foster care placements

    for the 13-18 age group across the 10 ARCH counties was $52.83 with an average daily

    administration rate of $18.99, an average daily provider rate of $33.83, and an overall ADP of

    652 children.

    Table 20

    CPA Average Daily Rate, Administration Rate, and Provider Rate by County for 13-18 Age Group

    County Average Daily Rate

    Average Daily Administration

    Rate

    Average Daily Provider Rate

    Average Daily Placement

    Adams 55.11 18.81 36.30 100.7 Arapahoe 55.95 18.70 37.25 54.9 Boulder 62.64 46.30 16.34 12.8 Broomfield 53.75 19.07 34.69 3.4 Denver 49.57 18.52 31.05 181.3 Douglas 49.20 27.83 21.37 6.5 El Paso 54.15 16.65* 37.50 130.9 Jefferson 56.84 18.51 38.33 48.0 Larimer 47.76 20.10 27.67 23.4 Pueblo 51.06 19.19 31.87 90.3 Total 52.83 18.99 33.83 652.2 *The administration rate for El Paso County does not include the $10 daily rate paid for casework services

    As displayed in Table 21, the average daily rate for county foster care placements for the

    13-18 age group across the seven counties that reported administrative costs was $59.52 with

    an average daily administration rate of $20.51, an average daily provider rate of $39.01, and an

    overall ADP of 329 children.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 24

    Table 21

    County Foster Care Average Daily Rate, Administration Rate, and Provider Rate by County for

    13-18 Age Group

    County Average Daily Rate

    Average Daily Administration

    Rate

    Average Daily Provider Rate

    Average Daily Placement

    Adams 61.51 22.82 38.69 24.1 Arapahoe 69.88 20.86 49.02 83.2 Boulder 82.12 42.91 39.21 13.3 Denver 48.29 18.04 30.25 102.9 Jefferson 68.98 18.21 50.77 44.5 Larimer 51.41 20.76 30.65 35.1 Pueblo 52.12 19.18 32.94 25.5 Total 59.52 20.51 39.01 328.6

    As displayed in Table 22, the average cost per episode of care was $1,116 lower for CPA

    foster as compared to county foster for the 13-18 age group. Based on a per episode length of

    stay of 238 days and an average daily rate of $52.83, CPA foster care averaged $12,574 per

    episode of care, as compared with $13,690 per episode of care for county foster care, which

    was based on a per episode length of stay of 230 days and an average daily rate of $59.52.

    Table 22

    Foster Care Comparison for Average Daily Rate, Administration Rate, Provider Rate, and

    Average Cost per Episode of Care for 13-18 Age Group

    Service Type Average Daily Rate

    Average Daily

    Admin. Rate

    Average Daily

    Provider Rate

    Full per Episode

    Length of Stay

    Average Cost per

    Episode of Care

    County Foster 59.52 21.25 39.01 230 $13,690 CPA Foster 52.83 18.99 33.83 238 $12,574

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 25

    7. Discussion

    The following discussion presents the conclusions, limitations, and implications of the

    outcome and cost findings for the comparison of public and private foster care in Colorado.

    7.1. Conclusions

    Similar to the studies cited in the literature, the results for the outcome and cost

    comparisons for public and private foster care were mixed and varied by age group. Overall, the

    outcomes for county foster care were greater than or comparable to the outcomes for CPA

    foster care, but CPA foster care had lower costs per episode of care than did county foster care.

    Furthermore, there were differences in client acuity but no definitive sense of which placement

    type was serving higher risk cases. It should be noted that the outcome results were very

    similar to an earlier analysis of data from fiscal years 2004-2007. This indicates some stability

    across a six year period, which gives more credence to these findings.

    For the 0-5 age group, there were statistically significant differences between the foster

    care groups on permanency at discharge, institutional abuse, subsequent any placements, and

    subsequent any placement days, with county foster care having an advantage for these

    outcomes. However, there was no difference between the foster care groups on reentry. The

    cost per episode of care was more expensive for county foster care, with county foster care

    having higher administration costs and CPA foster care having higher provider costs. The higher

    provider costs suggest that children in the 0-5 age range have higher needs as measured by the

    Needs Based Care (NBC) instrument which sets CPA foster parent rates. Furthermore, the CPA

    foster group was comprised of a higher percentage of non-Caucasian children. However, the

    county foster group was comprised of a higher percentage of children in Program Area 6. As a

    result, the difference in client acuity between the foster care groups for this age range is

    unclear.

    For the 6-12 age group, there were statistically significant differences between the

    foster care groups on subsequent any placements, subsequent any placement days, subsequent

    community placements, and subsequent community placement days, with county foster care

    having an advantage for these outcomes. There were no differences between the groups on

    permanency at discharge, reentry, institutional abuse, subsequent residential placements and

    placement days, and subsequent DYC involvement days and DYC commitment days. The cost

    per episode of care was more expensive for county foster care, with county foster care having

    higher administration costs and CPA foster care having higher provider rates. The two foster

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 26

    care groups appeared to be equivalent for the 6-12 age group, as there were no demographic

    differences.

    For the 13-18 age group, there were statistically significant differences between the

    foster care groups on permanency at discharge, subsequent any placement days, and

    subsequent community placements, with county foster care having an advantage for these

    outcomes. There also was a statistically significant difference for institutional abuse with CPA

    foster care having an advantage for this outcome. There were no differences between the

    groups on reentry, subsequent any placements, subsequent community placement days,

    subsequent residential placements and residential placement days, and subsequent DYC

    involvement days and DYC commitment days. The cost per episode of care was more expensive

    for county foster care, with CPA foster care having higher administration costs and county

    foster care having higher provider costs. However, CPA foster care appeared to serve higher

    risk cases based on ethnicity and prior placement.

    7.2. Limitations

    The outcome study faced numerous methodological challenges common to most

    research on child welfare interventions using comparative research designs and secondary

    data. The most notable limitation is the differences in populations served by the private and

    public foster care agencies due to client acuity. However, these differences were not consistent

    within and across age groups. Nonetheless, these observed differences in child characteristics

    along with other unobserved differences (e.g., behavioral functioning) likely explain some of

    the differences in outcomes between the foster care groups. Although the Colorado Client

    Assessment Record is used as a measure of initial functioning for children placed in residential

    settings, a pre-placement behavioral measure is needed for children placed in foster care

    settings.

    The integration of cost data into a child welfare outcome study brings its share of

    complications. It is important to note the context behind county foster care cost data, as

    counties have different practices that lead to different cost accounting. The major issue is the

    difficulty in capturing the administration costs to provide county foster care. These cost data

    are not available in Colorado Trails and it is up to the counties and the State to calculate and

    report this information. Thus, counties provided cost data from their own data systems for

    administration costs associated with county foster care placements. One limitation of this

    approach was that the daily administration rate for county foster care could not be calculated

    by age group because the cost data were not collected in that way. Thus, the same daily

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 27

    administration rate was used for each age group. This yielded a somewhat imprecise estimate

    in that administration costs are typically less expensive for younger children and more

    expensive for older children. The difference in administration rates between the 0-5 and 6-12

    age groups for CPA foster care was 5.9%, while the difference between the 6-12 and 13-18 age

    groups was 8.0%. As a result, the average cost per episode of care for county foster care may be

    slightly overestimated for the 0-5 age group and slightly underestimated for the 13-18 age

    group. Lastly, the cost comparisons have limitations because they are so heavily based on per

    episode length of stay.

    This study should be interpreted in light of the reality that foster care providers are only

    partially responsible for child outcomes, particularly length of stay, permanency at discharge,

    reentry, and subsequent placements. For example, child placement agencies typically do not

    provide services to biological parents who are directly responsible for whether and how quickly

    a child returns and/or stays home. Furthermore, caseworkers and courts often have the most

    influence in regard to where and when a child is placed.

    The generalizability of the findings to the larger child welfare population in Colorado is

    somewhat problematic because of the lack of random sampling and the criteria used to select

    the sample. In addition, the findings are most applicable to larger child placement agencies and

    larger counties, as smaller agencies and counties have a lower percentage of the overall sample

    size. Even with these limitations, this study has important implications for child welfare

    practice, policy, and research in Colorado.

    7.3. Implications

    Based on the overall comparability in outcomes and costs between public and private

    foster care, the main implication of this study is that caseworkers can be confident in placing

    children in the most appropriate foster home setting regardless of the placement type. Meezan

    and McBeath (2008) believe this suggests the need for child welfare decisions to be driven by

    the needs of children and families, not by financial decisions. However, the variability in

    demographics, outcomes, and costs by age group should also be considered in out-of-home

    placement decisions.

    Some researchers suggest (e.g., Fein & Staff, 1991) that establishing a close working

    relationship between private and public child welfare agencies can be critical to positive child

    welfare outcomes. The literature indicates that agencies can increase retention rates, have

    more satisfied foster parents, and improve child functioning within the system by providing

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 28

    better financial support to foster parents (Chamberlain, Moreland, & Reid, 1992; Lowry, 2004).

    Chamberlain et al. (1992) found that additional training and support as well as a small increase

    in the monthly stipend minimized dropout rates because the foster parents felt more valued.

    However, increasing foster care payments without adding additional support and training may

    not produce enough buy-in from foster parents (Chamberlain et al., 1992).

    This study added to the research on foster care by comparing administration costs for

    county and child placement agencies. Critics of private foster care claim that there is an

    increased administration cost due to greater demand for government oversight and increased

    labor cost per case in private agencies (Kamerman & Kahn 1998; Poole, 1999; Meezan &

    McBeath, 2008). The results from the cost comparison did not support this claim, as the daily

    administration rate was actually lower for private foster care placement in all age groups.

    Furthermore, the average cost per episode of care was lower for CPA foster care in all age

    groups. One caveat is that the cost data were collected from FY08 and FY09, so it is possible

    that the cost structure for foster care placements in some Colorado counties has shifted during

    the past three years. This speaks to the need for ongoing evaluation of foster care outcomes

    and costs. Furthermore, there is a need for more sophisticated outcome and cost analyses, so

    that comparisons among foster care placements can be made more precise.

    Another recommendation is for child placement agencies to have a more collaborative

    role in conducting child welfare research in Colorado. Specifically, CPAs should provide input on

    the measures used and the outcomes evaluated while providing data from their own systems to

    complement the administrative data collected in Colorado Trails. For example, it might be

    useful to collect agency leave reasons and compare them with county leave reasons. Lastly,

    child placement agencies are encouraged to conduct their own outcome and cost studies to

    contribute to the research-informed discussion about how best to design and implement a child

    welfare system that is responsive to the needs of children and families.

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 29

    8. References

    Adoption and Safe Families Act - P.L. 105-89 (1997).

    Center for Public Policy Priorities. (2008). Drawing the line between public and private responsibility in child welfare: The Texas debate. Retrieved from http://www.cppp.org/files/4/CPSreportweb.pdf.

    Chamberlain, P., Moreland, S., & Reid, K. (1992). Enhanced services and stipends for

    foster parents: Effects on retention rates and outcomes for children. Child Welfare, 71, 387-401.

    Else, J. F., Groze, V., Hornby, H., Mirr, R. K., & Wheelock, J. (1992). Performance-based contracting: The case of residential foster care. Child Welfare, 71, 513-526.

    Fein. E., & Staff, J. (1991). Implementing reunification services. Families in Society: The Journal

    of Contemporary Human Services, 72, 335-343. Freundlich, M., & Gerstenzang, S. (2003). An assessment of the privatization of child welfare

    services: Challenges and success. Washington, DC: Child Welfare League of America.

    George, R. M. (1990). The reunification process in substitute care. Social Service Review, 64, 422-457.

    Kamerman, S. B., & Kahn, A. J. (1998). Privatization, contracting, and reform of child and family

    social services. The Finance Project. Lowry, M. R. (2004). Putting teeth into ASFA: The need for statutory minimum standards.

    Children and Youth Services Review, 26, 1021-1031. McBeath, B., & Meezan, W. (2008). Market-based disparities in foster care service provision.

    Research on Social Work Practice, 18, 27-41. McBeath, B., & Meezan, W. (2009). Interorganizational disparities in foster care service

    provision. Children and Youth Services Review, 31, 513‐525. doi:10.1016/j.childyouth.2008.10.007

    Meezan, W., & McBeath, B. (2008). Market-based disparities in foster care outcomes. Children

    and Youth Services Review, 30, 388-406.

    http://www.cppp.org/files/4/CPSreportweb.pdf

  • Comparison of Public and Private Foster Care in Colorado | Social Work Research Center 30

    Paulson, R. I., Armstrong, M., Fitzpatrick, J., Jordon, N., Kershaw, M.