State Innovations in Child Welfare Financing

Submitted to:
U.S. Department of Health and Human Services
Assistant Secretary for Planning and Evaluation

Submitted by:
Westat and Chapin Hall Center for Children

"

Executive Summary

Recent fiscal reform initiatives have attempted to address some of the seemingly chronic problems of the child welfare system in the United States. This report describes how states are implementing fiscal reforms to contain costs or improve system performance. It also identifies issues that the implementation of fiscal reforms faces and describes how well fiscal reforms appear to be working. Many of these reforms are based on the managed care model that has been used in medicine for the past 30 years, while other reforms use approaches such as the privatization of services, performance contracting, and integrated funding.

Three general findings emerged from this review.

1. Despite a concern that focusing on fiscal aspects of child welfare systems will lessen the focus on children and families, that does not appear to be what happened in the states reviewed. An integral part of the initiatives seems to be a push to do things better for the children and families served, or at least not to allow things to get worse for them when money is being saved.
2. Available evidence does not support a conclusion that the fiscal reforms have had a major direct impact on outcomes, although impressionistic and anecdotal information points to some efficiencies and improvements in permanency outcomes. However, the fiscal reforms frequently encouraged agencies to develop creative and innovative approaches, which are improved upon over time, and changes in outcomes may not appear until much later.
3. Ongoing problems in child welfare are not necessarily eliminated by changes in fiscal relationships. Instead, these new relationships often highlight aspects of the system that need to be more clearly defined. For example, in establishing payment rates and incentives, a state must clearly identify what it wants to obtain and what is needed to obtain it: attaining basic safety for children requires a different set of services than does achieving improvements in longstanding situations, and the goal will greatly affect the design and costs of services.

The report describes 23 initiatives in 22 states. These initiatives focus on altering the financial relationships between public child welfare agencies (states or counties) and private organizations with which they contract for services (here called “contractors”). The altered relationships presumably lead to greater efficiency in the use of resources, improved services, and better outcomes for children and families. The motivation for experimenting with reforms such as managed care is a belief that the traditional mechanisms of payment for such services, fee for services, results in wastefulness of resources and suboptimal quality of service to families. Many believe that fee-for-service arrangements provide incentives for using higher levels of care than required and for extending care for longer than necessary. In some states, it appears (at least initially) that cost savings were achieved when alternate arrangements (such as case rates and performance contracting) were implemented, and they were achieved without declines in permanency outcomes.

The most universally acclaimed feature of the fiscal reform efforts reviewed is the flexibility they afford in the provision of service. Unlike traditional categorical approaches to funding, contractors are given the freedom to deliver a wide range of services and move children more freely among living arrangements. Funding follows the child rather than the service. Decisionmaking about services is, therefore, freed up, presumably to provide more appropriately for children's needs. While this flexibility is viewed as a major advantage, it does take place in the context of fiscal limitations, so the flexibility is constrained. There is more flexibility to use lower levels of care than higher.

The Scope of the Initiatives. The scopes of the fiscal reforms vary considerably across states. In two states, the initiatives cover most child welfare services across the state (except for the initial intake and child abuse and neglect reports, which were retained as responsibilities of state workers in all the states). In other states, the initiatives cover smaller numbers of cases in more limited geographical areas. Some states continue to expand the scope of their programs while others have pulled back. Some initiatives include children and families in systems other than child welfare.

Twelve of the 22 states with fiscal reforms have or are moving toward statewide programs, although these programs often do not cover the entire caseload. Boundaries of programs are usually defined in terms of particular services (most often foster care or other substitute care) or groups (e.g., children with severe needs for whom intensive residential services might be used). Overall, scopes of the initiatives are highly specific to the states’ particular situations and objectives.

The Target Populations. Many of the initiatives focus on particular groups of children or families. This can lead to potential problems in targeting. At one end, focusing on low-risk cases to prevent involvement with the child welfare system may result in the inclusion of a number of cases that might not have become involved in the child welfare system in any case. In contrast, some programs focus on high-end cases. These cases are often the focus of policymakers, since they have such extensive need, make the greatest demands on resources of the system, and are the most costly. But focusing on this group requires that there are ways to deal with their severe problems. Underlying the establishment of fiscal reforms in such cases is the assumption that the group can be adequately served with fewer resources, for example, by caring for them in less intensive placements and providing extensive supportive services. Success of these fiscal reform initiatives depends on the extent to which the assumption holds and the children can be maintained in less intensive arrangements. In several states, it appears (at least initially) that some children can be maintained in less intensive settings and can be served at a lower cost, even with extensive support services. It is important to note that providing support services requires flexibility in funding and delivering services that can be difficult to achieve under current child welfare funding mechanisms.

Organizational Models. The initiatives follow varying organizational models. The most common is a lead agency model, in which the public agency contracts with a private agency that assumes responsibility for contracting with other providers and providing case management and coordination. Lead agencies may provide some services (beyond case management) themselves. Some initiatives use managed care organizations, private for-profit or not-for-profit entities that assumed responsibility for fiscal administration, case management, and developing a network of contracted service providers. A few public agencies maintain their traditional management roles, incorporating fiscal strategies into their contracts with private agencies (e.g., performance contracts), and sometimes assuming the role of a managed care organization. Currently evidence is lacking regarding the relative effectiveness of the different organizational models.

Standardized Decision Protocols. Several of the initiatives used standardized decision protocols. Such protocols hold the promise of greater consistency in decisions made about cases, as well as higher conformity to policy intent; however problems arise when the protocols cannot fully account for individual circumstances. Although greater consistency occurs, the question of the correctness of the decisions remains. Other problems occur when protocols are complex and difficult to implement. Further study is needed of the use of decision protocols in child welfare.

Evaluation. Several of the initiatives are being evaluated, and reports are available for a few. The evidence was mixed regarding the initiatives’ effectiveness; however, the evaluations tended to look at outcomes that are measures of system performance (and perhaps consumer satisfaction), rather than longer-term issues of child and family functioning. Perhaps most important, evaluation studies have so far revealed little about the conditions that are necessary for success or about those circumstances that lead to disappointment. Clearly, more extensive and more searching evaluation is needed.

Risk Sharing and Risk Management. One of the characteristics of managed care programs is that they provide for the sharing of risk among organizations, those responsible for financing services and those providing them. Thus, some financial risk is shifted from public agencies to private contractors. Three sources of risk may be identified: (1) number of children and families served (volume), (2) level of care provided (intensity), and (3) length of service (duration).

Traditionally, payment for child welfare services has been fee-for-service, which does not expose contractors to any of the three sources of risk, although it may result in losses to a contractor if the established fees do not cover the costs of the services. In some initiatives, a lead agency or managed care organization receives payments based on managed care principles but pays service providers on the basis of fee-for-service. However, the most common payment arrangement is the case rate, in which contractors are paid a fixed amount for each case served, exposing them to intensity and duration risk. In any arrangement, risk sharing may be implemented through provisions for bonuses or penalties for performance.

Contracts may limit the private agency’s risk in various ways – for example, through stop-loss provisions (limiting the contractor’s loss to a certain percentage over the contract amount) or risk pools (funds established by the state which contractors can access if their costs exceed payments by a certain percentage). Fewer than half of the initiatives reviewed, and for which the necessary information is available, appear to incorporate limits to contractor risk. Several states adjusted the rates or payment model after a period of operation, to re-align the payments with actual cost experience

It is evident, however, that contractors have other ways to reduce their financial liabilities under managed care contracts. Lead agencies or managed care entities sometimes institute utilization review procedures, in which decisions on level of care and other services are subjected to second-guessing, attempting to assure that the decisions were appropriate. Beyond that, contractors often have some control over case referral, decisions on cases, and service planning. Some are able to regulate the number of children with expensive needs accepted into their programs. Flexibility in the use of resources is a crucial element in these initiatives and is used to provide lower levels of care (that are less expensive) than might have been the case without these initiatives. Of course, this is one of the main ideas behind these approaches, but it is largely unknown the extent to which lower levels of care are appropriate or inappropriate, given the child's needs. Fiscal considerations clearly enter into these decisions, raising the question of how children's needs should be balanced with financial pressures. Some contractors appear to use various forms of triaging of cases or rationing of services to help control risk.

Still another device that contractors use is to rely on community resources or other funding sources. This too is a central objective of many programs. This effort can be seen as an attempt to shift responsibility for child welfare cases away from the child welfare system. Many reformers hope to do just that, arguing that communities ought to take responsibility for the welfare of their children. There is, of course, the philosophical question of whether this responsibility ought to reside in the state or in communities. More practically, there is considerable variation in the capacity of communities, and difficulties arise when they do not have the resources to accept this responsibility.

Challenges Faced by the Initiatives. Several major challenges must be addressed if fiscal reform initiatives are to have a positive impact.

• Payment levels must be adequate and must take into account variations from expected levels of service. Risk and reward must be balanced and not too excessive on either end. There must be adequate resources for success, either within the agencies themselves or in the community.
• States must have flexibility in selecting and paying for services, in order to provide incentives to try different ways of serving children and families and establish more effective and efficient systems. States should be supported in incorporating this flexibility, which they can achieve by integrating funding from several public agencies and by implementing title IV-E waivers.
• Good data systems are important for successful management of any organization, but they are particularly critical in managed care arrangements. Substantial investment is needed in hardware, software, and training to ensure that information technology is available and used for system implementation and improvement.
• It is essential that fiscal considerations, and attention to proximate system performance indicators, not be allowed to overshadow objectives of improved wellbeing of children and families. Quality control mechanisms that assure continual attention to those objectives need to be enforced.
• The initiatives require complicated change processes, as states shift service delivery from public agencies to private contractors, implement team decisionmaking about cases, switch to a focus on outcomes rather than processes, and bring together a range of organizations to work on the initiatives. State and federal involvement to support development and implementation of fiscal reform initiatives should include providing training and technical assistance, disseminating written products, allowing sufficient start-up funds, adopting realistic implementation schedules, and convening forums to discuss emerging issues and policy decisions.

Chapter 1: Introduction

The provision of child welfare services has undergone significant changes during the 1990s. Some of the most important changes include a renewed focus on the rapid achievement of permanency goals for children, a large increase in the number of children in foster care, the continuing shift of responsibility for direct care to private agencies (both nonprofit and for-profit), the development of management information systems to monitor case progress more carefully, and the use of financial incentives to direct services toward desired goals. State child welfare systems have responded to these new circumstances in different ways and to varying degrees.

Generally, the provision of child welfare services is a partnership between government and private providers of service. Although states vary considerably in the division of responsibility, most states and localities contract with private nonprofit or for-profit organizations for the provision of at least some services to children and families. Until recently, these contracts were largely fee-for-service arrangements, in which the provider was paid by the state or county for delivering specific services.

This report describes the implementation by states of fiscal reforms in child welfare that replace traditional fee-for-service payment arrangements.¹ It also identifies issues that implementation of fiscal reforms faces and describes how well fiscal reforms are working. Many of these reforms, such as capitated rates, are based on the managed care model that has been used in medicine for the past 30 years. However, some reforms reflect other approaches, such as the privatization of services and performance contracting. Some states, perhaps most notably Kansas, have transformed their entire systems along these lines. Most states have chosen to implement fiscal reforms on a smaller scale, targeting specific populations or programs.

Emergence of Managed Care in Child Welfare

The fiscal reforms described in this report are all directed at changing the relationship between public child welfare authorities (states or counties) and private agencies by altering the financial arrangements between them. The intent is to influence the behavior of those private agencies. Although not all are managed care reforms, most have incorporated managed care strategies at least to some extent.

Origin of Managed Care

The concepts, principles, and tools of managed care were first developed in the medical field starting in the 1970s. Managed care was a response by the major payers of medical costs (employers and unions) to soaring medical care expenses. It was thought that a major source of higher costs was the fee-for-service financing system at the time. In that system, individual patients selected their health care providers, who had the sole discretion to set prices for their services. Third-party payers then footed the bill for any care that insured patients received. Critics charged that such a system was ripe for abuse by both patients and doctors. For patients, health care costs had become an abstraction represented by a bill that was paid by someone else. Doctors, it was argued, could increase their fees with little complaint from patients and could order unnecessary tests and procedures to increase revenue with no accountability.

The development of medical managed care was also driven by other dynamics. As medical malpractice suits began to proliferate, doctors responded by practicing “defensive medicine” in which they ordered tests or procedures in order to avoid accusations that they had been negligent. Further, in the fee-for-service system, doctors had little incentive to consult with one another and coordinate patient care. Patients could go directly to an expensive specialist or even multiple doctors at the same time. This could lead to overlapping treatments, dangerous prescription drug interactions, or other problems because no “gatekeeper” was aware of the full spectrum of the patient's medical history.

The American health care system underwent a revolution from a fee-for-service system to a predominately managed care system during the 1970s and 1980s. By 1998, three-quarters of privately insured Americans under the age of 65 were enrolled in some kind of managed care plan (Rosenbaum, 1998, p. 198).

Managed care organizations serve as “gatekeepers” that coordinate services for the patient. They seek to restrain doctors from ordering unnecessary tests and services by limiting certain reimbursements. They often require an authorization process that gives them the opportunity to review the diagnosis and medical recommendations and to suggest less expensive treatment. At the same time, they require patients to make a co-payment so that they have a financial incentive to avoid unnecessary procedures.

Managed Care in the Child Welfare System

In the early 1990s, some child welfare professionals began to advocate the adoption of managed care models, and many state child welfare systems began to try such arrangements. By the middle of the 1990s, some observers believed that managed care in child welfare was developing very rapidly (Scallet, Brach, and Steel, 1997) and was about to revolutionize the field (Emenhiser, Barker, and DeWoody, 1995). However, it appears that the adoption of managed care principles and tools in child welfare service systems has proceeded slowly. The Child Welfare League of America’s 1998 state and county managed care survey indicated that 29 states had some kind of managed care or privatization initiative (McCullough and Schmitt, 1999). It was estimated, however, that such initiatives targeted only as little as 10 percent of the nation’s child welfare population.

Perhaps the driving force in the development of managed care in child welfare was the rapidly escalating costs experienced by state child welfare systems in the late 1980s and early 1990s. This increase in costs was largely driven by increases in the numbers of child maltreatment reports and children entering out-of-home care. In 1984, a total of 1,727,000 children were reported as neglected or abused; this number had risen to 2,890,234 in 1993—an increase of 68 percent (Curtis et al., 1995). By 1996, that number had increased to 3,126,000 (Waldfogel, 1998). Reports declined in 1997 and 1998 before increasing again in 1999 (NCANDS, 1999, 2000, 2001). The number of children in out-of-home care grew by 65 percent between 1984 and 1993, from 270,000 to 445,000 (Curtis et al., 1995). By 1999, the number of children in care had increased to 581,000 (AFCARS, 2001).

Increases in the unit costs of services also added to the increase in the cost of foster care. Moreover, the substance abuse crisis contributed to an increase in children entering the system with multiple psychological and physical traumas, such as high rates of exposure to drugs in utero. In addition, improved diagnostic tools and treatment capability raised expectations for state agencies to provide service for complex conditions. Combined, these factors result in small numbers of children with very severe difficulties who may absorb a majority of resources.

All of these factors are likely to have contributed to increased foster care expenditures. Of course, increased costs may provide a greater benefit for the children in state care. Insofar as specialized foster care placements address the complex needs of children, better outcomes, if achieved, may justify the heftier price tag.

Managed Care Assumptions

Because managed care practices were developed in the medical field, they require some adaptation to be applied to child welfare. Whether managed care can be adapted sufficiently to operate effectively in the child welfare arena depends on the following assumptions:

1. Economic incentives are important determinants of service provision in child welfare. Increased expenditures in child welfare may be the result of perverse economic incentives. Private agencies can and should share some of the financial risk of increased foster care costs with state agencies.
2. Decisionmaking in child welfare is sufficiently sophisticated that the appropriate course of action can be determined in most cases.
3. It is possible to set rates of payment for services under managed care arrangements that will allow a well-managed agency to cover its costs. This implies that reasonable predictions of costs are possible.
4. Prevention of placement is possible but often requires the availability of other supports and services.
5. Services offered by community-based organizations are more effective than more traditional services. The task for contractors or other case managers is to develop and manage flexible provider networks within the client’s neighborhood and social networks.

How each of these assumptions plays out in child welfare is considered next.

Economic Incentives

Some researchers and policymakers claim that the child welfare system has labored under economic incentives that keep children in foster care longer than may be necessary. They argue that the structure of federal funding for child welfare is the source of this problem. The problem is threefold. First, differential federal funding may distort local decisions. Since the federal government reimburses states for a share of the costs of foster care but not for in-home or preventive services, serving a child at home may be more expensive for the state even if those services are more appropriate and cost less overall. Federal reimbursement for a state’s foster care costs ranges from 50 percent of costs to over 75 percent depending on the state’s concentration of poor families. So, for example, a particular child may be better served by in-home aftercare services for $500 a month rather than a continued foster care placement at $1000 a month, but if the federal government reimburses the state for none of the in-home services and 75 percent of the foster care placement, returning the child home may be more expensive for the state. Workers and even supervisors may not consciously think about the public policy impact on their case decisions, but the overarching structure of the system may exert subtle pressure nonetheless.

A second part of the problem involves the available service array that results from the federal emphasis on foster care funding. Because services flow to the funding, many believe that the current reimbursement structure has led to foster care services that are better developed and more available than alternative service models. Therefore caseworkers are unable to base decisions on a range of service alternatives. Congress intended to remove incentives for placing children in foster care by creating the title IV-B Child Welfare Services Program in 1980 and the Family Preservation and Family Support Program (now called the Promoting Safe and Stable Families Program) in 1993. These programs provide grants to states for a variety of child welfare services, including those to help prevent foster care placement. However, the federal funding for title IV-B child welfare services is far less than for title IV-E foster care payments. In FY2002, title IV-B appropriations totaled less than 13 percent of title IV-E foster care appropriations.² In addition, title IV-E funds are an uncapped entitlement that reimburses states for a portion of foster care costs, no matter how fast they grow, while title IV-B is a capped matching grant that has grown quite slowly.

A third aspect of the issue relates to the provider base for foster care and residential children’s services. Some believe that the entitlement reimbursement under a fee-for-service model has led to excess service capacity. In order to maintain revenue, service providers may continually seek out new populations of children who might benefit from their care, keeping beds full at a higher level of care than may be needed. If the state or county is not vigilant regarding the level of care needed by individual children, it becomes easy to over-use expensive services because they are more readily available than lower cost alternatives. This “structural flaw,” according to some, has created a child welfare system that must maintain large numbers of children in care in order to perpetuate itself (Wulczyn, 2000).

The arguments about how the child welfare funding structure may affect service delivery have been widely discussed for a number of years. They remain speculative, however, and considerably more evidence is needed to support them. Furthermore, there are alternative explanations for these problems. Prominent among them is the chronic underfunding of child welfare services, which results in high caseloads and the inadequacy of other resources needed to help families work toward reunification of children.

Decisionmaking

In medical managed care models, there is the assumption that, for most ailments, a correct method exists for determining the most effective treatment. This has proven to be a hard assumption to justify. In fact, Eddy (1994) reports that “in general, observers looking at the same thing will disagree with each other or even with themselves from 10 percent to 50 percent of the time.” The assumption is even more unlikely to translate into the field of child welfare because of the difficulties in problem definition and a lack of research on best practice and the correctness of decisions. There is evidence of considerable disagreement among experts in the child welfare field as to the proper decision in particular cases (Schuerman, Rossi, and Budde, 1999).

Another challenge to the assumptions that underlie managed care is that social workers and agencies are not the final decisionmakers. For children in state custody, judges have the ultimate decisionmaking authority. They may order additional services, refuse a recommendation to return a child home, or delay the termination of parental rights when the agency is trying to move the child toward adoption. Hence, social service agencies that contract under managed care have limited control over the amount of services that will be provided. Judges are less subject to fiscal incentive structures that are designed to implement policy intent. This gap between the risk that agencies assume and the control they have over decisions is a problem that is likely to plague child welfare managed care.

Rate Setting

Another significant problem for child welfare managed care is that prepayment or prospective payment systems often rely on historical utilization data to set payment rates. If it is known that there are 20 people with diabetes in a population and it takes X dollars to treat them, predictions can be made about the future costs of providing that treatment. In child welfare, data of this sort are rarely available, and even where they are, they are often deficient.

Case mix is also a problem. In the medical arena, both managed care and traditional insurance providers have developed mechanisms to limit their exposure to costly cases. They do this to minimize risk, since to enroll large numbers of people with expensive, chronic, or disabling conditions would quickly generate high costs and therefore financial losses (Master, 1998).³ Similar dynamics may occur in child welfare managed care. Insofar as child abuse and neglect are acute and episodic, a managed care approach is more likely to be successful; chronic, long-term conditions will cause difficulties. Ideally, there would be sufficient low-intensity users to balance out the risk involved with long-term conditions. However, child welfare cases are heavily weighted toward the chronic and long-term.

Prevention

Prevention advocacy is quite fashionable in child welfare, as in many other areas. However, there is scant evidence of the effects of most efforts at preventing child maltreatment (Littell and Schuerman, 1995). And there is substantial evidence that placement prevention programs do not have their intended effects (Schuerman, Rzepnicki, and Littell, 1994; U.S. Department of Health and Human Services, 2001).

If managed care agencies are unable to prevent entrance into foster care, they may attempt to limit service utilization by preventing recidivism. Little is known, however, about how to prevent a child who is discharged from foster care from reentering. The empirical data are scant in suggesting why some children who are discharged from care will ultimately re-enter (estimates are usually around 20 percent) (Goerge and Wulczyn, 1990). Interestingly, the strongest finding thus far is that placement duration is “strongly negatively associated with the rate of reentry” (Ibid). In other words, children who stay longer in foster care are the least likely to reenter care at a later point in time. This might suggest that a longer time in foster care gives parents the time they need to get on their feet and become stabilized before regaining custody of their children. Of course, the longer children are in foster care, the less time is available for them to experience either further maltreatment or re-entry into the system. In any event, the finding presents a problem to managed care efforts to reduce stays in foster care.

Community-Based and Faith-Based Organizations

Throughout American history, community-based organizations (CBOs) have provided assistance to families in need, although the emphasis placed on these services at different historical periods has varied. The past decade, however, has seen an unprecedented attempt to create a privileged role for CBOs in the social service delivery system. CBOs have considerable appeal. They combine themes such as reliance on private, local, and—frequently—religious agencies with an activist approach to addressing social problems with significant federal resources.

The presumed advantages of CBOs are numerous. One is the flexibility to enter into a variety of relationships with clients and with other service providers. Some states explicitly rely on the ability to develop provider networks that can respond to a family’s particular situation at the community level. A second advantage is the increased knowledge about available resources for the clients. Finally, there is the opportunity to develop more effective relationships with clients based on an intimate understanding of their circumstances (Kahn and Kamerman, 1996). However, disadvantages may include uneven distribution or unavailability of CBOs in some areas as well as the issue of the capacity of CBOs to provide extensive services or serve families and children with severe needs.

Limitations of Managed Care

As the field has matured, the American public has become increasingly critical of many aspects of medical managed care, such as the requirements by insurance companies that providers obtain authorization before ordering a variety of procedures and tests (sometimes denying the requested care) and mechanisms that limit the freedom of patients to choose providers. (Public demands have led both houses of Congress to pass Patient's Bills of Rights.) The results of the widespread implementation of managed medical care have been ambiguous, and it can be reasonably said that the jury is still out. It is not evident that medical managed care has saved money. And everyone agrees it has not led to increased health care coverage for the uninsured. Many people have, however, argued that it has spawned its own kind of abuses and conundrums. What remains to develop is a consensus on whether the abuses and conundrums under the old system are qualitatively and quantitatively worse than those observed under the new system (Hurley, 1998). Described as “neither poison nor panacea” commentator Robert Hurley states, “A balanced summary judgement would be difficult, but it can be safely asserted that in general the experience has been better than its critics would acknowledge but less beneficial than apologists would contend.”

In child welfare, in spite of the impetus of rising costs, many factors have contributed to the states’ reluctance to jump aboard the managed care bandwagon. First, a diverse set of federal and state initiatives throughout the 1990s competed with managed care, the Adoption and Safe Families Act (ASFA)⁴ clearly being the most important of these. While ASFA helped promote fiscal reform efforts in numerous states, the array of family preservation, reunification, and adoption-oriented initiatives it encouraged often competed with fiscal reforms for limited management resources. Second, state agencies have less leeway regarding protecting vulnerable children than third-party payers have in providing medical care. Most crucially, the ultimate decisionmaking authority in most cases remains with the courts rather than with the state agency, limiting the ability to make definitive case plans.

Methods for the Report

Information for this report was gathered through reviews of existing documentation about fiscal reforms in the states, including materials produced by the states themselves, results of research by other organizations such as the Child Welfare League of America, evaluation reports by independent evaluators, and conference presentations on the reforms made by state officials as well as their consultants and evaluators. Most of the existing materials used had been published during the period 1999 through 2002. From August 2000 through November 2001, states were contacted to fill in gaps in the publicly available information about their particular initiatives. At that time, officials were invited to verify the accuracy of the information that had been gathered from other sources.

Information was available about a considerable range of changes in the relationships between states and private agencies. In order to focus the report, the universe of interest was defined as those efforts that involved changes in financial arrangements between the state (or county) and private contractors⁵ designed to affect the behavior of the private organizations. The programs described below do not include or represent every such initiative across the United States because some initiatives were excluded due to time and space limitations. The programs do constitute the majority of such efforts.

The following chapters describe the fiscal reforms, identify issues that were encountered in implementation, and specify what is known about how well they are working. The descriptive information (Chapter 2) covers the scope of these programs and their target populations, their objectives, and their organizational models. Chapter 3 discusses in detail their financial arrangements, with particular attention to issues of risk. Finally, the conclusion (Chapter 4) discusses some of the ongoing challenges in implementing fiscal reforms in child welfare that may be of particular interest to federal policymakers and identifies how those challenges have been addressed by the initiatives. Individual summaries of each initiative are presented in the appendix.

Endnotes

¹  The examination of this subject was aided considerably by a previous survey of such programs: McCullough and Schmitt, 1999.

²  Title IV-E foster care appropriations in FY2002 were $5.06 billion, while the title IV-B appropriations were $597 million ($292 million for Subpart 1 and $375 million for Subpart 2).

³  Medicaid, on the other hand, is a federally sponsored medical coverage program aimed at low-income people, many of whom have long-term, chronic health conditions. In contrast to private health insurance, managed care concepts have only very recently been applied to the Medicaid program. Part of the reason for this is that managed care has had limited success in controlling costs when applied to the chronic conditions that plague much of the Medicaid population.

⁴  ASFA, passed in 1997, sought to achieve outcome goals in seven areas: reduce the recurrence of child abuse and/or neglect; reduce the incidence of child abuse and neglect in foster care; increase permanency for children in foster care; reduce time in foster care to reunification without increasing reentry to foster care; reduce time in foster care to adoption; increase placement stability; and reduce placement of young children in group homes or institutions (U.S. DHHS, 1999).

⁵  In this report, “contractor” is meant broadly, to cover any private nonprofit or for-profit organization that has a contract with the state to deliver services or manage networks of providers. It can refer to a lead agency or a managed care organization as well as a direct service provider. In some initiatives, the organization assuming financial risk may provide no services directly but contract out for them. The organization may receive a capitated rate, case rate, or block grant, then pay service providers fee-for-service or per diems; thus, the service providers themselves assume no risk. In other initiatives, risk is transferred to service providers.

Chapter 2: Overview of the Initiatives

This review focuses on 23 initiatives in 22 states. Each of the initiatives involves the implementation of financial arrangements to influence the behavior of private agencies. The fiscal strategies used by the states include capitated and case rates, risk sharing, performance contracting, performance incentives, privatization, and pooled or flexible funding. Although strictly speaking not all are managed care initiatives, most have incorporated managed care strategies (such as prospective payments, utilization management, and service coordination) at least to some extent.

This section provides an overview of the initiatives that were reviewed. Table 2-1 provides a descriptive summary of the initiatives. The information is “point-in-time” as of fall 2001, and much may have changed since it was collected, especially since most of the initiatives had been implemented relatively recently (see the table) and were still evolving. The information does illustrate the range of approaches states are using to better serve children and families. The report does not attempt to describe or draw any conclusions about the effectiveness of the initiatives; a few of the initiatives are being evaluated and will have that information (or do already) but evidence is still preliminary or lacking for most.

Scope of Initiatives

The “scope” of an initiative refers to the proportion of children and families in a state that are covered by the initiative and is defined by both the geographic area and the populations served. An initiative may be implemented in a small, defined area or the entire state. It may serve a subgroup of the child welfare population, such as children in traditional foster care only, or the entire child welfare population.

The initiatives investigated represented a variety of approaches, from small, contained projects that either stayed small (such as Kentucky’s initiative) or eventually expanded (Illinois, Tennessee), to projects covering, nearly from the onset, most (Massachusetts) or all (Kansas) of the statewide child welfare caseload. Several of the initiatives covered most or all of the state geographically but included a smaller proportion of the child welfare caseload (Arizona, Georgia, Oklahoma).⁶ In Florida, a statewide fiscal reform is being implemented district by district, so it currently covers only part of the state but will target the entire child welfare caseload when fully implemented. In Missouri, the initiative is limited both to a narrower segment of the child welfare caseload and to a smaller area of the state. County-administered states such as California, Colorado, Minnesota, and Pennsylvania had county-designed and county-implemented projects that varied considerably in terms of populations and services covered. Some initiatives were designed for urban areas with large proportions of the states’ child welfare caseloads and specific system characteristics or needs (Baltimore, Detroit, Milwaukee, New York City). Title IV-E waiver demonstrations were implemented in limited areas of the states and/or targeted narrower segments of the child welfare population due to their waiver designs and the experimental nature of the demonstrations (California, Connecticut, Maryland, Michigan, Ohio, and Washington).

Table 2-1
Summary of State Fiscal Reforms in Child Welfare
“W” indicates the initiative is based on a title IV-E waiver; “C” indicates the initiative has ceased; “NA” indicates that information is not available.

Initiative	Geographic Coverage	Objectives	Year Implemented	Structural Model	Target Population	Caseload
Arizona Family Builders	Most of the state	Provide voluntary services for families previously unserved	1998	Lead agency	Low-risk and potential-risk families with reports	6000 referrals/year, 1900 family assessments, 1600 families receive services
California Project Destiny (W)	Alameda County	Reduce length of stay in care, divert SED placement in residential facilities, serve children in the least restrictive environment	1997	Lead agency	SED children; those most at risk of placement in high level group homes	90 children enrolled at time of interview; budgeted capacity of 256 over entire waiver period
Colorado Boulder County Managed Care Pilot Project	Boulder County	Gain flexibility to enhance interagency partnerships and provide services in the community	1997	Public agency	Adolescents 12-18 in need of or at-risk of needing residential services	500 youth
Connecticut Continuum of Care (W)	North Central and South Central regions of the state	Reduce length of time in care, develop a localized network of services, improve outcomes by establishing flexible incentive-oriented environment	1999	Lead agency	Children ages 7-15 with severe behavioral, mental health, or educational problems	Maximum of 70 children
Florida Coalition for Children and Familiesa/	District 8 (Sarasota)	Provide services efficiently and effectively through community partnerships	1997	Lead agency	All children with founded neglect or abuse reports, regardless of whether in-home or out-of-home	1650 children
Georgia Metropolitan Atlanta Alliance for Children (MAAC)	Atlanta area	Place children in the best long-term and least-restrictive settings in a system that moves children out of high levels of care quickly and efficiently	1998	Managed care organization	Children needing residential care who are not eligible for Project Match	40 children
Illinois Performance Contracting	Statewide	Ensure more efficient use of limited resources, improve outcomes, control costs, increase permanency	1998 (across state)	Public agency	All children in relative, traditional, and specialized foster care	35,000 children and their families
Kansas Public Private Partnerships	Statewide	Improve client outcomes, increase permanency, better protect children at risk	1997 (total population)	Lead agency	All children in state custody and at risk of entering custody	3000 families
Kentucky Quality Care	Jefferson County	Decrease length of stay in care, improve outcomes, improve quality of services, provide individualized care	2000	Lead agency	Adolescent girls in residential placement, children transitioning home, children entering care and in need of intensive services	30 children
Maryland Baltimore Child Welfare Managed Care Project (W)	Baltimore City	Reduce congregate care and enhance permanency for children ages 0-5; decrease length of stay and recidivism for children in care	2000	Lead agency	Children ages 0-5 in out-of-home care and siblings; newly dispositioned children of any age and siblings; kinship conversions and siblings in care	500 children
Massachusetts Family-Based Services	Statewide	Implement a collaborative, community-based approach utilizing state resources and maximizing the use of other resources	2000	Managed care organization/ lead agency	Primarily, children at risk of placement and their families; some children in care	3300 families
Michigan Michigan Families (W)	St. Clair, Monroe, Livingston, Van Buren, Jackson, and Newaygo Counties	Find innovative ways to serve and improve outcomes for children without necessarily putting them in foster care	1999	Lead agency	Children in out-of-home care or at risk of being placed	190 children
Michigan Permanency Focused Reimbursement System	Wayne County	Keep children out of residential facilities, provide as many services as possible in the community, allow flexibility in treatment approaches	1997	Lead agency	Children in care	4000 children
Minnesota PACT 4 Families Collaborative	Four rural counties in western MN (Kandiyohi, Meeker, Renville, Yellow Medicine)	Ensure that children receive needed services, including mental health, and provide early intervention	1995	Lead agency	Children ages 0-21 and their families	NA
Missouri Interdepartmental Initiative for Children with Severe Needs.	Urban eastern region and rural central region	Provide better coordination of services to reduce barriers, enhance effectiveness and efficiency, and prevent children from falling through the cracks	1999	Managed care organization/ administrative services organization	Children ages 4-18 in or at risk of long-term residential placement and with serious behavioral health needs as measured by a standardized instrument	250 children
New York Safe and Timely Adoptions and Reunifications (STAR)	New York City	Enhance permanency outcomes by providing flexible dollars based on agencies’ improvement in outcomes	2000	Public agency	All children already in care	40 out of 44 providers participate
Ohio ProtectOhio (W)	Franklin County	Use performance bonuses and managed care to reduce length of stay in care and increase flexibility of services	1999	Public agency/lead agency	All children and families with reports	Performance bonuses (public agency): 5100 children; managed care (contractors): 1200 children
Oklahoma Oklahoma Children’s Services	Statewide	Keep families together, bring about reunification quickly, prevent disruption of placement	1992 (across state)	Lead agency	Children at home and at risk of placement; children in care working toward reunification	2000 families
Pennsylvania Berkserve (C)	Berks County	Develop an efficient public-private partnership model using a network of local agencies to provide services	1997 (ceased in 2000)	Lead agency	Any family with at least one child in the child welfare system	24 families
Tennessee Continuum of Care	Statewide	Provide services in the least restrictive and lower cost settings, as well as reduce length of stay and recidivism	1995	Lead agency	Children in state custody who require a level of care higher than regular foster care	4400 children
Texas Permanency Achieved Through Coordinated Efforts (PACE) (W)b/ (C)	10-county area around Fort Worth	Improve outcomes, ensure efficient use of limited resources, decrease lengths of stay, provide coordinated services	1999 (ceased in 2001)	Lead agency	Children needing a level of care higher than regular foster care	600 children at its peak
Washington IV-E Waiver Demonstration (W) (C)	Spokane County	Ensure placement in the least restrictive setting, decrease length of stay, improve permanency outcomes	2000 (operated 6 months)	Lead agency	Children ages 8-17 likely to enter high-cost care with a DSM diagnosis and with mental health or special education needs	30 children at its peak (50 overall)
Wisconsin Bureau of Milwaukee Child Welfare	Milwaukee County	Reform the child welfare system in Milwaukee County (the State took over the county's system)	1998	Managed care organization	All children in the county who are identified as at-risk of abuse or neglect, and all children in out-of-home care	Ongoing Case Management (children in out-of-home care): 6000 children; Safety Services: 200 children; Wrap-around: 1000 children
a/ Florida was granted a title IV-E waiver for its privatization initiative, but the waiver was never implemented. b/ PACE began under a title IV-E waiver, then it was withdrawn from the waiver in 2000.

Federal court orders or state legislation requiring changes in child welfare systems often prompted initiatives targeting a large proportion of the child welfare population. A court order or legislative mandate ensured that funds were appropriated to implement the changes. The earliest of these was in Oklahoma, where an initiative was designed in response to a consent decree involving adolescents in state custody. Kansas’s initiative was implemented as a result of a lawsuit regarding timely service provision as well as pressure from the governor and legislature to privatize services. In Florida, legislation requiring districts to contract with lead agencies for child welfare services was passed in a general climate of reducing government and providing more services at the local level. The initiative in Wisconsin was the result of a court order and a legislative change in which the state, previously entirely county administered, took over child welfare in Milwaukee County. In Berks County, Pennsylvania, the growing complexity of regulations and standards, as well as anticipation of state imposition of managed care requirements, prompted child welfare service providers to develop a local managed care pilot.

In several states (Arizona, Illinois, Kansas, Maryland, and Oklahoma), the initiatives started out as limited pilot projects. All but one have become permanent and expanded beyond the original geographic area to the entire state (although not the entire child welfare caseload, except for Kansas). The exception is the managed care initiative in Baltimore (a title IV-E waiver demonstration). The state plans to carefully examine indicators of service quality to assess its success before deciding about making the project permanent.

Objectives

The initiatives were motivated by a large range of factors. Some were the result of court and state mandates to change practice, improve outcomes, and/or spend less money. Others addressed large and growing permanency backlogs that persisted despite intensive efforts and in the face of ASFA requirements. Concerns about families in crisis and children who languished in foster care or overly restrictive placements for extended periods underlaid many initiatives. In response to these concerns, the initiatives were implemented to achieve two types of objectives: (1) better outcomes for children and families and (2) system goals such as service flexibility and spending the dollars more effectively—both of which often involved obtaining more or enhanced services for the same amount of money.

Improving outcomes for children and families usually entailed redirecting resources from maintaining children in care to achieving permanency outcomes--preventing placement, reunifying children with their families more quickly, shortening length of stay in placement, reducing recidivism. The initiatives in Illinois and New York City have this type of objective, and both focus on their entire foster care caseloads (excluding children in residential treatment centers and specialized foster care). These initiatives provide fiscal incentives or rewards to agencies that meet standards or show improvements in permanency outcomes for children in care. Arizona uses a different means to improve outcomes by preventing placement; its fiscal reform initiative provides services to potential- and low-risk families only. The objective of Kansas’s initiative, which involves its entire child welfare caseload, is to use performance-based contracts to enhance child safety and well-being.

One type of system goal involved gaining the flexibility to implement interagency or public/private partnerships and provide a broad array of services. These initiatives emphasized collaborations and community-based approaches as well as maximizing the use of other resources and enhancing federal reimbursements. For example, the initiative in Boulder County, Colorado, is an interagency collaboration established to provide the flexibility to “serve kids as Boulder County kids, not as DSS kids or juvenile corrections kids.” The initiative in Massachusetts provides a flexible collaborative response to family needs by customizing services based on community needs and resources. Minnesota’s PACT-4 collaborative pools funds from county agencies, school districts, and private partners in four counties to provide integrated, community-based services. Missouri’s initiative integrates funding from various state child-serving agencies to support comprehensive, coordinated services for children likely to be served by multiple state agencies. Often, the objective was cost neutrality, spending the dollars more effectively and providing flexibility to enhance outcomes for children and families.

Achieving system goals such as spending dollars more effectively usually involved implementing programs to prevent high-cost placements and ensure placement in the least intensive and least restrictive setting possible and appropriate. Developing local provider networks and enhancing community services were usually components of these initiatives, which generally targeted children requiring a level of care higher than regular foster care. For example, Tennessee’s initiative focuses on children who need a level of service higher than regular foster care; it provides fiscal incentives to agencies to provide services in the least restrictive settings and thus achieve savings for the state by avoiding high-cost therapeutic placements. Alameda County, California’s Project Destiny focuses on severely emotionally disturbed (SED) children. It provides wraparound services to shorten length of stay in expensive residential treatment. Money saved by preventing or shortening high-cost placements generally was not used to reduce child welfare spending; instead, it was used to enhance services, serve more children, or improve the system's capacity in another way.

Structural Models

The structural models of the various initiatives varied substantially regarding how many functions were retained by the public agency versus contracted out. In all the initiatives reviewed, the initial intake and child protective services (CPS) investigations were retained by the public child welfare agency. Beyond those initial functions, however, management and service delivery structures could be categorized into lead agency models, managed care organization models, public agency models, and administrative service organization models (see McCullough and Schmitt, 1999; and U.S. General Accounting Office, 2000). Most (15) of the initiatives followed the lead agency model, with three (Colorado, Illinois, and New York) following a public agency model and two (Georgia and Wisconsin) using a managed care organization model. Massachusetts utilized a mixed (managed care organization and lead agency) model, as did Missouri (managed care organization and administrative services organization) and Ohio (lead agency and public agency).

Lead Agency Model

In the lead agency model, the public child welfare agency contracts with a private nonprofit or for-profit agency to serve as a lead agency for a county, service area, or region. The lead agency then coordinates and provides all necessary services (either directly or by subcontracting with providers) and sometimes conducts utilization management and quality assurance. The goals in having a lead agency are to enable or encourage provider networks and provide accountability at the local level. In some cases, the lead agency assumed considerable financial risk (discussed later).

Every service district in Florida, for example, is required to contract with a lead agency (or, in the Miami area, more than one lead agency) that will take over all child welfare responsibilities beyond initial intake and investigative functions. In Kansas, the state child welfare agency maintained responsibility for administrative services (such as utilization management, monitoring services, and tracking performance and outcomes) and contracted with nonprofit lead agencies to coordinate and provide all child and family services. Maryland’s managed care initiative in Baltimore involves a vendor (a partnership of a nonprofit and a for-profit agency) that is responsible for all administrative functions, case management, and service delivery for children referred into the project.

Public Agency Model

Illinois and New York City followed a public agency model, which maintains the traditional management and service-delivery structure while incorporating managed care practices in its own practices or contracts with service providers. Both initiatives involve public agencies that maintain their previous management and service-delivery structure while incorporating financial incentives into their contracts with foster care agencies. In both cases, the public agencies closely monitor the agencies' performance and outcomes, and financial incentives are based on analysis of data on permanency outcomes. In Colorado’s Boulder County initiative (as well as in other managed care counties in Colorado), the public agency has joined with other public child-serving agencies to use managed care principles in case management and service delivery. Oklahoma also follows a public agency model in its capitated contracts with providers.

Managed Care Organization Model

The managed care organization model involves the public agency’s contracting with a private organization that incorporates managed care principles into its subcontracts with service providers. Generally the private organization does not itself provide direct services. For example, every service area in Massachusetts is covered by a community-based child welfare agency that receives a set amount of money each year to provide functions such as gatekeeping, utilization review, creating and maintaining provider networks, monitoring quality of services, and accessing third-party reimbursement. The lead agencies generally subcontract for services, although they are allowed to provide up to 20 percent of the services delivered. In Missouri’s model, a private for-profit organization, which was created for the purpose, manages a network of providers and monitors quality and utilization of services.

Mixed Model

At times more than one model was incorporated into the initiatives. For example, Missouri utilized both a managed care organization model and an administrative services organization model (in which a private contractor provides administrative services only). The state contracted with two agencies, one to manage the delivery of services and the other to provide operational support for its initiative, which targets children with severe behavioral health needs. Massachusetts also contracts with lead agencies to develop and operate provider networks (managed care organization model) and a separate vendor to develop and support a database for utilization management (administrative services organization model).

Target Populations

Prior research has found that most managed care initiatives targeted children in foster care, although there was a trend toward also including children at risk of placement (McCullough and Schmitt, 1999 and 2001). In the initiatives described in this report, the target populations range from a narrow population of children in care to the entire child welfare population. Many of the initiatives target children and families with high needs. The rationale for targeting a population with severe behavioral or mental health problems or special education needs is that often the outcomes are poor, which creates a need to find different ways to address the problems, and the costs are high, which creates a visible target and builds in incentives for reducing costs. These target populations include seriously emotionally disturbed children (California, Georgia); those with serious behavioral health needs as measured by a standardized instrument (Missouri, Texas, and Washington); those with placement needs higher than traditional foster care (Connecticut, Tennessee); and adolescents with high needs (Colorado, Kentucky). Generally these initiatives encourage service provision in the least restrictive and costly setting appropriate and often provide mechanisms for conducting enhanced assessments to better plan services.

Initiatives that target all or most of the foster care caseload hope to achieve widescale improvements in permanency outcomes. For example, the initiatives in Illinois and New York City were designed to encourage providers to achieve efficiencies and improve permanency outcomes for children in care, and both cover most of their foster care populations. Maryland’s managed care initiative in Baltimore was implemented to reduce placement length-of-stays for young children in care, and the target population is all children ages 0-5 in care and their siblings (and some other types of children). These initiatives attempt to address the economic incentives to keep children in care, discussed in Section 1, by offering economic incentives to shorten lengths of stay (with safeguards intended to ensure appropriate placements).

Some initiatives target children not in care but at risk of placement, both to avoid the costs of placement and to provide alternatives to removing children from their homes. The initiatives in Massachusetts and Oklahoma primarily serve children in their own homes who are at risk of placement, with some services provided to children in care (in Massachusetts, the initiative serves over 75 percent of all children in the child welfare system). Arizona’s initiative targets low-risk or potential-risk families in order to prevent escalation of maltreatment into a higher risk category that would require taking children into custody.

Other initiatives target the entire child welfare population, for all the reasons noted above. Florida’s and Kansas’s statewide privatization requirements include all children in the child welfare system. Ohio’s and Wisconsin’s initiatives also target all children in the child welfare system in their geographic areas (Franklin County, Ohio, and Milwaukee County, Wisconsin); Ohio has other managed care initiatives in several other counties (both as part of the title IV-E waiver and outside the waiver).

Referral Process

In most of the states, a caseworker or other child welfare worker refers children or families into the initiative by using guidelines or protocols (as in Arizona, Massachusetts, Oklahoma, Pennsylvania, Tennessee, and Wisconsin). The most complex guidelines were used by the managed care initiative in Berks County, Pennsylvania, which developed a detailed protocol for county intake workers to follow. After going through selection/admission criteria step by step, if the protocol indicated that a case was appropriate for the managed care initiative, the case was referred. Additional decision trees to be followed by service providers accompanied the case. The complexity of the protocol was one factor in the demise of the initiative, as both county workers and providers found the process daunting.

A few initiatives use other referral procedures. For example, initiatives incorporating title IV-E waivers (those in California, Connecticut, Maryland, Michigan, Ohio, and Washington) involve referral through random assignment after workers applied screening criteria, which is a component of the required evaluation design of these demonstrations. Other initiatives have automatic referrals if a report was founded (Florida) or if a case meets criteria (Missouri and Texas).

Sometimes cases are screened or even referred by an interagency team. Often the initiatives involve interagency teams developing treatment plans and, in effect, pre-authorizing services. Thus one of the major features of the fiscal reforms involves implementing a team approach to referring cases, identifying family need, and specifying services, taking that responsibility away from the individual caseworker. For example, the managed care initiative in Boulder County, Colorado, institutionalized the interagency approach by developing a new organizational entity comprising representatives of all local child- and youth-serving agencies (corrections, probation, mental health, social services, public health, substance abuse services, and other community agencies); each agency contributes funding that is pooled. The new entity handles case management and contracts with private providers for services. This approach takes a child or family out of a specific system, provides for collaborative decisionmaking, reduces cost-shifting, allows flexibility in services, helps to identify and address gaps in services, and eliminates duplication of services. A challenge is that cooperation and service integration require the development of trust and clear role definition. Although some caseworkers oppose the shifting of responsibility to a team, an advantage is that it gets more agencies invested in the care of the children and aware of the issues that need to be addressed.

Most of the initiatives have a "no reject, no eject" requirement whereby contractors cannot refuse any referral from the public agency or disenroll any child until all objectives are met. Kentucky’s initiative has a provision that allows contractors to protest a referral and a third party to decide whether the case is appropriate for the initiative. Michigan's experience with a title IV-E waiver initiative highlights the effect that risk aversion can have when contractors can choose whom they serve. In that initiative, community contractors are allowed to develop their own screening criteria, which has the effect of only relatively "easy" children being accepted for services. Since the capitated rate is based on historical averages of payments for all children (including children in specialized treatment foster care and residential placement), the contractors are able to minimize their risk and accumulate money in their "risk pool."

Level-of-Care Assessments [7]

One of the promises of managed care is that it can promote efficiencies of time and money by providing more accurate assessments of client problems and the appropriate services for them through more rigorous assessment protocols. This assumes that there are a significant number of cases where the child is receiving services that are more intensive than necessary to reach desired outcomes and that it is possible to determine who these children are. These assumptions are supported by the extensive research on outcomes since the passage of the Adoption Assistance and Child Welfare Act (P.L. 96-272) in 1980, by the growing experience that social workers have in managing permanency outcomes, and by the ongoing refinement of assessment tools.

As with all aspects of managed care reforms, there is a wide variation in assessment protocols and their use in the initiatives. Numerous states require that contractors accept the state’s predetermined level of care for individual clients. Some states use independent third-party contractors to conduct a binding assessment. Most of the states allow a contractor to complete a postreferral assessment as the basis for determining a treatment plan. In some states where the target population is children with high-end needs, level-of-care assessments are often used as a screening mechanism to ensure the least restrictive setting for children. The rationale for this is that such children represent a small part of the total population but a significant part of expenditures.

Independent Assessments

Assessments can be performed by the state agency, the contractor, or an independent third party. Some of the most innovative approaches involve the use of independent third-party assessments. In Texas, the contractor performs a battery of assessments and then turns the material over to an independent third party who determines the level of care, which is absolutely binding on the contractor. Because these two agencies had extensive previous interactions, they are able to achieve agreement about the level of care for approximately 95 percent of the cases. There are real financial stakes for the contractor in correctly assessing the level of care because the flat-rate case payment is based on a historical average of the level of care needed for the target population. This average is about 3.6 on a scale of 1 to 6 where 1 is regular foster care and 6 is an intensive residential care facility. In the first year of the program, the average level of care was approximately 3.2. The second year, however, the average level of care was about 3.8-3.9, which exacerbated other financial strains facing the contractor.

In Kentucky, as well, an independent agency assesses level of care. The primary function of this agency is to adjudicate conflicts between the state and the contractor about the appropriateness of a referral. If the contractor disagrees about the appropriateness of the referral, it reviews the case with the state. If this review does not resolve the disagreement, the contractor can bring the case to the independent review agency, which examines the records and makes its own determination. As of the beginning of the second year of operation, this independent review had been used four times, with each party winning twice. Although the contractor is allowed to reject a limited number of cases over the course of a year, the contractor has continued to offer services to the children even when the level-of-care review supports their argument that the referral is inappropriate.

State Assessments

Some states perform assessments themselves. Upon referral in Connecticut, for example, a child is assessed for functionality, ability, behavior in the community, behavior in the family, and behavior in school and assigned a score which corresponds to a case rate. Children then are randomly assigned to either the experimental group (the initiative) or the control group (traditional public agency services). The state does not determine services; instead, the contractors develop treatment plans.

Contractor Assessments

Many states (such as Maryland and Tennessee) allow the contractors to conduct their own client assessments so that they can develop their own service plans. In Maryland, the contractor takes the service plan originally developed by Baltimore caseworkers and other case records, meets with the family, and then uses the Structured Decision Making assessment tool to see if the original service plan needs to be revised. In Tennessee, contractor caseworkers have 15 days to conduct a thorough assessment. They use a triage system to place the child initially while they perform the assessment. Their assessments include a social history; an Early and Periodic Screening, Diagnostic and Treatment screen; a community risk assessment to assess the risk the child poses to the community; and family strength and weakness screens. This assessment then feeds into the continuum of services the contractor offers. As a result of this system, Tennessee has been able to greatly reduce its use of emergency shelters.

There are degrees of integration of assessment and case planning. Some states (such as Connecticut) keep assessments separate from the development of a treatment plan. Connecticut initially assesses a child using a set of four-point scales that determine the child’s functioning. The assessment is then given to the contractor, which has the responsibility to develop a treatment plan for how it will broker the services. The reason for giving the contractor this responsibility is that the state does not want the assessment to limit the flexibility that the contractor has. However, many states seem to integrate assessment and case planning tightly. In Oklahoma, for example, all long-term cases requiring prevention, reunification, and placement maintenance services are referred to a contractor, which then conducts a battery of assessments and develops an intervention plan.

Pre-authorization

Traditional fee-for-service arrangements with contractors required that states purchase a specific bundle of services; any services not explicitly mentioned in the contract required the state’s permission before the contractor could provide them. Such arrangements reproduce traditional power arrangements between the state and agencies and have tended to restrict the flexibility of contractors in serving their clients. Many of the managed care initiatives undertaken over the past decade have sought to ease these limitations and to empower contractors to take more responsibility for their cases. Re-thinking pre-authorization procedures has been a feature of many?though not all?of the managed care initiatives.

Understanding this aspect of service provision sheds light on the autonomy of the contractor and how important decisions about a case are made. It is striking that none of the programs surveyed used the traditional model of having a formal process in which private caseworkers had to consult with state child welfare supervisors before initiating a new service for the client. Pre-authorization for some services remained necessary in at least three states (California, Connecticut, and Oklahoma). However, all of these services involve medical and mental health services that are paid by Medicaid. The pre-authorization is needed to meet federal Medicaid requirements, not because any of these states’ child welfare agencies mandated this process.

Based on the information collected from 22 states, there appears to be a continuum of collaboration. One pole of this continuum is represented by those states that only monitor outcomes, the other end by states that meet regularly with contractors to consult on case decisions and service provision. Table 2-2 below places each state along this continuum (referring to child welfare services only). No information is available on how child and family outcomes differ depending on the degree of collaboration.

Table 2-2
Degree of Collaboration

Only Monitor Outcomes		Some court or administrative involvement	Monthly or quarterly reviews	Frequent or continuous collaboration
Florida Illinois Kansas Michigan Missouri	New York Ohio Oklahoma Wisconsin	Connecticut Georgia Kentucky Texas	Arizona California Maryland Massachusetts	Colorado Minnesota Pennsylvania Tennessee Washington

The category “Only monitor outcomes” includes projects that privatized previously public services (Florida and Kansas) and implemented performance contracting (Illinois and Ohio). In most of these initiatives, no caseworkers from public agencies are assigned to a specific child; instead, contract monitors from the state or county evaluate overall contractor performance. In other initiatives, caseworkers from the state agency are assigned to monitor specific cases and are available for court appearances. For legal reasons, several states must provide caseworkers for court appearances even if their involvement in actually providing services is minimal (California, Michigan, Tennessee, and Texas). The category “monthly or quarterly reviews” refers not to administrative case reviews or other mandated procedures but to regular processes where various service providers (both public and private) meet to formulate case services. Finally, some states have devised projects in which there is constant collaboration between public and private agencies and often with the family as well. One example of this is Colorado’s Boulder County Managed Care Pilot Project. It has two interagency utilization review meetings a week, during which staff from state, county, and private agencies discuss cases and the appropriate services and outcomes for the clients.

It appears that in many cases the monitoring mechanisms of the state or county agencies have become embedded in the service provision process through these collaborative review processes. Minnesota’s collaboratives, for example, bring together various state and county agencies and private providers to pool funds and do assessment, coordination, planning, and purchasing of services. In Pennsylvania’s initiative, the extensive collaboration was a byproduct of the difficulty of the referral and assessment protocols, rather than an intentional feature of the pilot. The process never reached the point of cases making a smooth transition from county to contractor responsibility.

Contractor Monitoring

In keeping with the broader reorganization of the relationship between the state agencies and the contractors that is occurring in many of these initiatives, the process of monitoring contractor performance is being revised to accommodate greater flexibility.

Performance and Outcome Measures

Perhaps the most important change is in what gets monitored. In many traditional child welfare programs, monitoring mechanisms focused on process issues, i.e., were certain tasks performed (evaluations, number of visits and therapy sessions, etc.)? The new initiatives are part of a broader trend in child welfare that seeks to follow client outcomes instead of process. Performance contracting is the most direct example of this, but outcome measurement is integrated into almost every initiative. However, despite this shift in emphasis toward outcome measurement, no state has thus far completely abandoned process measures because of the continuing state responsibility to ensure quality services.

Specific outcome measures used by the states vary according to the target population served by the initiative. Initiatives that work with the general child welfare population have outcome measures such as the numbers of adoptions, children returned home without re-entering the system, and at-risk children safely maintained in their own homes. Programs that work with children with high-end needs have outcome measures that focus on placing the child in the least restrictive setting (including returning the child home).

Both state administrators and contractors indicated that this shift toward outcome measurement has been a positive step but that the process has been uneven. Some administrators noted that this has caused temporary difficulties for state personnel. As one administrator observed:

It’s been hard to get the monitoring staff around the state to look at things differently than how they’re used to. They’re used to just checking off on a list whether or not a provider did a service. Now they have to look at the providers’ work; did they do a good job? And they may need to interview kids, caseworkers, and parents. It’s a different mindset.

In addition to the changes in the mindset of workers, there has been a real revision in the responsibilities of state agency caseworkers. Another administrator noted that it is difficult for some caseworkers to relinquish the actual case decisionmaking authority in favor of a strictly monitoring relationship.

Monitoring Mechanisms

Many state administrators seem to expect that the reduced role in case decisionmaking will allow state agencies to focus more of their energies on ensuring child well-being through more rigorous monitoring processes. Many of the states retain traditional modes of accountability, including monthly reports to contract monitors or quarterly case review meetings. Some contractor supervisors have noted that the monitoring mechanisms impose another level of bureaucratic paperwork on their workers. In one state, a contractor complained that the state monitors focused on items like staffing patterns and turning case plans in on time instead of whether the child was safe.

In spite of these enduring processes and the complaints that go with them, new monitoring mechanisms have been devised by some of the states, although little is conclusively known about the effectiveness of the various monitoring systems. The two most prominent features of these systems are collaborative reviews and the integration of management information systems. As noted earlier in regard to the pre-authorization process, collaboration between state and contractor has become quite common in both case decisionmaking and performance monitoring. In Massachusetts, for example, there are several sets of meetings to discuss case issues. There is a weekly meeting between core team members to discuss current case issues. Then every 6 weeks, all of the relevant staff and the family review case plans and goals. The contractor is responsible for monitoring the specific aspects of the case such as level of involvement and the use of community resources.

Another variation on collaborative monitoring was established in Maryland, which has two committees to review program progress. The Managed Care Committee looks at global program issues and includes staff from Baltimore City’s child welfare staff, state child welfare workers, lawyers from the state attorney general’s office and Baltimore City, as well as the evaluators. The Partners Committee meets regularly to review case plans and other details of service provision. This committee consists of city and state child welfare workers plus staff from the lead agency.

Some states have contract monitors that work with the contractors on an ongoing basis. For example, in Wisconsin there are two types of program evaluation monitors--one examines service provision, and the other looks at the fiscal component of the program. There is a formal quarterly review for each aspect of the program where they discuss program quality, patterns of expenditures, and permanency plans.

Another feature of some of these programs, as in Arizona, New York, and Ohio, is the greater importance given to management information systems for case decisionmaking and contractor monitoring and accountability. Our information is tentative, but some administrators indicated a frustration that current information systems did not put useful information in the hands of the workers. Sometimes the system is difficult to access, other times the data are not broken down in such a way as to guide decisionmaking in immediate case situations. New York’s initiative probably has the most highly developed data system; it incorporates a unique interactive system that allows the public agency to tie agency reimbursement to the outcomes for children.

Evaluations

All title IV-E waiver projects are required to have independent, third-party evaluations that examine both cost and quality concerns. Some of the programs did not have waivers and did not have such evaluations underway. Most of the programs had secured contracts, usually with local schools of social work, but had not yet produced reports. Several of the states were just beginning to implement their programs and had not made arrangements for evaluations. Four states where evaluation reports are available include Arizona (Arizona Office of the Auditor General, 2000), Colorado (Mercer, 2000), Florida (Paulson et al., 2002), and Kansas (James Bell Associates, 1999).

The next section describes in some detail the various financial arrangements that states adopted to achieve their objectives, including the specific managed care strategies used, administrators’ impressions of the effects on child welfare systems, and contractors’ reports of their resulting financial status.

Endnotes

⁶  In states where less than the entire child welfare caseload is targeted by the fiscal reform, initiatives often target particular types of cases based on expected services or service intensity needed, such as intact at-risk families or seriously emotionally disturbed (SED) children in care. These differences in types of target populations are further discussed later in this report.

⁷  This section describes the assessment process used by initiatives in initial referrals and placements. Ongoing assessment of children and families is a critical part of delivering services but is not described here because it was not a focus of the review.

Chapter 3: Financial Arrangements of the Initiatives

The states' fiscal reform initiatives incorporate a variety of approaches to address accountability concerns, enhance financial flexibility, and achieve better performance. Financial arrangements--payment mechanisms, risk sharing, and risk management--attempted to redirect resources, encourage comprehensive services, and serve more children and families with the same funding levels as under the previous financial arrangements. The initiatives varied in the extent to which financial risk was transferred to private organizations, but most hoped to achieve better outcomes or cost savings through relying on contractors for much of the work that once was the responsibility of public agencies. In general, it is not yet known whether better outcomes were actually achieved by the initiatives. Cost savings were rare. Several initiatives provided financial rewards for contractors that achieved outcome standards or improved their performance and imposed penalties for contractors that did not.

Some initiatives reported concerns about potential or actual conflicts between fiscal and treatment considerations. Indeed, nearly all initiatives had or were working toward mechanisms for monitoring contractor performance and outcomes to prevent decisions that reduced costs by reducing treatment effectiveness. Many emphasized the importance of balancing the pressure to reduce costs, or to do more with the same amount of money, with an emphasis on improving child and family outcomes. As one state child welfare administrator said: “Privatization is a double-edged sword. We must never lose sight of our mission--to protect kids, not to save money.”

Payment Basis

Traditionally, public child welfare systems’ payment arrangements with private-sector service providers have been fee-for-service. Payments depend on both the type and the amount of service delivered. Embedded in this system, it has been claimed, is a perverse incentive for providers to deliver more reimbursable services than are needed or to prolong treatment beyond what is necessary. The crux of the argument against the fee-for-service system is that it encourages providers to use scarce resources inefficiently. Evidence used to support this argument includes long stays in foster care and lengthy wait lists for some services. Per diem payments, in which providers are paid for each day that service is delivered to a client, are based on the length of time that services are delivered (and often the type or intensity of services). As with fee-for-service payments, per diem payments may encourage the inefficient use of scarce resources when clients are provided services for longer than might be necessary if alternative arrangements were available.⁸ Under both these payment schemes, there is no financial incentive to change a service from one that is reimbursable to one that is not or that is reimbursable at a lower level. The states carry the financial risk for charges billed retrospectively for services already delivered.⁹

The perverse incentive argument underlies many states’ experimentation with alternative payments such as capitated rates, case rates, and block grants, which basically are prepayments for a service package. These payment methods allow some or all financial risk to be transferred to a private contractor, as payments are fixed and based on historical averages (and are sometimes dependent on geographic area and expected severity of need for services). They are made prospectively to cover all or a defined spread of services, which provides an incentive for contractors to control expenses in order to avoid losses and realize financial gains. Shifting from retrospective payment methods (fee-for-service and per diem) to prospective payments (capitated rates, case rates, and block grants) fundamentally changes the incentive system from one that offers incentives to retain cases on the caseload, to one with incentives for avoiding unnecessary placements or lengths of stay.¹⁰

Risk Source

Prospective payment systems in effect force the contractor to operate within a given budget or face financial loss?in managed care terminology, these schemes impose a financial risk on the contractor. The risks can be due to intensity, duration, or volume, all of which are discussed in more detail later in this section.¹¹ To shift the risk from the public child welfare agency to the private contractor, payments are fixed at a specified rate. The risk facing the contractor is that the costs of meeting the service needs of a group of clients may be greater than the payments for those services.

The types of risk-shifting payment methods that are most commonly used in states experimenting with managed care fiscal reforms are capitated rates, case rates, and block grants. Table 3-1 summarizes the source of risk faced by contractors in each payment method.

Table 3-1
Financial Risk Associated with Payment Methods

Payment Method	Retrospective	Prospective	Source of Risk to Contractor
Fee-for-service	X		None
Per diem	X		Intensity
Capitated rate		X	Intensity
Case rate		X	Intensity, duration
Block grant		X	Intensity, duration, volume

Capitated rates are paid on a per-case per-month basis?the contractor is paid monthly for all contracted services for an enrolled population. The contractor receives the predetermined monthly amount, based on a specified number of cases to be served, regardless of the level of services that the enrolled population requires. If the population requires more services or more intensive services than projected, the contractor faces financial risks. If there is an increase in the number of cases served, there would be an increase in payments; thus the contractor is not at risk for volume. And since the contractor is paid as long as services are provided, the contractor is not at risk based on duration of services; payments do not stop until cases are disenrolled. Similar in some ways to per diem payments, in that contractors under both payment mechanisms avoid volume and duration risk, capitated rates offer a flexibility that per diems do not. Contractors can change service intensity more easily and usually can offer wraparound services and other supports to enable a switch to lower-cost services or placements.

Case rates are a fixed fee paid to a contractor for all services delivered to a client over a treatment period or an episode of care. Contractors with case rate contracts are at financial risk if the intensity and duration of care are greater than expected. But they are not at financial risk if there is an increase in the number of cases served, since there would be an increase in payments. For example, a contractor may receive a flat case rate of $5,000 for each family referred; some families receive services for 3 months, and some receive services for 9 months, but the case rate is the same. The contractor receives the same payment amount for all the families.

Kansas’s initiative clarifies the difference between capitated rates and case rates. Kansas paid lead agencies an initial episode of care case rate for foster care and adoption. However, lead agencies experienced losses, and the state realized that some factors affecting permanency were beyond the control of the lead agencies. Kansas then changed to a capitated per child/per month payment system for foster care and adoption so that lead agencies no longer experience risk based on duration or lose money on children who do not move to permanency in a timely fashion. Contractors receive the monthly rate as long as a child receives services.

Unlike capitated and case rates in which contractors receive a payment for each case served, a block grant is a single payment that is made for a specified period, usually annually, for all cases served during the payment period. These types of payments are also called allocations, budget transfers, or capitation payments (not to be confused with capitated rates, described above). Under block grants, contractors may experience financial losses if the intensity, duration, or volume of service is greater than anticipated. For example, a contractor may receive an annual block grant and then must serve all referred cases in its jurisdiction, regardless of the number of cases or their intensity or duration of services.

Table 3-2 summarizes the fiscal characteristics of the initiatives. It shows, for each initiative, the payment basis, risk source, rate-setting method, risk management features, fiscal incentives, adequacy of payment (as reported by the contractor), and contractor’s financial status. Each of these is discussed below, except for payment basis and risk source, which were described previously. The overall cost of each initiative is not addressed, since complete information was not available.

Table 3-2
Fiscal Features of Initiatives
“NA” indicates that information is not available.

Initiative	Payment Basis	Risk Source	Rate-Setting Method	Risk Management	Fiscal Incentives	Payment Adequate	Contractor’s Financial Statusa/
Arizona Family Builders	Lead agencies receive a case rate, paid in three installments.	None (case closed if no progress)	Lead agencies’ cost estimates	Lead agencies close case if no progress after 6 months.	Case rate savings	Yes	Neutral
California Project Destiny	Lead agencies receive a monthly case rate for 2 years.	Intensity Duration	Historical costs for highest levels of care	Providers bear full risk but have some discretion over case decisionmaking.	Case rate savings	Yes	Neutral
Colorado Boulder County Managed Care Pilot Project	County receives block grant and negotiates providers’ allocations and fee-for-service rates.	Intensity Duration Volume	Historical data	County bears full risk.	Block grant savings	Yes	Neutral
Connecticut Continuum of Care	Lead agencies receive a case rate, paid in four installments; they pay providers fee-for-service.	Intensity Duration	State’s historical cost for residential treatment	Lead agencies bear full risk.	None; savings are returned to the State.	No	Losses
Florida Coalition for Children and Families	Lead agencies receive a block grant and must maintain time logs and justify their expenditures.	Intensity Duration Volume	Prorated based on case counts	A statewide risk pool can be tapped in cases of excess referrals or catastrophic service costs.	Lead agencies can receive “excess earnings” of federal reimbursements as bonuses.	Yes	Neutral
Georgia Metropolitan Atlanta Alliance for Children (MAAC)	Managed care organization receives a single per diem rate and pays providers per diems that were negotiated with the State.	Intensity	Average per diem for all levels of care	Managed care organization bears full risk but can refuse referrals.	Per diem savings	No	Losses
Illinois Performance Contracting	Providers receive monthly administrative payments based on expected caseload ratios.	Volume	Historical data	State bears full risk.	Providers surpassing permanency standards can receive incentives; those not achieving standards lose referrals.	NA	NA
Kansas Public Private Partnerships	Lead agencies receive capitated rates for foster care and adoption, and case rates for family preservation.	Intensity (foster care and adoption); Intensity and Duration (family preservation)	Historical data	Lead agencies bear full risk except that there is a risk corridor for foster care.	Capitated and case rate savings	NA	NA
Kentucky Quality Care	Lead agency receives a case rate.	Intensity	Lead agency’s cost estimate	A stop-loss provision protects the lead agency.	Case rate savings	Yes	NA
Maryland Baltimore Child Welfare Managed Care Project	Vendor receives a case rate.	Intensity Duration	State’s historical cost	A stop-loss provision protects the vendor.	Case rate savings	No	Neutral
Massachusetts Family-Based Services	Lead agencies receive block grant; service providers receive fee-for-service and per diem rates directly from the state; developing case rates.	Intensity Duration Volume	NA	Lead agencies bear full risk for cost of case management; state bears full risk for costs of services.	Block grant savings; lead agencies may lose their contracts if they spend the services budget too quickly.	Yes	Neutral
Michigan Michigan Families	Lead agencies receive case rates.	Intensity Duration	State’s average cost over all children in all levels of care	Lead agencies have a risk corridor, can accumulate dollars in a risk pool and have discretion over cases accepted.	Case rate savings	Yes	Gains
Michigan Permanency Focused Reimbursement System	Lead agencies receive case rates, partially based on performance, plus administrative per diems.	Intensity Duration	State’s overall average costs for 5 years + 15 percent	State bears financial risk. Providers risk not receiving incentive payments if placements are not successful.	Case rate savings, plus lead agencies can receive incentive payments for successful placements.	Yes	Gains
Minnesota PACT 4	Lead agency receives block grant from pooled funds.	Intensity Duration Volume	Based on county size and school enrollment	The counties bear full risk and can tap county reserves.	None	Yes	Neutral
Missouri Interdepartmental Initiative for Children with Severe Needs.	Managed care organization receives a monthly case rate for 6 months plus fixed case management payment.	Intensity	Historical costs of highest level of care	The state covers part of any loss experienced by the managed care organization.	Case rate savings	No	Losses
New York Safe and Timely Adoptions and Reunifications (STAR)	Provider agencies receive per diems.	None to providers	Historical data	The city bears full risk.	Providers can receive fiscal rewards based on permanency outcomes.	NA	NA
Ohio ProtectOhio	Managed care organizations receive case rates.	Intensity	Historical data	Risk corridors established; county bears risk beyond the corridors.	Case rate savings	No	NA
Oklahoma Oklahoma Children’s Services	Lead agencies and providers receive block grants.	None; state keeps referrals within established limits to avoid excessive costs to contractors.	Historical data on costs, number of children in care, and legislature’s allocation	Lead agencies stop serving families when block grant is spent.	None; savings are returned to the state.	No	Neutral
Pennsylvania Berkserve	County paid providers fee-for-service and paid the lead agency a percentage of billable services for administrative costs.	Administrative cost	Historical data	Planned to establish a risk corridor and case rates (initiative has ceased).	Case rate savings, when fully implemented	No	Losses
Tennessee Continuum of Care	Lead agencies receive per diems based on level of care.	Intensity	Independent time and cost study	Lead agencies bear full risk but can receive augmented rates in special cases and have some discretion over case decision-making.	For children that “step down,” agencies continue to receive initial per diem rate.	No	Losses
Texas Permanency Achieved Through Coordinated Efforts (PACE)	Lead agency received a per diem.	Intensity Volume	Average per diem	Lead agency bore full risk.	Lead agency could keep 10 percent of per diem savings.	No	Losses
Washington IV-E Waiver Demonstration	Lead agency received a monthly case rate.	Intensity	Negotiated with provider	Lead agency bore full risk.	Case rate savings	No	Losses
Wisconsin Bureau of Milwaukee Child Welfare	Lead agencies receive a case rate for in-home services and a block grant for out-of-home services (developing a case rate).	Duration	Case sample and needs assessment	If a lead agency experiences a deficit, the state will cover it as long as the agency is trying to control costs.	Lead agencies can receive fiscal rewards and penalties.	Yes	Neutral
a/ “Contractor” includes lead agencies, managed care organization, and service providers--any private nonprofit or for-profit organization that has a contract with the state to manage the delivery of services in order to achieve the objectives of the fiscal reform.

Rate Setting

Contractors’ financial risk arises from three sources: intensity (the level or costliness of services that must be provided), duration (the length of time that services must be provided), and volume (the number of clients who must be served). The principal source for the estimate of costs is historical data on the patterns of service usage and costs of providing services. The reliability of such data is clearly critical and notoriously poor. Beyond that, there is no consensus on a best method or formula for establishing payments that guarantees that the payment level itself will pose no financial risk to the contractor.

As can be seen in Table 3-2, states use a wide variety of methods to set payment rates, ranging from states’ historical costs for specific types of services, particular populations, or bundles of services across a sample of cases, to time and cost studies conducted by an independent entity. Generally, states use an average of some historical cost data for the populations and services, and sometimes the geographical area, that the payment is intended to cover. Whether the payment is adequate depends not only on the accuracy of the historical data but also on the appropriate selection of representative populations and services. For instance, in Baltimore’s Child Welfare Managed Care initiative, twice as many children received therapeutic care than had been included in the case sample on which the payment rate was based; as a result, the payment rate was lower than the actual cost of providing services. When this happens, contractors can attempt to get the rates raised or receive supplemental funding from the public agency, or they can cover the shortfall through other means such as fundraising from private sources. Otherwise the agencies may experience such financial losses that the initiatives cease operating, as did the initiatives in Texas and Washington.

In addition to using historical costs to set the payment rate, some states settle on a final rate after negotiating with the contractor. Other states increase the payment rate obtained from historical data by some percentage to take into account the possibility that rate-setting methods underestimate the cost. For instance, Michigan increased by 15 percent the payment that was based on the state’s overall average costs for 5 years.

States also base payments to contractors on their payments to non-initiative programs delivering services to a similar population. For example, California’s Project Destiny pays the same case rate to initiative contractors for the delivery of community-based wraparound services to children at risk of residential placement as they pay for residential placement. The objective of this payment system is to achieve cost neutrality. California also uses a control group to adjust case rates every 6 months and to ensure cost neutrality.¹²

Georgia and Texas use a variation of this payment-setting method. Both states’ initiative contractors are paid (or, in Texas, were paid, since the PACE initiative is no longer in operation) an average of the level-of-care per diems paid to non-initiative contractors. In Georgia, the managed care organization (MCO) contracts with a network of providers to deliver services in a variety of settings that range from regular foster care to residential treatment. The MCO receives the average of the range of per diems that the state pays directly to service providers. Then the MCO pays to its network members the same per diem rate that the state would pay them if the MCO were not the intermediary. In this payment arrangement, the MCO attempts to ensure that services are provided in the least costly setting. When services are delivered in higher-cost settings, the MCO pays the provider a higher per diem rate than the state pays the MCO. At the time of the interview, the MCO administrator reported that the initiative had a larger number of children in high-level care than had been anticipated. As a result, the MCO was facing a financial shortfall and taking proactive measures to reduce further loss. The administrator was both seeking an increase in the MCO’s per diem and avoiding entry of children with high-level needs into the initiative.

Since there clearly is no consensus on the best rate-setting method, the question arises as to how well the states have estimated the cost of services delivered by or through their contractors. One way to explore this question is to examine contractor reports of payment adequacy and the extent to which they have sustained financial gains or losses or achieved cost neutrality.¹³ However, this analysis strategy is somewhat problematic because, as will be shown later, contractors typically have at their disposal a variety of ways to manage whatever budget is given to them.

Risk and Risk Management

Contractors entering into non-fee-for-service payment contracts may take on either full or partial financial risk. In a full-risk contract, the contractor absorbs all losses incurred as a result of providing services above those covered by the state payment, regardless of whether additional services or higher levels of care are deemed necessary. In the states using this approach, the amount of risk that the contractor is subject to is not explicit in the contract, and, in fact, neither the state nor the contractor is able to estimate accurately the extent of potential risk. As is depicted in Table 3-2, several of the initiatives feature contracts in which the contractors bear substantial or full risk. Somewhat more often, the contracts either explicitly limit contractors’ financial risk or contain risk-sharing agreements. Many of the states acknowledged that contractors are reluctant to take on full financial risk due to the inability to estimate accurately what that would cost. Requiring that they take full risk would likely result in contractors’ being unwilling or unable to participate in the initiatives.

Partial-risk contracts either explicitly limit contractors’ financial risk or contain risk-sharing agreements. Of these two types of partial-risk contracts, risk sharing is more common. The terms of partial-risk contracts vary considerably. Some states establish a risk pool from which contractors may draw down additional funds if their total service expenditures exceed the overall payment by a stipulated percentage. For instance, Florida’s lead agencies can access the risk pool if the number of children entering care is 5 percent more than expected. Other states’ contracts contain stop-loss provisions that stipulate the percentages of the total loss for which the contractor and state are liable. Maryland, for example, is responsible for 90 percent of the costs that exceed the contractor’s payments. Another variation is a risk corridor, in which a contractor is liable for a percentage of excessive costs. Beyond this percentage the state picks up the costs, and the contractor keeps a similar percentage of savings and returns the rest to the state. For example, in the first year of Ohio’s initiative, the contractors were responsible for the first 5 percent of costs that exceeded revenues and could keep the first 5 percent of “excess” revenues; that percentage rose to 10 percent in the second year, and 15 percent in the third and subsequent years. The next 10 percent of excess costs or revenues are shared equally by the contractors and the county, and beyond that the county is responsible.

Risk Regulation

Each of the three types of risk (volume, intensity, and duration) can be regulated by the contract. For instance, a contract can stipulate the number of children to be served by the contractor for the contract period and thereby limit the contractor’s exposure to volume risk. The contract may, on the other hand, contain a no-reject, no-eject clause. That is, the contractor may be prohibited from refusing referrals or discharging clients without state approval. This arrangement obviously places the contractor at greater volume risk. Similarly, the contractors’ exposure to duration risk can be limited by stipulating the length of time that treatment is to be provided. For instance, many states’ contracts stipulate that contractors are responsible for children’s care for a specified period regardless of the level of care needed. This type of contract exposes the contractor to greater risk if, overall, the level of provided care costs more than the total payment.

Not all administrators provided enough detailed information about their contracts to determine the extent to which states are using these types of contractual mechanisms to limit contractors’ financial risk. Typically, however, initiative contractors are not protected from risk due to delivering higher levels of care. In fact, the primary objective of using fiscal risk arrangements in many of the initiatives is to reduce the level of care that is provided. Contractors usually attempt to accomplish this objective by providing services in the least restrictive and least costly setting, usually in the community. In these types of arrangements, contractors are somewhat protected from intensity risk if they have partial- rather than full-risk contracts.

Managing Risk

Besides the risk-sharing provisions of their contracts, a number of other features of state initiatives enable contractors to better manage fiscal risk. Generally, initiative contractors that have some control over case referral, decisionmaking, and service planning are able to use these features to stay within their limited budgets. For instance, contractors that have authority to refuse case referrals can regulate the number of children with high-end needs that enter their programs. According to several contractors, the authority to refuse referrals has been critical to their ability to manage expenditures. In Georgia’s MAAC, for example, refusing a referral, if it appears that the child is at risk of needing high-end services, is a primary mechanism for managing financial risk because MAAC remains responsible for providing whatever level of care children need after they are accepted. MAAC is more likely to refuse high-end service users if a large number of children already in its care are receiving intensive services such as residential treatment. Conversely, contractors with no-reject, no-eject contracts may receive more children with high-end needs than their fixed budgets can support. No-reject, no-eject contracts contributed to financial losses for Missouri’s contractor and the demise of Texas’ PACE initiative.

Although having some authority in the referral process may enable contractors to better manage their budgets, unless the target population is clearly identified and the state and the contractor agree on the target population, contractors may mis-target their selection of cases. For instance, Michigan is currently revising its Michigan Families contracts to clarify the target population in response to selection by contractors of lower-need families into the program instead of the high-need children that the state had intended the program to target.

The extent to which contractors have discretion over case decisionmaking, including level of care and services provided, also influences how well they can manage fixed budgets. For example, in state initiatives such as California’s Project Destiny and Tennessee’s Continuum of Care, contractor discretion over level of care and services is particularly important because payment rates are based on an average level of care, and the program objective is to reduce the level of care. In these types of initiatives, contractors typically have substantial decisionmaking discretion over both level of care and service planning. One way that contractors use their decisionmaking discretion to step children down to lower levels of care is by delivering intensive services in settings that are less expensive than residential facilities or group homes. In turn, the contractors’ ability to deliver services in alternative settings such as foster or biological parents’ homes is closely linked to flexibility in funding.

Flexible Funding

Unlike categorical funding that requires providers to use child welfare dollars to deliver specific services in particular treatment settings, flexible funding gives contractors freedom to deliver a wider range of services and move children more freely between treatment settings. With flexible funding, instead of applying a limited set of categorical services to every case, contractors can develop an individualized treatment plan for each child. Hence, not every child receives a set of expensive services when more limited services may meet the individual child’s needs. Also, since flexible dollars follow the child rather than the service, contractors can more easily shift the child between service settings. For example, contractors may decide to deliver intensive in-home services instead of placing a child in an expensive residential treatment setting. Alternatively, the child may be placed for a short time in residential treatment but then be quickly moved into a community setting with intensive services.

Among interviewed contractors, flexible funding and the individualized treatment that it makes possible was one of the more popular features of the initiatives. From the contractors’ perspective, flexible funding and individualized treatment are necessary conditions for making the best treatment decisions. But in managed care, fiscal constraint is also intended to influence contractors’ decisions. When fiscal constraint enters into treatment decisions, contractors may, perhaps unconsciously, use individualized treatment planning as a tool to manage their budgets.

Although most contractors reported that their clients’ essential service needs are usually met, other comments they made reveal an apparent conflict between treatment and fiscal considerations. For instance, one lead agency reported that at the start of the initiative, its strategy had been to provide intensive community-based services at the beginning of a case to avoid placing children in higher levels of care. However, some children ultimately entered residential treatment. Consequently, the lead agency incurred losses. From this experience, the lead agency “learned to hold back on up-front services in case a child needed residential treatment later in the case.” Another contractor said that it could work within its budget only if cases are triaged as the agency approaches its budget limit. Other contractors told us that although they don’t require case managers to work within a set case budget, if the agency is “headed for financial trouble, that [information] is shared with workers.” These contractors’ comments suggest that treatment decisions depend not only on individual service needs but also on a contractor’s financial status at the time a decision is made. Hence, a child entering into care at the beginning of a budget cycle may have a different treatment plan than a child with similar needs but who enters care when the budget is closer to depletion.

Community Resources

Similar to flexible funding and individualized treatment plans, many contractors rely on community resources and informal supports to both meet some of the needs of children and reduce the level of their own resources that would otherwise be used to meet those needs. Contractors often reported that one of their major roles was to assist the family to “build up their own community support” or “set up [community] services.” In fact, some states’ case rates are based on the assumption that the contractor will rely extensively on existing community resources. For instance, the state administrator of an initiative designed to move children from residential care into communities said that the case rate would be adequate if the MCO used existing community resources and natural supports. If, on the other hand, the MCO was unable to tap into other resources, it is presumed that the state payment would not cover the purchase of needed services. In another initiative that provides services to families with children at risk of entering placement, a major objective of case management is to link families with informal support--family, friends, churches, community organizations--so that overall the state would cover only about 25 percent of the costs of services, with 75 percent coming from local resources.

Many child welfare advocates have pointed to the importance of linking families to ongoing community and informal supports in maintaining children in their local communities. However, it is not clear that these community resources are good substitutes for child welfare services. In addition, the strategy of reducing child welfare expenditures by relying more on community resources assumes that communities are well equipped to assist troubled families. This may not be the case, and if not, contractors who count on community resources to reduce their expenditures could face budget shortfalls. Indeed, a lack of appropriate community resources could be one reason that some contractors have been unable to prevent residential placements and, as a result, have experienced financial losses.

Of course contractors do not rely solely on the mechanisms discussed above to manage their budgets. Many contractors have developed utilization management systems to help them regulate expenditures; these systems range in sophistication from simple to complex. The more simple systems consist of frequent case reviews that examine lengths of stay and levels of care and develop plans to reduce both. On the more complex end, one contractor (in Kentucky) has developed software to predict costs based on a family assessment and the types and lengths of services needed. This same contractor tracks all of the costs of providing services to a family and the balance of the case rate. One of Michigan’s contractors created a new position of utilization manager. The manager tracks how many children are receiving various services, the length of time children receive services, and the number of slots that are open. Also, the manager is responsible for approving services that the caseworkers provide. Despite the differing levels of sophistication of contractors’ utilization management systems, there is unanimous agreement among contractors that budget oversight receives greater attention under the new payment arrangements.

Fiscal Incentives, Payment Adequacy, and Contractors’ Financial Status

The last three columns of Table 3-2 summarize the initiatives’ fiscal incentives and contractors’ reports on the adequacy of the state payment rates and whether the agencies achieved cost neutrality, suffered losses, or reaped gains. Where prospective payments (capitated rate, case rate, or block grant) are used, contractors generally can experience savings when their costs are less than the payments (which offsets the risk they bear when their costs are more than the payments), although some states (such as Connecticut and Texas) limit the amount of cost savings that contractors can keep. Other contractor incentives are tied to contractor performance (Illinois) or permanency outcomes (Michigan, New York, and Wisconsin).

When assessing the contractors’ reports, it should be kept in mind that the initiatives have been operating over different time periods. Some initiatives were implemented in the mid-1990s, and others began as recently as January 2000. Those contractors with a longer operating history may have achieved equilibrium between losses and gains either as a result of their greater experience with a payment system or adjustments to the original payment rates and other aspects of their contracts. Recently implemented initiatives, on the other hand, may not have had enough time to master the new fiscal arrangements or make appropriate modifications to problematic aspects of their contracts. This explanation of the variety of experience with fiscal arrangements, however, is not consistently supported across states. Administrators of both older and more recently implemented initiatives report inadequate payment levels and financial shortfalls.

With regard to payment adequacy and financial status, states generally fall into four classes:

• those reporting adequate payment levels without reservation;
• those reporting adequate payment but with reservations;
• those reporting inadequate payment but still operating; and
• those reporting inadequate payments and whose initiatives are now defunct.

Of the contractors reporting that their payments were inadequate, those who had financial losses and abandoned the initiative are obviously the more severe instances.

Pennsylvania, Texas, and Washington are the three state initiatives that had large financial losses and were abandoned. In Pennsylvania, the lead agency administrator reported that the agency lost as much as $1,500 per month; however, the reason given for the initiative’s demise was not financial. Instead, both the state and lead agency administrators said that the program ended because of extraordinarily cumbersome administrative procedures, high lead agency staff attrition that necessitated recurrent intensive training, and public agency workers’ resistance. Of course, all of the cited reasons are likely to have driven the lead agency’s administrative costs upward.

The Texas initiative was terminated as a direct consequence of the payment rate. Generally, the average state reimbursement is 87 percent of any contractor’s actual cost. Contractors make up the difference through private fundraising. Because the initiative’s contractor received many more cases than it had anticipated, its board decided that the agency would be unable to make up the difference with private funding and asked the state to increase the per diem from $77 to $92. Because the state was willing to increase the rate to only $82/day, the contractor’s board did not renew the contract.

Washington’s initiative was terminated because the case rate was half of what the contractor actually needed to cover the costs of delivering services. In fact, the contractor estimated that the agency lost $80,000 a month on the program. According to the contractor, this large loss occurred primarily because many more children needing intensive services were referred to the program than had been anticipated.

Six other states report inadequate payment rates (Maryland and Oklahoma) and financial losses (Connecticut, Georgia, Missouri, and Tennessee), but continue to deliver services under their fiscal reform contracts. Among these six states, the reason cited most often for financial losses was the unexpectedly high volume of children needing more expensive services such as placement in residential treatment facilities and therapeutic foster care. Only one contractor reported sustaining financial losses because the agency delivered services that were not covered by the contract. The contractor believed that the services-in-home aftercare?were necessary to reduce children’s lengths of stay in residential treatment. However, he speculated that if these services had not been provided, the agency’s losses might have been even larger.

Only one of the four currently operating contractors with losses had explicit risk-sharing provisions in their contracts at the time of the loss, and that state (Missouri) covered part of the contractor’s financial loss. The remaining contractors address their financial losses in a number of other ways. Some attempt to make up losses through private fundraising. More often, agencies sustaining or anticipating a financial loss try to reduce their costs over the remainder of the year by providing fewer high-cost services or reducing the number of children in their care who need high-level services. At any rate, most contractors with or anticipating losses managed to stay in business.

Contractors reporting that their payments are adequate often did so with some reservation. For instance, contractors in three states commented that they were able to work within their budgets by triaging or rationing services. Two other contractors said that their payments were adequate to cover the costs of “essential” services but that additional funding would be needed in order to deliver longer-term rehabilitation or to reduce caseloads.

Contractors in only one state had actually gained financially from managed care fiscal reforms. In one of the two initiatives that Michigan has implemented, Permanency Focused Reimbursement System, the contractor attributed its gains to the fact that the performance-based case rates had increased faster than the alternative per diem rates that non-initiative contractors received. The agency used its unexpended payments to purchase independent research on the impact of services on different populations. Pending a review of other program outcomes, such as reducing placement disruption and lengths of stay, both the contractor and state consider this program to be successful.

Michigan’s other program, Michigan Families, has also resulted in financial gains for at least one contractor but is not considered to have been completely successful because the savings were achieved by serving a less needy population than the state had intended the program to target. Contractors put any unexpended funds into an agency risk pool, but it is very unlikely that these funds will be used for services unless the program serves more needy families. In response to the current situation, the state is revising the program contract and plans to seek bids from new contractors.

Conclusion

The above description of states’ attempts to reform their payment arrangements with service contractors to achieve multiple goals such as improving services, shortening lengths of stay in foster care, and reducing foster care costs, highlights the complexities that such initiatives engender. Reform efforts generally are not limited to changes in contractor reimbursements; they also encompass a shift from categorical to flexible funding. Flexible funding, in turn, enables contractors to individualize treatment plans and deliver different types of services in non-residential treatment settings. Components of the reform efforts--fiscal arrangements, flexible use of funding, and community-based services--are tightly linked. Flexible funding allows contractors greater discretion over treatment decisions, including what and where services are provided, and fixed budgets are intended to encourage contractors to deliver less expensive services in lower cost settings whenever appropriate given the child and family’s needs. To some extent then, the greater flexibility that these new arrangements afford enable contractors to control the costs of delivering services.

A majority of contractors managed to operate within the limited budgets imposed by fixed payment rates. Many of these contractors had greater discretion over intake, discharge, and treatment and were therefore in a better position to control the types and duration of services delivered and hence the agencies’ costs. Some contractors were also able to reduce their costs by supplementing the services that were provided directly with those available from other community agencies and informal supporters.

Nevertheless, in those states where contractors incurred substantial financial losses, payment rates obviously did not cover the cost of services that contractors determined were necessary to address the needs of the populations served. In some instances, for example Texas, the discrepancy between the payment rate and the contractor’s costs may have been a consequence of mis-targeting--serving a population with a larger proportion of high-end service users than was intended. On the other hand, state administrators argue that the payment rates would have been adequate if only the contractors had made the right decisions with regard to the types and duration of services delivered. This conflict in perspectives underscores the difficulty that both state administrators and contractors often face in accurately targeting particular services and predicting case trajectories.

Finally, there is the question of whether the fiscal reforms address the alleged perverse economic incentives of the fee for service system. Overall, contractors report that they are monitoring their budgets more carefully at both the program and case levels. At the case level, contractors generally attempt to avoid more costly services and residential settings whenever possible. If permitted and other circumstances are conducive, some contractors avoid serving potentially high cost cases in order to limit their financial risk. So, as intended, the fiscal reforms do appear to have some effect on contractors’ cost consciousness.

However, the fiscal reforms may introduce their own set of perverse outcomes. In particular, over-emphasis on the cost of providing services may unduly influence contractors’ decisions regarding problem assessment and treatment at the expense of effectiveness. For instance, if budgetary concerns blur providers’ judgments, they may tend to minimize the extensiveness of families’ problems and overestimate the effectiveness of weak treatments. This is, however, a potential dynamic, not a documented one. Assessing any perverse effects that the fiscal reforms may introduce will depend on implementing ongoing systems to carefully monitor child and family outcomes. Many states have thus far been hesitant to implement such systems primarily because there is no consensus about what outcomes should be monitored, and measurement tools are still evolving.

Endnotes

⁸  In addition, these payment mechanisms afford little flexibility in treatment; the services provided must be on the predefined list of reimbursable services. Many fiscal reforms attempted to open up the range of services through more flexible funding mechanisms.

⁹  Under per diem payments, providers may bear some financial risk if the services needed cost more than had been anticipated in setting the per diem, and “level-of-care” per diems are not available.

¹⁰  It is important to note that federal reimbursement under title IV-E for foster care days is viewed by many as a major impediment to implementing fiscal reforms in child welfare, due to IV-E’s categorical per diem reimbursement structure.

¹¹  In his article “Federal Fiscal Reform in Child Welfare Services,” Wulczyn (2000) identified volume, duration, and unit cost as the three variables that both determine the total cost and financial risk of providing child welfare services. In this report, level of care (intensity) is substituted for unit cost because none of the state-provider contracts in the states interviewed were based on the unit costs of individual services. Instead, payment rates are typically based on the average cost of providing bundles of services or levels of care to specific populations.

¹²  The purpose of Project Destiny’s financing arrangements are not to reduce costs, but to provide quality services in the least restrictive settings at no additional cost?to spend no more on nonresidential settings than they would have on residential treatment. In the long term, administrators hope to reduce costs as a result of fewer re-entries into the system and lower use of residential treatment facilities. They see the initiative as a long-term investment.

¹³  In some states, there was disagreement between state administrators and contractors’ reports regarding the adequacy of payment rates. Where there were such differences, state administrators typically reported that payment adequacy was contingent on good financial management. On the other hand, although they may generally be more inclined to report that payment rates are inadequate, most contractors substantiated their claims with specific ways in which underfunding had negatively affected the management and delivery of services and the financial status of programs.

Chapter 4: What Are the Challenges?

Recent fiscal reform initiatives have attempted to address some of the seemingly chronic problems of the child welfare system in the United States. The purposes of this review of fiscal reforms in child welfare are to identify how states are addressing the need to contain costs or improve system performance, and to see whether the changes appeared to be positive or negative for children and families. The 23 initiatives reviewed focus on altering the financial relationships between public child welfare agencies (states or counties) and organizations with which they contract for services. The altered relationships are intended to lead to greater efficiency in the use of resources, improved services, and better outcomes for children and families.

Unlike other child welfare initiatives, these reforms are not focused on technologies of service (e.g., family preservation). They do not, in and of themselves, provide new ways to solve the problems of families. Rather, they rely on a reorganization of existing methods. Also unlike other initiatives, these efforts do not tilt the system one way or the other on the continuum of protecting children from harm vs. upholding the integrity of the family. Rather, the initiatives are overlaid upon the existing system orientation.

How widespread is the adoption of fiscal reforms in child welfare? A few years ago, it was thought that managed care was about to sweep through the system, revolutionizing the field. It is evident that the impact of these ideas has been somewhat more modest. In a Child Welfare League of America survey in 1998, it was estimated that in the 29 states with initiatives identified at that time, only about 10 percent of the nation’s child welfare population was affected.¹⁴ The scope of fiscal reform efforts varies considerably across states. In some states (e.g., Kansas), the program covers most child welfare services across the state (except for initial intake and investigation, which was retained as a responsibility of state workers in all states). In other states, programs deal with small numbers of cases in limited geographical areas. Some states continue to expand the scope of their programs while others have pulled back. Several programs include children and families in systems other than child welfare.

One finding that emerged from the review of fiscal reforms was that, despite a concern that focusing on fiscal aspects of child welfare systems will lessen the focus on children and families, that does not appear to be what happened in the states reviewed. An integral part of the initiatives seems to be a push to do things better for the children and families served, or at least not to allow things to get worse for them when money is being saved. Many of the initiatives have implemented close monitoring of outcomes, quality-assurance procedures, performance contracting, client satisfaction surveys, evaluations, and other mechanisms for helping to ensure that saving money does not become more important than helping children and families. However, if some initiatives do not live up to early expectations that they will reduce costs, or if early success creates pressure to save more money or do even more with the same amount of money, it will be important to emphasize accountability and effectiveness. Fiscal considerations must not be allowed to overshadow objectives of improved well-being of children and families.

Available evidence does not support the conclusion that the fiscal reforms have had a major direct impact on outcomes, although impressionistic and anecdotal information points to some efficiencies and improvements in permanency outcomes. However, the initiatives’ overall impact may be difficult to quantify and document. Many states experienced a “jump-start” in child welfare when they implemented their fiscal reforms; as one administrator stated, the fiscal reform “got them out of the box and encouraged them to discover creative ways to do things.” This may be the greatest effect of the fiscal reforms, and it may not influence outcomes until much later, as states experiment with innovations and make improvements over time.

Another finding was that ongoing problems in child welfare are not necessarily eliminated by changes in fiscal relationships. In fact, the fiscal reforms may even highlight or exacerbate these problems, some of which are discussed later in this chapter. An example is in establishing rates and incentives (further discussed below); a state must clearly identify what it wants to buy and what is needed to buy it. Although the child welfare outcomes of safety, permanency, and well-being are often the goals, there is little agreement on specific outcomes. For example, does a state want to buy basic safety or does it want to achieve improvement in longstanding situations? The answer will greatly affect the design and costs of services.

At any rate, it appears that fiscal reforms are likely to be part of child welfare’s future, as agencies strive to improve child and family outcomes and reduce inefficiencies in the face of escalating costs. As fiscal strategies are adopted, there are many challenges to address; some of the major ones (discussed below) include the following:

• structuring risks and rewards;
• gaining flexibility;
• obtaining data;
• maintaining accountability and effectiveness; and
• changing people and organizations.

Since it seems that fiscal strategies are here to stay, at least in the foreseeable future, these issues need further consideration at both program and policy levels if the initiatives are to have a positive impact.

Structuring Risks and Rewards

How much should states pay for services? How should costs be apportioned?

The success of a fiscal reform initiative in a child welfare system depends to a great extent on the effective structuring of risks and rewards. The issues are both practical and philosophical. The practical issue involves establishing adequate rates and incentives for contractors and service providers: if the rates and incentives are inadequate, or if the risks and penalties are too severe, services will be inadequate or will not be available at all. Of course, the intent of most of the initiatives is to reduce inefficiencies, so the rates and incentives must not be set too high or they will not exert the necessary pressure to squeeze out excesses or increase efficiencies. The philosophical issue addresses the question of who should bear the responsibility for child welfare services: some fiscal reforms incorporate the expectation (or experience the result) that part of child welfare services is supported by sources outside the public child welfare system. Obviously, practical and philosophical issues are intertwined; setting inadequate payment rates can lead to incentives to shift costs to other systems such as mental health, juvenile justice, or special education systems. Here the issues are separated for the purposes of discussion.

Establishing realistic and adequate rates and incentives for contractors and service providers requires the ability to project utilization and risk. Medical managed care has actuarial and health care utilization data to use in pricing services; child welfare has no such database and no proven formula to use in projecting the costs of serving children and families. In addition, the pricing of child welfare reforms can be restricted by the categorical nature of much child welfare funding, which may preclude certain services, prevent collaborative arrangements, and create cash-flow problems for states due to retrospective payment methods.¹⁵

In setting payment rates and incentives under their initiatives, states generally relied on historical data on the costs of providing services for the targeted population in the covered geographic area. However, this often proved inadequate in projecting future costs. For example, the rate paid under the Baltimore managed care pilot was based on 3 years of historical data; however, the data were misleading because (1) twice as many children in the initiative needed therapeutic care as children in the historical data, and (2) the contractor was required to provide some services that were not in the historical data (including an intermediate level of care). Thus the actual caseload has required more expensive services than were predicted by the historical data.

A related issue is that states must balance the goal of limiting costs with the need for establishing and maintaining a stable and responsive provider community. In many communities, states must build provider capacity to address the full spectrum of needed services or serve the entire range of targeted children or families before a fiscal reform can be successful. And clearly it will not help states achieve their goals if the provider network fluctuates widely over time in the number of providers and their ability to deliver effective services. Although the adequacy of the provider community is influenced by many factors, payment rates are central. Rates must reflect both the actual costs of providing services and the investment required to support an effective provider community. In Florida, for instance, the legislature’s requirement that child welfare be privatized statewide has presented the challenge of building local provider capacity to take on complex child welfare responsibilities.

States found ways to address the rate-setting challenge. Recognizing that historical data may not produce realistic current rates, many initiatives limit the costs that contractors have to cover beyond a threshold (through risk sharing, risk corridors, and risk pools), or limit or exclude children likely to require extremely high levels of services. This can help maintain contractors’ fiscal viability and a stable provider community, but may result in higher costs to the state or in children or families not receiving needed services. Sometimes funding levels are sufficient to achieve safety and permanency at a minimal level, but true rehabilitation or long-term turnaround of chronic situations would require a much higher rate.

Several initiatives adjusted the rates or payment model after a period of operation, to re-align the rates with actual cost experience. Kansas, for example, changed from a case rate model to a capited per/child, per/month rate for foster care and adoption contracts, when contractors experienced fiscal difficulties under the case rate system. The state realized that achieving timely permanency was influenced by many factors beyond the contractors’ control.

Other states assume that the rates paid under the initiatives will not cover the costs of services and that contractors will make other provisions. This brings the discussion to the philosophical issue concerning responsibility for child welfare costs. In some of these states, contractors are expected to obtain funding from other sources to cover shortfalls; these sources include federal funding streams (especially TANF and Medicaid) and grants from foundations and philanthropic organizations. The demands placed by time-consuming fundraising activities or complex reporting requirements of other funding sources can greatly affect contractors’ operations (and were a factor in the demise of the initiative in Texas), while expecting a contractor to cover larger shortfalls than anticipated may force the contractor out of the initiative (as in Washington).

Some initiatives (such as the one in Massachusetts) incorporate expectations that community resources and natural supports will be utilized extensively. Whether these expectations can be met depends on high-quality resources being available. Although in the United States there is a long history of private-sector and community support for child welfare activities, the actual availability and amounts of contributions vary considerably across states and localities. Some reformers argue that communities ought to take responsibility for the welfare of their children, but there remains the question regarding whether this responsibility ought to reside in the state or in local communities.

It appears that, until adequate historical data and cost formulas are available, states will need to incorporate some flexibility in their initiatives. This includes risk sharing arrangements and mechanisms for adjusting rates periodically. States will have to address individually the philosophical question regarding responsibility for child welfare, although it is an issue that also should be addressed nationally.

Gaining Flexibility

How can states do things differently to obtain better results?

Some believe that there are major constraints with current federal title IV-E funding that make it difficult for states to experiment with innovative reforms and achieve better outcomes or improved system functioning. For example, title IV-E is limited in terms of what it will cover; it reimburses for out-of-home care and not for prevention or wraparound services that may eliminate the need for costly out-of-home care. It also reimburses only for services already delivered, which can create cash-flow problems for states that want to try prospective payment mechanisms. The results are a fiscal incentive to place children in foster care and keep them there in order to have a steady federal funding stream and little incentive to put resources into supporting families to keep children at home or “stepping down” children into less intensive (and lower reimbursed) levels of care, when necessary supports would not be reimbursed.

Most of the initiatives found ways to incorporate flexibility in the use of resources. One way to both avoid categorical funding restrictions and increase available funds is to pool or integrate funding from several public agencies and finance the initiative through this more flexible and comprehensive mechanism. Reflecting the networking and collaborative activity seen at the provider level over the past several years, collaboration at the funding level can result in an integrated system of care that addresses a full array of child and family needs. Service integration is not a new approach to providing services to families – it has been around for a while – but many of the fiscal reform initiatives re-energized the concept. This type of collaboration requires clear role definition and the development of trust, which takes considerable time and effort, but the result can be new levels of responsiveness to complex needs, an increased flexibility to focus on outcomes, and broader cooperation and enhanced relationships across agencies generally. As one program administrator described the integrated funding initiative in his state, “this is the wave of the future—the way child welfare should be run.”

Colorado’s initiative in Boulder County, for example, established a new entity that institutionalizes interagency partnerships and integrates county funding from mental health, corrections, social services, public health, substance abuse services, and other community services. This allows the interagency treatment planning team to reduce duplication and provide a variety of flexible wraparound and in-home services. Minnesota’s collaborative initiatives allow the integration of funding from two state departments (human services and education) to provide services for children prenatally through age 21. And Missouri’s Interdepartmental Initiative for Children with Severe Needs integrates funding from state social services, mental health, health, and education departments to support comprehensive, unified care for children likely to need services funded by multiple state agencies.

Another way to avoid categorical funding restrictions is to obtain a title IV-E waiver, in which the federal government waives certain restrictions and allows a broader and more flexible use of federal funds. The purpose is to test innovative strategies (including but not limited to managed care initiatives). However, waivers have their own set of restrictions, including a government preference for an experimental design (control group and random assignment), a requirement for an independent evaluation, and a limited authorization period (at the end of the waiver period, states have to revert to their previous way of operating their system unless they receive extensions). These restrictions have discouraged some states from developing initiatives that would allow funding flexibility, or at least discouraged them from operating such initiatives under a waiver.

Florida, for example, was granted a IV-E waiver for its privatization initiatives; however, the state decided to proceed with privatization without implementing the waiver, due at least in part to the stringent evaluation requirements (Florida Department of Children and Families, 2001). Washington’s waiver initiative has ceased operation, and one factor in its demise was the evaluation requirements. Texas also was granted a IV-E waiver, but by the time the waiver was granted the initiative had already been in operation for a year and some shifting in services had occurred, so that the waiver demonstration was no longer perceived as financially advantageous and the state withdrew it. Five other states currently utilize IV-E waivers as platforms for their fiscal reforms and the flexibility afforded by their initiatives (California, Connecticut, Maryland, Michigan, and Ohio).¹⁶

In general, states that did not utilize IV-E waivers for their fiscal reforms did not need waivers. Some initiatives targeted children not in foster care; e.g., Massachusetts covers 75 percent of all children and families in the child welfare system but excludes most children in care. Others tapped state or local general revenue funds and other funding streams rather than IV-E funds, such as New York City’s initiative in which the flexible dollars are funded through general operating revenue. Others operated initiatives that did not require deviation from IV-E requirements, such as Illinois’s initiative in which foster care payments are still covered by IV-E (Illinois operates three separate IV-E waivers in other child welfare reform efforts).

Obtaining Data

How will states know they’re doing better?

Inadequate data on service needs, utilization, costs, performance, and outcomes plague states’ attempts to implement child welfare fiscal reforms. Public agencies need data to design and manage their initiatives, and contractors need data to improve performance and satisfy reporting requirements. In particular, as child welfare agencies move from documenting process to measuring results, information on outcomes and performance is needed. And in order for public agencies to establish reasonable payment rates and performance standards, they must have good service, cost, and outcome data.

However, few initiatives have the necessary management information systems (MIS’s) to provide timely access to all the needed data. Although numerous initiatives relied on MIS data for contract monitoring and/or case decisionmaking, the data produced tended to be too limited and the systems too inflexible to be useful for assessing the impact of the reforms. In many cases, public agencies and contractors are working inefficiently with incompatible systems, and both have difficulty buying or developing systems that respond to their needs. Substantial investment is needed in hardware, software, and training to ensure that information technology is available and used for system improvement. To be most useful, an MIS should:

• Be performance-based and capable of tracking children, families, and providers on a timely basis;
• Incorporate the perspective of various users (e.g., able to track service utilization, costs, client status, and outcomes; handle billing and reimbursement; and provide user-driven reports);
• Link with or be part of the state’s Statewide Automated Child Welfare Information System (SACWIS)¹⁷;
• Include privacy protections for clients and families; and
• Be compatible between various government and service system entities.

Maintaining Accountability and Effectiveness

Are things getting better for children and families?

In child welfare, there appears to be an awareness that fiscal considerations should not be allowed to outweigh objectives of improved well-being of children and families; i.e., a more important goal than controlling costs is improving services for children and families. In fact, spending increased under several of the initiatives. Although data systems are rarely adequate for allowing state and local agencies to monitor results, establish performance standards, and link performance to financial incentives (see previous subsection), many states are taking steps toward establishing accountability for performance and results. However, not much is known definitively about the effects of the initiatives.

Some states have mechanisms in place to monitor their initiatives. Wisconsin, for example, tracks its initiative in Milwaukee County through extensive state monitoring and fiscal reviews that include analyses of permanency plans to ensure goals are being met. As previously noted, it will be important to maintain the focus on child and family well-being as fiscal reforms continue to try to improve efficiency and system performance.

Most of the initiatives track at least some measures covering established child welfare outcomes (i.e., safety, permanency, and well-being). The way those outcomes are measured depends on the specific initiative’s objectives and target population. For example, Georgia’s Project MATCH’s standards are that 40 percent of the children will improve their functioning and be discharged to a less restrictive placement setting, and a 20-percent reduction will occur in the frequency with which children harm others. A few states have established systems for linking financial rewards and penalties to outcomes. In New York, for example, the Safe and Timely Adoptions and Reunifications (STAR) program provides flexible dollars that agencies can allocate to a broad array of services to achieve timely permanency for children in placement. Agencies can obtain flexible dollars if they are able to show improvements in the length of stay for children in their care; improvement is defined as an increased discharge rate into permanent homes without a corresponding increase in re-entries and transfers to other agencies.

Although positive outcome changes have been documented in several of the states with fiscal reform initiatives, it usually is unknown whether the changes can be attributed to the fiscal reforms because the initiatives have not been rigorously evaluated. Illinois experienced a dramatic reduction in caseload, from over 51,000 to less than 30,000 in 3 years, after implementing the Foster Care Performance Contracting initiative. Milwaukee County, Wisconsin, has begun to see shorter placements for new families and some cost savings. However, these trends have not yet been documented or analyzed by an evaluation, and isolating the impact of each initiative is difficult, especially in places where multiple initiatives are operating.

Where initiatives have been evaluated, findings are mixed regarding outcomes. These initiatives include:

• Arizona’s Family Builders Program, where an evaluation showed that the initiative did not increase children’s safety in their homes overall, although families who accepted services experienced a slight but statistically significant decrease in the risk for child abuse and neglect (Arizona Office of the Auditor General, 2000);
• Colorado’s Boulder County initiative, where evaluation findings show that managed care has not yet had a discernable effect on outcomes for children and families (Mercer, 2000);
• Florida’s Community-Based Care initiative, where early evaluation findings show that initiative counties did at least as well as the comparison counties on the outcome indicators and they spent fewer dollars on direct child welfare services (Paulson et al., 2002); and
• Kansas’s privatization initiative, where the evaluation has found that most of the state’s performance standards have been achieved by the contractors, with the notable exceptions of (1) rates of re-entry into care within 12 months of reunification, and (2) permanent placement within 6 months of referral, where performance has been consistently below standards (James Bell Associates, 2001).

Several other initiatives have evaluations currently under way, and more will be known about the effects of fiscal reforms when those evaluations are concluded. Fully ensuring accountability for outcomes and attributing outcomes to fiscal reforms requires rigorous evaluation. However, performance assessment (carefully defining and consistently tracking specific outcomes and performance indicators) or performance contracting (tying performance on outcomes to financial rewards and penalties) can serve as an interim step on the way to evaluation or as a way to monitor overall trends in outcomes. Fundamentally, it is essential to stress that fiscal considerations, and attention to proximate system performance indicators, not be allowed to overshadow objectives of improved well-being of children and families. Quality-control mechanisms that ensure continual attention to those objectives must be enforced.

Changing People and Organizations

How do states get there from here?

The fiscal reform initiatives described here often required complex system changes and resolution of major interorganizational issues. Reforming the fiscal structure does not automatically bring about changes in performance, efficiency, or cost. Those require fundamental changes in both people and organizations. When reforms are implemented, many forces are put into play, not all of which are under the control of policymakers. Organizations and individuals often seek adaptations to altered structures that result in minimizing changes. They seek to regain the previous equilibrium and revert to doing what they are accustomed to rather than really creating change. Success depends on the receptiveness and capacity of organizations and people to make the desired changes.

Privatization, team decisionmaking, outcome measurement, and organizational collaboration were the primary changes incorporated into the 23 initiatives. Each requires complicated change processes as relationships, roles, and responsibilities evolve. This process includes the following:

• Shifting service delivery from public agencies to private contractors requires that (1) public agency staff switch their role from delivering services to monitoring service delivery and (2) provider staff change their service delivery, funding mechanisms, and reporting requirements. Often different sets of staff skills are required, and retention may become a problem as staff decide that the change process is too difficult and move on to other jobs.
• Implementing collaborative or team decisionmaking about cases takes the responsibility away from individual caseworkers, who often are then expected to take on responsibility for case management and provider monitoring. Again, a different set of staff skills is required and a different “culture” for case decisionmaking is created. Sometimes staff turnover is required in order to bring in new staff who support the innovations.
• Switching from a system that measures processes and outputs to one that emphasizes outcomes represents a paradigm shift. It requires letting go of a preoccupation with prescribed processes and moving to a focus on achieving desired outcomes and results. It often requires changes in priorities and business procedures, including major redesigns of data collection systems. Initiatives that feature performance contracting in which payments are tied to outcomes most explicitly require this shift, but other types of initiatives also monitored outcomes and had expectations for performance and results. Although for several years a focus on outcomes has been expected⁽¹⁸⁾, the actual operationalization has been inconsistent and is still in process.
• Bringing together a range of participating organizations to design and implement an initiative requires clear communication channels and mechanisms for negotiating disputes. Depending on the range of stakeholders involved in the initiative, there may be major philosophical differences to be resolved. This process can take a long time and require extensive negotiation. Collaboration is popular in social services today; it is thought to help in making the best use of available resources, leveraging an agency’s resources, and expanding the range of organizations concerned with child welfare--and working together is valued as a good thing to do. However, all parties must perceive benefits, give up some autonomy, and be willing to invest time and energy.

State and federal involvement can support the necessary change processes by providing training and technical assistance, disseminating written products, allowing sufficient start-up funds (including funds for MIS development and implementation), adopting realistic implementation schedules, and convening forums to discuss emerging issues and needed policy decisions. This support will help the child welfare system of the future evolve in ways that correct current inadequacies and benefit children and families.

Endnotes

¹⁴  CWLA has updated its data through a survey in 2000 (McCullough and Schmitt, 2001). That study identified 39 initiatives in 25 states. CWLA did not estimate the national scope of initiatives in 2000.

¹⁵  States receive reimbursement for the federal share of costs only after the delivery of services. Federal title IV-E funding, which finances most foster care, cannot be used for in-home or prevention services; when a child returns home, federal reimbursement for foster care costs ceases even if the child and family continue receiving in-home services. Title IV-B, the primary federal child welfare funding source for in-home interventions, is capped, and the amount available is limited.

¹⁶  Several other states also operate IV-E waivers, but the fiscal reforms reviewed in this report were not based on those waivers.

¹⁷  A recent GAO study (U.S. GAO, 2000) showed that none of the 12 initiatives contacted were using their SACWIS to manage information on clients, services, or outcomes, although some states hoped to eventually either link their initiatives’ data with SACWIS or incorporate SACWIS into their initiatives.

¹⁸  An outcomes focus for federal agencies has been required since the passage of the Government Performance and Results Act of 1993 (GPRA), which forced a shift away from traditional concerns such as staffing and activity levels and toward the single overriding issue of results.

Bibliography

AFCARS (2001). The AFCARS report interim FY 1999 estimates as of June 2001. Available at: http://www.acf.dhhs.gov/programs/cb/publications/afcars/june2001.htm.

Arizona Office of the Auditor General (2000). Evaluation: Family Builders pilot program. Report No. 00-4. Phoenix, AZ: Officer of the Auditor General.

Curtis, P. A. (1999). Introduction: The chronic nature of the foster care crisis. In P.A. Curtis, G. Dale, Jr., and J.C. Kendall (Eds.), The foster care crisis: Translating research into policy and practice (pp. 1-14). Lincoln, NE: University of Nebraska Press.

Curtis, P. A., Boyd, J.D., Liepold, M., and Petit, M. (1995). Child abuse and neglect: A look at the states – The CWLA stat book. Washington, DC: Child Welfare League of America.

Eddy, D. M. (1994). Clinical decision making: From theory to practice. In P. R. Lee, C. L. Estes, and N. Ramsay (Eds.), The nation’s health. Boston, MA: Jones and Bartlett.

Emenhiser, D., Barker, R., and DeWoody, M. (1995). Managed care: An agency guide to surviving and thriving. Washington, DC: Child Welfare League of America.

Florida Department of Children and Families (2001). Issue summaries: Health and human services. Tallahassee, FL: Department of Children and Families.

Goerge, R., and Wulczyn, F. (1990). Placement duration and foster care reentry in New York and Illinois. Chicago, IL: Chapin Hall Center for Children at the University of Chicago.

Hurley, R. E. (1998). Have we overdosed on a panacea?: Reflections on the evolution of Medicaid managed care. In S. M. Davidson and S. Somers (Eds.), Remaking Medicaid: Managed care for the public good (pp. 20-40). San Francisco, CA: Jossey-Bass.

James Bell Associates (2001). External evaluation of the Kansas child welfare system: July 2000-June 2001 FY 2001 final report. Unpublished.

Kahn, A. J., and Kamerman, S. B. (1996). Children and their families in big cities: Strategies for service reform. New York, NY: Columbia University School of Social Work.

Littell, J., and Schuerman, J. (1995). A synthesis of research on family preservation and family reunification programs. Washington, DC: U.S. Department of Health and Human Services.

Master, R. J. (1998). Medicaid managed care and disabled populations. In S. M. Davidson and S. Somers (Eds.), Remaking Medicaid: Managed care for the public good (pp. 100-117). San Francisco, CA: Jossey-Bass.

McCullough, C., and Schmitt, B. (2001). 2000 management, finance, and contracting survey. Washington, DC: Child Welfare League of America.

McCullough, C., and Schmitt, B. (1999). Managed care and privatization child welfare tracking project: 1998 state and county survey results. Washington, DC: Child Welfare League of America.

NCANDS, U.S. Department of Health and Human Services, Administration on Children, Youth and Families (1999). Child maltreatment 1997: Reports from the states to the national child abuse and neglect data system. Washington, DC: U.S. Government Printing Office.

NCANDS, U.S. Department of Health and Human Services, Administration on Children, Youth and Families (2000). Child maltreatment 1998: Reports from the states to the national child abuse and neglect data system. Washington, DC: U.S. Government Printing Office.

NCANDS, U.S. Department of Health and Human Services, Administration on Children, Youth and Families (2001). Child maltreatment 1999. Washington, DC: U.S. Government Printing Office.

Paulson, R.I., Armstrong, M., Brown, E., Jordan, N., Kershaw, M.A., Vargo, A., and Yampolskaya, S. (2002). Submission to the Florida legislature: Evaluation of the Florida department of children and families community-based care initiative in Manatee, Sarasota, Pinellas and Pasco counties. Tampa, FL: University of South Florida.

Rosenbaum, S. (1998). Negotiating the new health system: A nationwide analysis of Medicaid managed care contracts. In S. M. Davidson and S. Somers (Eds.), Remaking Medicaid: Managed care for the public good (pp. 197-218). San Francisco, CA: Jossey-Bass.

Scallet, L., Brach, C., and Steel, E. (1997). Managed care: Challenges for children and family services. Baltimore, MD: The Annie E. Casey Foundation.

Schuerman, J.R., Rossi, P.H., and Budde, S. (1999). Decisions on placement and family preservation: Agreement and targeting. Evaluation Review, 23(6), 599-618.

Schuerman, J.R., Rzepnicki, T.L., and Littell, J.H. (1994). Putting families first: An experiment in family preservation. Hawthorne, NY: Gruyter, Inc.

U.S. Department of Health and Human Services (2001). Evaluation of family preservation and reunification programs: Interim report. Available at: http://aspe.hhs.gov/hsp/hspyoung.htm.

U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau (1999). Child welfare outcomes 1999: Annual report. Washington, DC: U.S. Government Printing Office.

U.S. General Accounting Office (GAO) (July 2000). Child welfare: New financing and service strategies hold promise, but effects unknown. GAO/T-HEHS-00-158.

Waldfogel, J. (1998). The future of child protection: How to break the cycle of abuse and neglect. Cambridge, MA: Harvard University Press.

Westat and Chapin Hall. (1998). Evaluation of the New York City HomeRebuilders demonstration: Final report. Available at: http://aspe.hhs.gov/hsp/NYC-HomeRebuilders99/index.htm

William M. Mercer (2001). Colorado Child Welfare Evaluation: Second Interim Implementation Status Report. Available at: http://www.cdhs.state.co.us/cyf/cwelfare/2nd percent20interim percent20report.pdf

Wulczyn, F.W. (2000). Federal fiscal reform in child welfare services. Children and Youth Services Review, 22(2): 131-159.

Wulczyn, F.W., Hislop, K.B., and Goerge, R.M. (2000). Foster care dynamics 1983-1993: A report from the multistate foster care data archive. Chicago, IL: Chapin Hall Center for Children at the University of Chicago.

Appendix A: Summaries of Fiscal Reform Initiatives

Arizona: Family Builders

The goal of Arizona’s initiative Family Builders is to enhance parents’ ability to create safe, stable, and nurturing home environments that promote the safety of all family members and healthy child development. The initiative was implemented first in two urban areas, in January 1998. Other districts were added in 1999; currently it covers four districts out of six. The initiative is a public/private partnership that provides an alternative response system to families with low or potential risk child abuse or neglect reports. (Prior to Family Builders, most of these families were not being served at all.)

When a report comes in to the state, a state employee screens the report, and if it is judged to be low or potential risk, decides which Family Builders contractor the report should go to. The contractor then has 48 hours to contact the family. The family participates voluntarily, and many (as many as two thirds) decline to participate. The initiative uses a strength-based, family-centered approach to reduce the re-occurrences of subsequent child abuse and neglect reports.

The state has contracts with eight community providers to receive referrals, assess families, and provide the services identified in the service plans. Services that contractors must have available include family assessment, housing search and relocation, emergency services, intensive family preservation services, case management, parenting skills, parent aide services, child day care, transportation, respite services, shelter services, and supportive intervention/guidance counseling. The state monitors the contracts to ensure that services are provided as specified in the treatment plan. Contractors are paid a case rate, and bill the state at three points: referral, assessment, and completion of service plan.

In the first 18 months of operation, a total of 8,335 families were referred to Family Builders contractors. Of those families, 5,600 families declined services, 2,800 families received an assessment, and 2,300 families received services and completed a service plan. An evaluation conducted by the Arizona Office of the Auditor General (2000) found that families participating in the program were just as likely to have a new child maltreatment report after entering the program as two similar groups of families not participating in the program. However, caseworker assessments indicated that the risk of child maltreatment among participants decreased slightly.

California: Project DESTINY

Alameda County Department of Children and Family Services’ Project DESTINY (Demonstrating Effective Strategies for Intensive Placement Youth) was first implemented in 1997 as a small pilot project to support children living in group homes and residential treatment, and attempt to move them to the least restrictive environment whenever possible. The pilot project was expanded in 1999 as part of the California Title IV-E Child Welfare Waiver Demonstration Project – Intensive Services Component. Under the waiver, federal title IV-E funds may be used flexibly to pay for services not traditionally covered by title IV-E.

The project targets children between the ages of six and seventeen who:

• Are in highest level of care group homes, whose placement is disrupting, and who are likely to be placed in another high level residential facility;
• Are below age ten and who are living in group homes;
• Are probation wards between the ages of 10 and 14 and living in group homes;
• Are entering placement and are referred for a high level of care group home;
• Require 24-hour supervision to ensure safety;
• Need intensive services for more than six months; or
• Require short-term intensive hospitalization or hospital alternative assessment and diagnostic services.

Alameda County contracts with three private agencies (Seneca Center, Lincoln Child Center, and Fred Finch Youth Center), collectively called the FlexCare Consortium, to provide case management and wraparound services to eligible children. Generally, the county screens every high-level group home referral to determine whether the child is eligible for Project Destiny. If eligible, the case is referred to the FlexCare Consortium for review and randomly assigned to one of the three private agencies. Currently, the three agencies serve approximately 90 children. However, the county has appropriated funds to serve 256 children over the entire waiver period.

Service providers use a child and family team model to engage families and their natural support systems (i.e., extended family, neighbors, and friends) in service planning and delivery. Flexible funding enables providers to deliver an array of services such as parent advocacy, social and recreational rehabilitation in the community, 24-hour crisis response, respite, parenting skills activities, individual and family therapy in the home or community, and collaboration between community agencies, family, and mental health counselors.

The county reimburses providers a capitated rate of $4,082 per child per month, for a two-year period. During that time, providers remain responsible for delivering all services that the child may need regardless of the placement setting. Hence, the payment system encourages providers to deliver services in least restrictive community settings and to minimize the length of stay in high level and expensive residential treatment.

Case oversight is provided through the Case Management Workgroup, which is composed of both county and FlexCare case management staff. This group meets weekly to discuss ongoing cases and review new referrals. In addition, the University of California-Berkeley has been engaged to conduct an experimental evaluation of the project. UC Berkeley will rely on Department of Children and Family Services’ administrative data, a number of family assessment tools, and in-depth family interviews to evaluate the project.

Colorado: Boulder County Managed Care Pilot

From 1991-1997, the State of Colorado funded 90 percent of the costs for each child who met out-of-home placement criteria. In 1991-1992, out-of-home expenditures were $40.9 million; by 1996-1997, that had risen to $125.9 million. The legislature that year denied a supplemental funding request and demanded that the state child welfare agency roll back provider rates by 10 percent for three months. The agency had been discussing privatizing the entire child welfare system; the legislature told the agency to study managed care for six months and come back with a report.

Then the legislature decided to support a public child welfare managed care model, but not privatization. In 1997, the legislature block-granted child welfare funding and capped it (the legislation is referred to as “Capped Allocation”). That legislation also authorized three managed care pilots, and legislation in 1998 authorized three more pilots. The state’s objectives for the managed care pilots are to control costs and to assist children and families to achieve safety and permanency within ASFA timeframes. The requirements are: (1) there is a single entry point (i.e., the client was not referred out for services); (2) there are interagency agreements with community partners regarding funding of services; and (3) there are utilization review mechanisms (which often resulted in front-end approval mechanisms).

The managed care initiative is testing the principles of child welfare managed care; the counties keep savings that, if it were a private corporation, would go to shareholders, and they put the savings back into child welfare services. In addition, pilots are required to have performance agreements regarding outcomes, and some rules and regulations are waived. Managed care counties also have more flexibility for spending within the lines; they can negotiate provider rates, services, and outcomes except for some rates such as for Residential Treatment Centers. One result has been that counties with children in high-end placements moved relatively quickly toward getting those children adopted. The counties can opt to apply up to 5 percent of their savings toward their 20 percent portion of child welfare costs (there’s an 80/20 split with the state), but so far they all have opted to use their savings instead for creative ways to do child welfare.

Establishing the right allocation is difficult, and there is little incentive to become a managed care pilot site unless a county is sure it would have savings. For the first couple of years, managed care counties underspent their allocations, and in 2000 four out of six overspent their allocations. They were trying to find a balance, but the rule for blocked allocations is that since managed care counties can keep their savings, they’re last in line to receive any surplus child welfare funds. So those counties ended up covering their overspending with county-only funds. Eight of the 10 big counties overspent; both of the 2 who did not overspend were managed care counties.

Boulder County was one of the original three pilots and initiated its managed care initiative in July 1997. Several collaborative efforts were already in place in the county, and the pilot was seen as an opportunity to implement managed care activities by building on the existing collaborations, without using a for-profit managed care organization. The objective of the Boulder County Managed Care Pilot was to gain the flexibility to enhance the interagency partnership.

Boulder’s initiative involved developing a new organizational entity called Integrated Managed Partnership for Adolescent Community Treatment (IMPACT). The local community mental health center, the Mental Health Center of Boulder County, Inc., served as the fiscal agent, while the pilot’s board comprised representatives of all of the local child- and youth-serving agencies (Department of Youth Corrections or DYC, probation, mental health, Department of Social Services or DSS, public health, substance abuse services provider, and other community members). IMPACT’s primary function is to manage out-of-home placements for DYC, mental health, and social services; each of these agencies provides funding to IMPACT based on their historical costs for providing out-of-home placements. In addition, each participating organization contributes funding for IMPACT.

IMPACT is organizationally a part of the county mental health department. The county DSS has a contract with the mental health department for IMPACT services, and DSS supervises IMPACT staff. DSS is responsible for all administrative functions for the initiative, and usually DSS does intake and works with the family. Out-of-home case management is the responsibility of IMPACT, and service delivery usually is done by contract providers but some foster homes are provided by the county or through private foster care agencies.

The initiative’s focus is adolescents ages 12-18 in need of, or at-risk of needing, residential services. Also all DYC youth are included. The rationale for focusing on these youth is that they are the most expensive; also, the model is very interagency-focused, and probation is an important partner. Any of the partners (schools, mental health centers, probation, child welfare) can refer cases, and all youth who fit the criteria are referred. IMPACT serves about 500 youth per year.

After referral, an interagency meeting assesses needs and plans the treatment. Staff present at the interagency meeting can authorize services; they also decide whether the case will be in the child welfare or the juvenile justice system. The flexibility of the initiative allows services to be “out of the box” and also allows the county to avoid duplication. Ongoing utilization review identifies and addresses gaps in services.

The funding for services varies according to the type of provider. For example, the mental health center is a partner, and its services are paid for by one lump sum spread over 12 months. For that amount, they provide services to everyone referred. Residential treatment centers and day treatment providers are fee-for-service. The rates are determined by how the amount of funding appropriated by the state. The child welfare block allocation is flexible, and the county negotiates rates based on utilization patterns and how much money is available. The core services allocation is less flexible; the county negotiates with providers and sets the rates.

In the first couple of years after implementation, DSS had substantial savings because of managed care, and those savings were shared with the partners – they go into the IMPACT “pot” and are divided among the partners. In 2000 DSS experienced significant losses, due in large part to the State granting a 15 percent increase for Medicaid but only 1.5 percent to the counties, so the county allocation was not sufficient to fund residential treatment centers. Also, staff were given a larger raise than 1.5 percent. So DSS overspent, but the county was able to save money through DYC and “bail out” part of the financial liabilities. There is no risk sharing with the residential providers.

Connecticut: Continuum of Care

Connecticut implemented the Continuum of Care pilot project in 1999 to reduce the entry and length of stay of children in residential treatment facilities. The state contracts with two lead service agencies (LSAs), one each in the North Central and South Central regions, to develop a network of providers and coordinate community-based services for children assigned to the program. The program targets children between the ages of seven and fifteen who are referred to the state’s central placement team for residential treatment. Overall, the program can serve a maximum of 70 children.

Cases are referred from the two state regional offices to the central placement team for review. Based on a review of information pertaining to the child’s functioning, the placement team determines whether the child is eligible for the program. Aside from determining that the child’s acuity level is high enough to warrant treatment in a residential facility, the placement team does not recommend specific service plans. Eligible children are randomly assigned to residential treatment or one of the two LSAs.

LSAs are paid a case rate based on the average annual cost of in-state residential treatment. The case rate, $5000, is intended to cover all administrative, case management, and service costs for 15 months. LSAs receive payment in four installments -- at case assignment, 50 days, 180 days, and discharge. Payments may be withheld if a case review determines that the treatment plan, services being provided, or progress toward goals are unsatisfactory. Providers are also required to maintain invoices that are audited by the state’s independent evaluator. To discourage providers from under-serving children and families, LSAs are required to use 90 percent of the case rate, and any unexpended funds are returned to the state.

Within two weeks of receiving a referral, the LSA meets with the family and other service providers to assess the child and develop a treatment plan. LSAs may use the case rate flexibly to deliver, or purchase from other providers, a range of services needed to maintain children in community settings. Ideally, children will be returned to their communities prior to the end of the 15 months so that after care can be provided in the last several months of the treatment period. The decision to close a case is made jointly between the state and the LSA.

An independent consultant is evaluating the pilot project. The evaluators are monitoring both program processes and outcomes including child and family well-being, child functioning, education, safety, health, caseload size, number of client contacts, and costs.

Florida: Coalition for Children and Families

In 1996, the Florida legislature adopted the privatization strategy for the state child welfare system by calling for pilot “Community Based Care (CBC)” projects that would implement the lead agency model on a county-by-county basis. In 1999, the legislature expanded the CBC requirement statewide; the original target date for privatizing all 67 counties was Jan. 2003, but the public agency has requested a two-year extension on that target date and it is likely to be granted. Across the state, 14 counties are in some stage of privatization, representing about 50 percent of the active child welfare cases in the state. Initially Florida planned to operate the initiative under a title IV-E waiver, which was granted by the federal government, but the state decided not to use the waiver, due at least in part to the stringent research requirements.

Under CBC, state child welfare workers still take initial calls of alleged maltreatment and conduct investigations, but one private agency is selected for each county to take over the primary responsibility for foster care, adoption, and family support. The agency can either provide services directly or subcontract with service providers.

Unlike any other child welfare privatization initiative in the United States, Florida’s CBC initiative is funded through “global budget transfers” (like block grants) to lead agencies. Other privatization initiatives use capitated or case rates. Florida’s initiative thus potentially transfers a large amount of risk (based on intensity, duration, and volume) to contractors. A statewide risk pool can be tapped in cases of excess referrals or catastrophic service costs. The initiative is supported through IV-B, IV-E, Medicaid, TANF, social services block grant, CAPTA, adoption incentive grants, and state funds.

The first pilot project implemented was the Coalition for Children and Families in Sarasota County, which began serving clients in January 1997. The lead agency is the Sarasota YMCA. All children in the county who are determined by state child welfare workers to be abused or neglected and in need of services, whether in-home or out-of-home, are referred to the Coalition, which then provides for all case management and services. A multidisciplinary team develops the case plans and participates in monitoring and decisionmaking. The Coalition has an active caseload of 1,600-1,700 children. Lead agencies are allowed, under state law, to receive “excess earnings” of federal reimbursements as bonuses, and the Coalition has been successful in doing that. The money has been used for maintenance adoption subsidies and out-of-home care for IV-E-eligible children.

Georgia: Metropolitan Atlanta Alliance for Children (MAAC)

Children in the Atlanta region who are not eligible for Georgia’s program for Severely Emotionally Disturbed (SED) children are referred to Metropolitan Atlanta Alliance for Children (MAAC). Established in 1998, MAAC is an MCO that contracts with a network of providers to deliver a range of services. In addition to residential treatment and foster care, providers in the MAAC network also deliver reunification and adoption services. The range of levels of care provided by the network members enables MAAC to move children out of high levels of care more quickly than would be possible if only the agency providing the intensive care were managing the case. Over the past year, MAAC served approximately 30 children and its capacity was recently expanded to serve 40 children.

MAAC is responsible for placing referred children with a network provider according to the level of care that is needed, and remains responsible for providing care until the case is closed. Referrals to MAAC are reviewed by a six-member professional committee made up of representatives from each of the six provider agencies. This committee determines the level of care that the child needs and assigns the case to a network provider. The placement decision is made by consensus among the six-member committee. For children in state custody, the service agency caseworker, the state child welfare worker, and any other professional assigned to the case attend all case staffings and court hearings.

Rather than increasing the array of services that any one provider delivers, the effect of MAAC has been to make available a wider array of services to any child served by the network. Through the provider network, children can more easily move between levels of care and service. Furthermore, because providers can rely on each other for support services, placement disruption is less likely. For instance, the network’s intensive in-patient hospital provides short-term stabilization for network members’ clients. Otherwise, the hospital only provides long-term care in chronic cases. As a result of this collaborative effort among network members, individual providers are more willing to maintain care for a child for longer periods because the other providers are more available to them if short-term supportive services are needed to avoid placement disruption.

The state pays MAAC a flat per diem rate of $159.60 for each child referred to the network by Match. MAAC then pays the provider at the following per diem rates that are set by the state: basic care $70-80, intermediate care $133 to $230, and intensive care $313. Hence, the payment system imposes a financial incentive for MAAC to avoid placing children in higher-level, higher-cost care and to keep the length of stay in this type of care to a minimum.

MAAC’s administrative entity has taken on the financial risk of the network arrangement since member agencies also receive referrals from other sources that pay the same per diem rate. Hence, the providers can survive financially without MAAC, but the administrative component of MAAC can’t survive if it overspends its budget. As it is, the state per diem only covers services provided by network members, not MAAC’s administrative costs. MAAC conducts private fundraising efforts to pay its administrative costs.

Illinois: Foster Care Performance Contracting

The overall goal of this initiative was to redirect resources from maintaining children in the system to a focus on gaining permanency for children. Its objectives were to provide the resources necessary to improve permanency outcomes for children, hold the system more accountable, improve the quality of care, and create incentives to perform in the best interest of children and families. When the initiative was implemented, a top priority was addressing a large and growing permanency backlog. It was implemented for Cook County relative care in July 1997, and expanded in July 1998 to all traditional foster care statewide. The budget for FY99 was about $334 million, including $192 million for case management and $141 million for maintenance payments; this represents about one-fourth of Illinois' child welfare budget.

The initiative accomplishes the objectives by linking provider reimbursements to performance. Agencies are no longer guaranteed a caseload but are expected to manage their cases by balancing the cases flowing in with those flowing out; if the standards are not met, caseloads will increase but the level of payment remains steady. Agencies that move more than the contracted number of children (24 percent of their caseload) into permanent living arrangements do not experience a reduction in case management payments, and they may receive a bonus above the standard payment. Each placement that results in a "pay-out" will be replaced with another referral. However, agencies that do not achieve their contracted performance standards may lose up to 33 percent of their contract because they will not continue to receive referrals. The initiative includes new investments in reunification services, permanency capacity, subsidized guardianship, emergency care, recruitment, counseling, and case assignment.

The target population is all children in relative care and traditional foster care statewide. Approximately 23,000 children (over 70 percent of Illinois' substitute care population) were expected to be served in 2000 through the performance contracts. The initiative excludes children in specialized foster care, independent living programs, and residential placements. Services provided include case management, family preservation and support services, family foster care, kinship care, adoption, and respite care.

Monitoring teams review the agencies' performance and provide feedback for continuous quality improvement. Benchmarks measure the quality of care provided to children and families. The performance standards negotiated with providers are substantially higher than the previous average, and agencies that fail to achieve the standards set under the contract risk having their intake placed on hold. Throughout the year aggregate permanency performance data (managed by Chapin Hall) are shared with all providers and each agency is ranked from high to low; agencies can reconcile their numbers if they believe the numbers are not accurate.

Under performance contracting, payments to providers are made in two parts: maintenance payments, which are passed through to foster parents and relatives caring for children, and administrative payments, which provide for services to the child, the child's family, the foster family/relative caregiver, and administration costs. The monthly administrative payments provide for case management; a permanency worker, recruitment worker, and education liaison for each team, counseling and therapy; and emergency care. Agencies also receive lump-sum payments for reunification/aftercare, and traditional (non-relative) foster care agencies receive additional resources for the recruitment and training of foster parents and the provision of emergency foster care. Administrative payment rates are based on expected caseload ratios (22.5 cases per caseworker), with differential expectations of intake, permanency outcomes, and non-permanency outcomes depending on whether the agency is in or outside of Cook County and whether the case is relative or traditional foster care. The current effective monthly rate for traditional foster care administrative payments is $569 per child; for relative foster care, administrative payments range from $600 to $714 per child, depending on the agency's previous performance. Maintenance payments range from $361 to $445, depending on age of child. The difference between the administrative payment level and the actual caseload represents the fiscal incentive/disincentive of the initiative.

An important result of the initiative has been a dramatic reduction in caseload, from over 51,000 children in care to less than 30,000 in three years. In Cook County, caseworker relative care caseloads declined from 25 to 22 within existing spending levels. A byproduct of this trend was that there were not enough referrals to meet the contracted intake obligations of performance contracts, especially in downstate Illinois. The state responded by transferring cases from "lower performing" cases to the rest of the system, which is not viewed as a long-term solution to the systemic problem of "intake-dependency."

Kansas: Public-Private Partnership

Of all the state fiscal reform initiatives in child welfare, the Kansas Public-Private Partnership may be the most comprehensive in its scope. Over the course of fiscal year 1997, Kansas completely privatized all family preservation, adoption, and foster care services.

The state was divided into five regions and contracts for each of the services were given to private agencies. For family preservation services, monthly case rates were $2,200 for 12 months. Agencies were paid this rate regardless of the extent of the services they provided but they were expected to stabilize the family within 90 days. If a child was brought back into the system during the first year after case completion, the private agency had to service the case without any additional case payments. The state imposed the following outcome measures on the agencies: there must be no confirmed reports of neglect or abuse in at least 90 percent of the families; children must remain home in at least 80 percent of the cases; and finally, clients should be satisfied with the services.

The first adoption contract went into effect on October 1, 1996, with one agency (Lutheran Social Services) in charge of adoptions throughout the entire state. The provider is responsible for recruiting and training adoptive families, matching the adoptive family with an available child, and providing support services for the family for the first eighteen months after the adoption is completed. The agency was paid a case rate of $13,556 in the first contract. The main outcome was originally to place 70 percent of the children within 180 days. However, after the first year, this goal was revised downward to 55 percent. Other important outcomes are to finalize 90 percent of the placements by the end of the first year after placement, to ensure that 90 percent of the placements are intact 18 months after the adoption is finalized, to have at least 90 percent of children stay in only one placement from the time parental rights are terminated until the adoption is finalized, and to have a 90 percent family satisfaction rate.

The privatization of foster care services began in early 1997 with contracts to three agencies. These contractors are required to accept all referrals from the state and to provide post-reunification services for 12 months after the reunification of the child with its family. The case rates originally started at $12,860 for the lowest level of need and went up to $15,504. After the first year, the case rates were increased and a number of catastrophic risk pool slots were created for each provider to deal with extremely high-end cases. The payments are based on the expected achievement of case milestones. The contracts specify that 60 percent of the children will be returned to their families within six months. Other important outcomes are that 98 percent of the children will not be the subject of reports of abuse or neglect, that 90 percent will have no more than three foster care placements, and that 65 percent of sibling groups would be placed together.

Kansas has now changed from a case rate model to a capitated per/child, per/month system for foster care and adoption contracts. The monthly payment, which varies by region, begins with referral and ends when the child returns home, is adopted, ages out, or is referred to another agency. The shift from case rate to capitated rate was due to the financial shortages experienced by contractors under the case rate system; the state acknowledged that contractors did not have control over all factors that influenced permanency and should not bear all the risk.

Kentucky: Quality Care Initiative

Kentucky’s Quality Care Initiative started in early 2000 after an extended dialogue between the Mary Hearst agency and the state Department of Community Based Services. This pilot project is currently only available in Jefferson County. Under this initiative the agency is given a case rate and is expected to achieve designated permanency goals for the children. There is a stop-loss provision in order to limit potential losses to the provider. The provider covers any loses up to 125% of the case rate and the state covers any additional costs beyond that point.

The program is designed to serve three distinct target populations: teenage girls in need of residential placement; children in the foster care system who are transitioning home; and children who have just entered foster care in the last 45 days. At the time of the interviews with QCI staff, there were 5 youth from population 1, 21 youth from population 2, and 5 youth from population 3. There are no immediate plans to increase the number of youth involved in the program. The total number of youth served for the course of the program is 60.

This initiative has increased the provider’s involvement in the decision-making process. The provider determines the level of services that they offer to the family and they also attend court hearings to give reports on case progress. According to the provider administrator, the main advantage for this program is the additional freedom to provide wraparound services rather than the strictly prescribed services that are offered by other state programs.

Maryland: Baltimore Child Welfare Managed Care Project

The Baltimore Child Welfare Managed Care Project, Maryland’s title IV-E waiver demonstration, began serving clients in January 2000. It targets children age 0 to 5 in out-of-home care and their siblings and families. It is testing cost neutrality – i.e., can the contractor obtain better outcomes without costing more money.

The Baltimore child welfare agency refers cases that fit the criteria to the evaluator, who randomly assigns them to the initiative or to regular services. Upon referral, the contractor takes over all functions: case management, assessment, foster care payments, service plan, service delivery, wraparound care. The contractor is receiving $24 million to serve 500 children over three years. There is a stop-loss provision: if the cost for a child exceeds $3,500 per month, the state pays 90 percent of the costs over $3,500 and the contractor pays 10 percent.

The initiative’s evaluation will include a process evaluation, a cost-benefit study, and a cost-neutrality study. The contractor has agreed to improve benchmarks on length-of-stay, time to permanency, and re-entries. Early results indicated reduced length-of-stay and re-entry rate and increased adoption rate for children in the initiative.

Massachusetts: Family-Based Services Initiative

The goal of Massachusetts’s Family-Based Services Initiative is to implement a collaborative, community-based approach utilizing Department of Social Services (DSS) financial resources in coordination with other community, state, and private providers, systems, and funding sources in order to address family needs in a comprehensive and efficient manner. The impetus for the initiative came from the State child welfare staff wanting to be able to provide more flexible, collaborative, and responsive services to improve outcomes for children. The previous system of purchasing services through contracts prohibited the local customizing of services. The initiative allows the customizing of services based on community needs and resources; it enlarges the pool of service providers available; and it allows the State to purchase services on an as-needed basis. In addition, it provides a new emphasis on maximizing third party reimbursement and access to services not funded by DSS.

The target population for the initiative is primarily intact families who are at-risk of their children entering custody due to supported assessments of abuse and neglect. About 75 percent of the 38,000 children in the child welfare system are still at home, so that is the primary population targeted, although children in out-of-home care are not excluded.

The initiative, which began serving clients in January 2000, involves the State contracting with nonprofit community-based child welfare agencies to serve as lead agencies for DSS service areas. There are 29 service areas and 18 lead agency contracts – some area offices joined together to be covered by a single lead agency. The lead agencies receive a specified amount of money (initially $100,000 per year) to provide lead agency functions – gatekeeping, conducting utilization review, creating and coordinating provider networks, monitoring quality of services, and accessing third-party reimbursement. Each service area office now has a staff person from the lead agency on-site full-time. A separate contractor (Community for People) is tracking the initiative and developing and supporting a database for utilization management. There is no formal evaluation in place, but the State recently completed a six-month self-assessment report.

The lead agencies develop local networks comprising a broad array of formal and informal social service, educational, housing, and cultural resources serving families. These agencies recommend a core group of network providers from whom DSS purchases family-based services through contracts and at rates the providers negotiate with DSS. The providers are paid primarily on per-diem and hourly rates. The state is encouraging case rates, as the initiative utilizes models that vary in intensity according to the needs of the families, but case rates require time to develop necessary administrative structures.

The array of services required include family stabilization and reunification services, family support services, respite and short-term placement services, and miscellaneous resources (the last category is capped at $500 per family, and is flexible money to purchase either goods or services). Most of an area’s service budget is obligated in an “open order encumbrance” assigned to the lead agency, and payments to the network providers are made against and “draw down” on the open order encumbrance. Thus, the service budget is flexible and able to respond to changing client needs without the burden of administrative contract amendments. However, it is necessary to manage the service budget closely, pay attention to the services being provided, and be careful not to duplicate services. Although lead agencies do not receive incentive payments for exceeding goals or financial penalties for underperformance, one agency has already ceased being a lead agency for the initiative, at least partially because it went through the service budget too quickly. The initiative emphasizes the importance of informal family and neighborhood supports as primary components of every family’s treatment plan; the rough rule-of-thumb is that 75 percent of the treatment should come from family, faith, and friends, with 25 percent funded services.

The lead agencies convene and facilitate treatment planning teams, which consist of small groups of providers, other persons representing systems and/or community resources, and families. After a social worker brings a case to the attention of the area DSS office, the team (including the family) meets to develop and discuss a treatment plan. Two important changes the Initiative has brought about are (1) the involvement of the family right from the beginning as key players in planning the treatment, and (2) the involvement of the team in planning and coordinating services, rather than the social worker directly calling service providers.

Quality of services is monitored by lead agencies through weekly meetings with the core team and six-week review meetings with the families. Lead agencies meet quarterly with DSS to discuss quality issues. Providers are supposed to administer client satisfaction surveys, but it is not administered consistently.

It appears that the initiative is able to serve more families than were served under the previous system because the families now turn over more quickly. The initial funding for the lead agencies ($100,000 per service area, for a total of $2.9 million) was taken from the direct services budget rather than staffing budgets, which was not popular among providers but obviated the need to reduce public agencies’ staff sizes. (There were some staffing reductions, but this was handled by eliminating a few vacant positions so that nobody was laid off.) There is no formal evaluation being conducted, but the State developing a database, reviewing every three months and retooling as needed.

The initiative has broadened and enhanced the pool of providers and provides opportunities for newcomers. A major goal is to include minority agencies and community-based agencies, and the state is providing capacity-building for these agencies.

Michigan: Michigan Families

Michigan Families is a IV-E waiver demonstration project that was launched in 1999 and currently operates in 6 of the state’s 83 counties (St. Clair, Monroe, Livingston, Van Buren, Jackson, and Nuwego). The objectives of Michigan Families are to reduce foster care placements, shorten lengths of stay in foster care, and accelerate family reunification. Providers have adopted a wraparound model to deliver services in the home and the community.

When first implemented, the program targeted children who were in foster care or at risk of placement. However, the majority of children served are not in foster care and receive placement prevention services. Because providers have tended to serve lower-risk children who do not meet the criteria established for federal match dollars, recently the state lost a half million dollars in federal reimbursement. The program now has been modified to target only children in foster care.

In five of the six sites, Michigan’s Family Independence Agency (FIA) has contracts with community mental health centers that act as lead agencies, but in one county the FIA is the lead agency. Lead agencies deliver most services directly, but may subcontract with other providers for services that they don’t provide. The FIA retains responsibility for administrative functions — contracts, billing, reimbursement, and training — and oversees the provision of services. Case management is a collaborative effort between the FIA worker and the community and family team at the lead agency. The community team reviews referrals and develops service plans in collaboration with the FIA, the school system, and the family’s informal supports. FIA workers have discretion over placement decisions and maintain the role of making recommendations to the court regarding the case.

Lead agencies are reimbursed $1586 dollars per month per child to provide all services that are needed. This payment rate is based on the state’s average payment for all foster care, including residential and specialized, in 1997. Lead agencies are allowed to deposit any unexpended funds into an agency risk pool. The lead agency can draw from the risk pool to provide higher levels of care that cost more than the monthly reimbursement. Alternatively, unexpended funds may be used to expand the array of services that are provided by the agency. However, 50 percent of unexpended funds that are above 10 percent of the budget is returned to the state. Likewise, costs exceeding 10 percent of the lead agency budget are also shared 50/50.

Flexible funding enables providers to deliver a wider array of services to families in community settings including the provision of in-home services to prevent foster care placement or facilitate reunification. The state had expected agencies to use flexible dollars to provide nontraditional services such as parent mentoring, neighborhood-based services, school-based services, and after-school programs to address families’ unique needs. But generally, lead agencies have delivered more traditional services like parenting classes and substance abuse treatment.

FIA staff review cases that are assigned to the lead agencies quarterly to ensure that appropriate services are provided and that providers are operating in accordance with the wraparound model. No performance standards are currently in place, but the state plans to include standards such as family functioning, length of stay, and recidivism when contracts are renewed. The initiative is being evaluated by a third-party contractor.

Michigan: Permanency Focused Reimbursement System

Michigan’s Permanency Focused Reimbursement System was implemented in 1997 to reduce children’s lengths of stay in foster care and to increase the number of children placed in permanent homes with parents, relatives, legal guardians or adopted within specified timeframes. The primary mechanism for inducing providers to achieve permanency goals within specified time frames is an incentive payment system.

Under this system, at initial case referral, providers receive a lump sum payment of $2,150 and an ongoing per diem administrative rate of $13.20. An additional payment of $1,850 is made to the provider if the child is returned home, placed with a relative or legal guardian, or moved into supervised or independent living arrangements within 290 days; or if parental rights are terminated within 515 days. For placement cases, providers receive $1,250 if the child remains in the home for six consecutive months, and an additional $1,550 for placements that are stable for twelve consecutive months. Providers who place a child in an adoptive home within seven months of termination of parental rights receive a payment of $1,250.

These payment rates were based on average foster care costs. Hence, providers are expected to have financial losses on some cases and gains on others, but should break even overall. Providers may use payments flexibly to deliver and purchase whatever services are deemed necessary to hasten permanency.

Currently, under this pilot project, the state contracts with six providers that together serve 45 percent of foster care population in Wayne County (approximately 8,900 children). Providers receive cases on a rotational basis and may not refuse referrals. Also, once a case is assigned, the provider remains responsible for delivering services for 365 days following placement in a permanent home without additional incentive payments. However, if children do re-enter foster care prior to the 365-day period, the provider continues to receive the per diem administrative rate. The incentive payment cycle is restarted for children who reenter foster care after remaining in a placement for 365 days.

At present, state monitoring of the pilot system focuses on whether providers achieve permanency goals within specified timeframes. Contracts with the providers stipulate that the program goals will be achieved in 80 percent of cases. Based on a recent analysis of incomplete data, providers had met permanency goals in about 40 percent of their cases. According to the program administrators, barriers such as a scarcity of foster homes and high staff attrition in Wayne County have prevented higher success rates. In addition to monitoring permanency goals, the Purchase of Care Division conducts site visits to review cases and evaluate the quality and timeliness of services.

In terms of utilization management, one provider employs an access manager. The manager tracks how many children are receiving various services, the length of time children receive services, and the number of slots that are open. Also, the manager is responsible for approving services that the caseworkers provide.

Michigan has contracted with the University of Michigan, Ann Arbor, MI, to evaluate the pilot project.

Minnesota: PACT 4 Families Collaborative

Beginning in the mid-1990s, Minnesota (where child welfare is county-administered) began funding two types of collaboratives: (1) the Children’s Mental Health Collaboratives, which are coordinated by the state Department of Human Services, and (2) the Family Service Collaboratives (covering children in the child welfare system), which are coordinated by the state Department of Children, Families, and Learning (the equivalent of an education department). Both types of collaboratives bring partners together to do coordination, planning, and purchasing of services, and allow a more holistic and locally-designed approach. Also, the collaborative structure allows access to federal funding for IV-E and Medicaid. Local public partners include school districts, public health, corrections, and human services, and private partners include organizations such as Boy Scouts, Girl Scouts, 4-H, and YMCAs, as well as substance abuse treatment and mental health agencies.

The Putting All Communities Together (PACT) 4 Families Collaborative is one of the earliest and possibly most successful examples. It is a four-county, multi-agency partnership that operates both types of collaboratives in west-central Minnesota. The four counties cover 3,700 square miles and have a total population of 93,000 people. PACT 4 emphasizes prevention and early intervention, targets the needs of children ages 0 to 21 with severe emotional problems, and provides support to families so that they can keep their children. In addition to state funding, it receives local funding: the four counties each fund it at $1 per capita, schools contribute $1 per student, and organizations contribute $1,500-3,000, depending on their budget. PACT 4 has also received some foundation funding. The program helps to create awareness of all the services available and encourage services other than probation and corrections for the targeted children. Overall, the early intervention has helped prevent out-of-home placements, infused additional resources into child welfare, and put children’s mental health needs in the forefront of the child welfare systems.

Missouri: Interdepartmental Initiative for Children with Severe Needs

The State of Missouri has formed the Missouri Interdepartmental Initiative for Children with Severe Needs, which is a consortium of child-serving divisions from the Department of Social Services, Department of Mental Health, Department of Health, and Department of Elementary and Secondary Education. The Initiative integrates funding from all the participating Departments to support comprehensive, integrated Plans of Care for children with severe behavioral health needs and their families, implemented under a single, unified care management process.

The Initiative operates in two regions of the State, an eastern region (an urban area comprising St. Louis and surrounding suburbs) and a central region (a more rural area in the center of the State). It targets children and families with disruption likely to result in long-term residential care. Eligible children for enrollment are between the ages of 4 and 18 who (1) reside in those regions, (2) are currently in or at serious risk of long-term residential placement, and (3) have serious behavioral health needs as measured by a standardized instrument (Childhood Severity of Psychiatric Illness, or CSPI). This population of children is likely to need services funded by multiple State agencies, and the State realized that better coordination of services was needed to reduce barriers, enhance effectiveness and efficiency, and prevent children from “falling through the cracks.” In addition, the State hoped to enable many of the children to stay in their homes or return home or to their home communities by providing community-based wraparound support services, thus avoiding unnecessarily long and expensive residential placements.

The State contracts with two agencies to deliver services and provide operational support. The Care Management Organization (CMO), currently the Missouri Alliance for Children and Families, has responsibility for developing a network of resources, assigning care managers, purchasing services, and monitoring the quality and utilization of services. The Technical Support Organization (TSO), currently ValueOption, supports the initiative by managing information, finances, quality improvement, and communication. An interdepartmental management team manages the planning, procurement and contracting, start-up, and implementation of the Initiative and oversees system operation.

On March 1, 1999 the first children were enrolled. Currently there are about 250 children being served at any given time. The Initiative was originally designed to enroll 1,000 children in multiple CMO’s, but there is only one CMO and the maximum capacity of that CMO is 250. Child welfare workers, juvenile corrections workers, and mental health workers refer children to the Initiative through interagency teams (IT’s), which consists of local representatives from each participating State agency. About 60 percent of the referrals are from child welfare, 20 percent from juvenile corrections, and 20 percent from mental health. The IT enrolls and assigns children and families, reaching all decisions by consensus. Each eligible child must meet the criteria specified previously; there is a cut-off score for the CSPI, then all children who meet the criteria are enrolled as long as slots are available. Teams are notified each month how many openings are available.

The date that the IT enrolls a child is the day that State services cease and the CMO becomes responsible for all care and costs. The State’s practice is to give the CMO about two weeks lead time to review the referral and become acquainted with the child and family before becoming responsible for services. The minimum services required for each child include care management, team meetings, and development of a Plan of Care. The contract specifies that the CMO may not refuse any referral or disenroll any child from the Initiative until all plan of care objectives are met and the IT approves disenrollment (a “no reject, no eject” policy). The CMO is required to deliver services required by the treatment plan and is responsible for all costs except physical health, which is covered by Medicaid. Children are enrolled for six months; if necessary, an additional three-month period may be authorized.

The CMO is required to assess the child and family within 7 days of enrollment, and have in place an individualized, community-based Plan of Care within 14 days of enrollment. A family support team, consisting of the care manager, the family, the child, providers, neighbors, school personnel, the referring agency, juvenile court) develops the Plan of Care, monitors progress, and recommends disenrollment. The list of services that the CMO is required to make available is extensive, and includes inpatient psychiatric evaluation and treatment, residential treatment, residential sexual abuse and offender treatment, alcohol and drug abuse treatment (inpatient and outpatient), foster care, respite care, adoption services, educational services, outpatient psychiatric services, psychological consultation services, medication management/monitoring, individual/group/family therapy, crisis intervention stabilization, assessment, case management, intensive in-home services, day treatment, family support group services, wraparound services, crisis intervention access, residential support services, community integration support services, transitional living services, school-based behavioral support services, transportation, recreation, parent aide, supported work services, and mentor services.

The CMO, the Missouri Alliance for Children and Families, is a limited liability corporation (private for-profit) set up specifically for the contract under the Initiative. Its primary mission is to provide wraparound services under case rates to move children from restrictive placements to their home or other places within the community. It consists of nine member agencies – eight residential treatment providers and one psychological counseling agency – and was established in response to the growing movement of child welfare into managed care.

The CMO is paid a case rate that begins when a child is enrolled. Currently the case rate is $3,329 per month, with $226 per child for case management. The case rate was originally set up based on the 1,000 most expensive children in the two regions, and the Initiative was designed to include at least four CMO’s, each of which would receive a cross-section of the 1,000 children. As it turned out, only one CMO bid on the contract (probably due at least in part to companies concerned that the cost would exceed the case rate), and the most expensive (rather than a cross-section) of the children were referred to the initiative. Accordingly, the case rate was low. After the CMO lost money the first year, the state covered part of the loss. The CMO contract specifies 3 percent risk sharing – 3 percent of the budget was set aside to handle high-cost children – but there was no procedure established to access that risk pool.

The case rate is financed by contributions from divisions within the Department of Mental Health (DMH) and the Department of Social Services (DSS). The percentage at which each division participates was calculated from the actual expenditures identified in the historical costs for the target population. The contributing percentages are as follows:

• DMH/Alcohol & Drug Abuse – 0.57%
• DMH/Mental Retardation & Developmental Disabilities – 1.72%
• DMH/Comprehensive Psychiatric Services – 12%
• DSS/Youth Services – 7.22%
• DSS/Medical Services – 17.51%
• DSS/Family Services – 60.99%

The TSO, ValuOptions, documents performance and collects data pertaining to quality and outcomes. The CMO is required to report data on outcome, but that process is still being refined. The CMO monitors quality of services – the care managers monitor children’s progress and CMO staff conduct site visits to the residential facilities. The CMO reviews service utilization weekly, especially for the most expensive children, to make sure the teams understand that the goal is to move the children into the communities. Researchers at Washington University are evaluating the initiative, with a comparison group of children, but the evaluation is just beginning.

The goal of the initiative was to design a more efficient and effective way to spend the dollars, and provide flexibility to enhance outcomes. The creation of interagency teams resulted in broader cooperation generally between the member public agencies, and they now enjoy enhanced relationships across the board, not just with initiative cases. When the system moved from fee-for-service to case rates, a different composition of services was needed – especially more workers to work directly with families. And the initiative has not had a drastic impact on the court’s role, but it does divert some children who came into the child welfare system only because placement with the State could get them services, not because there were child abuse and neglect issues. Now the children do not have to be in the custody of one of the agencies to receive services.

New York City: STAR (Safe and Timely Adoptions and Reunifications) Program

This initiative, implemented in April 2000 by the city’s Administration for Children’s Services (ACS), provides flexible dollars that agencies can allocate to a broad array of services to achieve timely permanency for children in placement. Agencies can obtain flexible dollars if they are able to show improvements in the length of stay for children in their care. Improvement is defined as an increased discharge rate into permanent homes without a corresponding increase in re-entries and transfers to other agencies. As an agency demonstrates improvement compared with its own past performance, ACS returns to it all or a portion of the savings generated by reduced care day utilization. ACS also monitors re-entries and reduces the fiscal awards for agencies that exceed their own past re-entry rates. ACS must approve the agency's spending plan so the saved funds can be reinvested in enhanced foster care or preventive and aftercare services. The primary target population is children who are waiting to be adopted or reunified with their families -- a large percentage of the children in foster care.

ACS provides per diem costs for all foster care services delivered to children in the care of ACS contracted agencies. Under STAR, ACS assesses an agency's history and past performance on outcomes such as time to reunification, time to adoption, and likelihood of re-entry into foster care after discharge. ACS then projects the size and characteristics (in terms of special populations) of the agency's total caseload over the next five years. For each agency, ACS projects a set of discharge rates based on six case types that reflect the agency's past experience in moving children out of foster care and into permanent homes. The historical baseline was calculated on children in the system during 1993-1998, following them as far forward as possible to capture re-entries.

An agency's acceptable range of performance is calculated as the number of re-entries or transfers in the agency's historical average rate plus and minus 15 percent of that average. If an agency demonstrates a reduction in care days, there will be corresponding per diem savings that will be available to the agency for reinvestment, although if the agency exceeds the allowable number of re-entries or transfers, ACS deducts funds from the agency's reinvestment award. ACS then calculates how much money could be saved assuming a 10 percent improvement in discharge rates and days in care. Across all STAR providers, the potential savings in 2000 was $9 million; the actual savings was $8 million, with a reduction of 185,000 care days based on tracking 23,000 children. The potential savings for 2001 is $17 million. There is no penalty for agencies that do not reduce average length of stay, although they do not receive any savings.

Currently 40 out of 44 ACS providers participate in STAR. Initially 41 opted to participate, but one agency has closed. Three providers are too new to participate (STAR requires historical data), and one opted out on ethical grounds. Providers can apply to receive start-up funds based on their projected savings (each agency is eligible to apply for up to 50 percent of their projected savings). Last year 16 of the agencies applied and received a total of $1.1 million for starting up new services; the funds were used primarily for hiring staff such as housing specialists, parent mentors/advocates, and adoption/discharge expeditors. The downside is that if an agency receives start-up funds and does not achieve the expected savings, the agency has to pay back the funds.

Based on last year's performance data, preliminary results indicate that 5 agencies will probably have to repay their start-up funds.

During the first year of the initiative, ACS produced three performance reports, but it was only at the end of the year that a report was produced that presented final performance results. This year, ACS plans to produce quarterly reports. They are still working out what kind of data to give back to the agencies, and how often.

STAR is funded by federal, state, and city dollars. Reinvestment savings that are spent for IV-E-allowable purposes can be claimed; otherwise, only the state and city portion can be reinvested. ACS expects to spend the same amount of dollars under STAR that they would have otherwise, but agencies will provide more services and additional in-home aftercare and preventive services.

ACS has developed a management information system, STARDAT, that was implemented in June 2001. The system is accessed through a secure Internet site that stores data back to 1993 on each STAR provider. Providers can query the system to extract data (both STAR data and other types of data) for management and planning purposes.

There has not been a noticeable or differential impact on minority providers. The ACS provider contracts run for 9 years, so the list of providers has not changed. And agencies who did not do particularly well on the performance reports were not minority providers; in fact, several minority providers had large savings returns -- one received over $500,000 and one received $344,000.

In terms of STAR's impact on ACS, a department was created, the contract management unit, to implement and monitor STAR and other programs. That department has 9 staff. ACS anticipates that more changes to ACS will be coming, based on the reinvestment savings budgets due on 9/10, in which providers present their plans for using the funds. One lesson learned in the past year was that ACS will need to monitor the spending of the start-up funds; in the past year, providers did not always implement the changes they had received funds for.

STAR's impact on providers has been: (1) the agencies are thinking more creatively, knowing that there is a pool of money they can tap; (2) STAR has gotten across the message about the link between length of stay and funding, and that ACS is interested in reducing care days; and (3) the agencies are more attentive to using data to inform practice, since ACS is giving them a lot of information about their cases and asking them to look at how to improve outcomes.

STAR is not being formally evaluated. Outcomes are being monitored as described above, and ACS sponsors conferences about innovative practices.

Ohio: ProtectOhio (Franklin County)

Ohio’s title IV-E waiver demonstration was implemented in October 1997, and Franklin county’s initiative (part of the waiver) was implemented in July 1999. This initiative combines performance contracting and managed care contracts. Its goal is to reduce length of stay and increase flexibility of services.

The target population is all children and families with reports. When a case comes in, it is randomly assigned (a requirement of the waiver) to the public agency or to one of two managed care companies. The public agency is using performance bonuses and incentives linked to success. The two managed care organizations include the Ohio Youth Advocate Program (a full-service provider) and the Permanent Family Solutions Network (a collaborative of three large services providers). Currently the managed care organizations are serving about 17.5 percent of cases (1200 children out of 6300).

The contractors have a "no reject/eject" requirement for all children randomly assigned to them. They are required to provide intake, assessment, and case management for all cases, plus whatever services indicated by the treatment plan. Services that the contractors must make available include crisis intervention, homemaker, home health, parenting skills education, home-based services, transportation, outpatient and inpatient mental health, partial hospitalization for children, treatment for children who commit sexual abuse, substance abuse assessment and treatment, programming for children with developmental disabilities, assistance in accessing services, access to medical services for children, emergency aid for household expenses, protective day care, crisis placements, foster care services at all levels, residential treatment for children, intermediate residential programs for children, reunification services, independent living arrangements for adolescents, and linkage to community services when families are reunified.

The public agency is using a staff "bonus" system as incentive to increase permanency in cases handled by public agency staff. They receive yearly raises based on success at achieving outcome goals including visits and recidivism. An interagency team (the Intersystem Program) is used to develop wraparound services for children in the care of the public agency.

The primary payment to the contractor is made on a Continuum of Services (COS) basis (i.e., case rate), which is $23,074. This covers all services provided to the child and family. Partial payments are made at three points: 50 percent is paid at referral, 40 percent is paid at three months from the referral, and 10 percent is paid at case closing. When the costs of a case exceed four times the COS payment amount, with prior written approval the County begins paying 50 percent of the direct service costs excluding the costs of case management.

The contractors utilize case management, outcomes monitoring, and service coordination. Contractors are responsible for all services needed for up to six months after the case is closed. After that, contractors can be paid in "fractional" payments after case closure. Risk corridors were established as follows: In the first year, the contractors are responsible for the first 5 percent of costs that exceed revenues and may retain the first 5 percent of "excess" revenue. In the second year, the risk corridor rises to 10 percent. In the third and subsequent years, the corridor is 15 percent. The next 10 percent of excess costs or revenues each year will be shared by the contractors and the County equally up to 5 percent each. The County is responsible for excess of costs beyond the corridors.

The County's quality assurance department monitors the contractors, and service provision is monitored against indicators of increased risk to children and expectations for client contact and service documentation. Outcomes are monitored, and each outcome indicator has a goal threshold. Client data are entered into the MIS daily. Franklin County is part of the IV-E waiver, which is being evaluated by independent evaluators.

Oklahoma: Oklahoma Children’s Services

The Oklahoma Children’s Services initiative was begun in a limited area in 1989 and expanded statewide in 1992. The “backbone” of the program is Comprehensive Home-Based Services, with the objectives of bringing about reunification quickly for families with children in care; preventing family breakdown for intact families with children at risk of placement; and preventing disruption for youth in placement who are at risk of disruption. It focuses on preventive, supportive, and wraparound services, and has put resources into developing local capacity to provide services and enhancing community ownership of child welfare.

The initiative has a caseload of about 2,000 families and includes 12 contracts with 8 contractors, based on geographic area. The contractors do all case management, assessments, case planning, referrals, and service delivery. Funding is allotted to each district based on the number of children in care in that district, and limited by the amount the legislature appropriates. Annual contracts specify the funding amount and the number of children and families the funding is supposed to cover, and each contractor serves everyone referred. The state closely monitors contractor spending and services delivered, and can take funding away if it not used on services to families.

Pennsylvania: Berkserve

In Pennsylvania (with a State-supervised, county-administered system), Berks County piloted a fiscal reform initiative that operated from July 1997-February 2000. This initiative, Berkserve, was designed to implement managed care principles in the delivery of child welfare services. It was designed and implemented by the Laurel Group, a statewide group of about 15 child welfare agencies that decided to pilot their own managed care initiative to avoid having managed care requirements imposed by the State without provider input. The providers also wished to develop an efficient public-private partnership model that relied on a network of local agencies to provide services.

The design phase lasted about two years, with numerous meetings involving county commissioners and the provider agencies' boards, as well as a study completed by a consultant. The Laurel Group canvassed for counties to participate in the pilot and Berks County was selected. Concern, Inc., which was the largest provider agency in Berks County and one of the four Berks County providers involved in developing the Berkserve concept, became the lead agency.

Concern, Inc. is a 501(c)3 agency that provides foster care services through 10 offices located throughout eastern Pennsylvania and Maryland. It serves about 525 children per day, providing treatment foster care primarily for children with severe needs who would otherwise be institutionalized. Although most of the agency staff were not employed by the Berkserve initiative, as the lead agency a few Concern staff members spent a large amount of time and effort trying to make the initiative work. The agency ended up losing money in the process, about $1500 a month, which was an indication of the agency's level of commitment to Berkserve. The Laurel Group also committed resources to keep the initiative going and helped "navigate the politics.”

Referral into the Berkserve initiative involved utilizing an extensive set of protocols developed by the four local private agencies and the county child welfare agency. According to the program administrator, a great deal of effort and thought went into the protocols to decide "what's reasonable and what's possible" in developing service plans and delivering services. The protocols specified for each step in the process how long the task should take and who was responsible for completing the task.

The protocols in essence formed a decision tree for county intake workers to follow, which involved going through selection/admission criteria step by step. If an intake worker worked through the protocol and decided that a case was appropriate for Berkserve, the worker called the lead agency (Concern) and referred the child into the initiative. The protocol also indicated which services the child needed to receive. Then the lead agency made the arrangements for services to be delivered, either through the four private agencies or through the network of 12 local providers that Concern had already established. The lead agency provided administrative, case management, and utilization review functions. All services except inpatient hospitalization could be provided through the network; inpatient hospitalization had to be handled in the traditional way.

The service providers were paid fee-for-service. The lead agency was responsible for monitoring services and verifying that they had occurred. The providers invoiced Concern, who paid the providers and invoiced the State for the costs. The invoices included 3-4 percent of billable services as administrative costs, which was a pool of money set aside by the county for administrative services. Concern and Laurel each got a portion of that 3-4 percent.

The plan was to reimburse providers on a fee-for-service basis for their actual costs during the first year. In the second year, a risk corridor would be established -- providers would not make or lose more than 5 or 10 percent. Then in the third year, case rates would be implemented and there would be full risk on the part of the providers. However, the plan never progressed beyond the first step.

Another plan was to set performance goals and quality standards as they went along. In the first phase, they were analyzing the costs of providing services through the initiative compared with providing services the traditional way, as well as monitoring timeliness and safety. There was a mechanism designed to assess client feedback and satisfaction, but it was not utilized. The initiative was not evaluated.

Initially, the children eligible to be served through Berkserve were families that had at least one child in care and had been involved with the county child welfare agency less than six months (they wanted to test the initiative with families that had not been "contaminated" with system-wide experience). However, there were such a limited number of families being enrolled under those criteria that they opened the initiative up to any case. Throughout the 2-1/2 years of Berkserve's operation, only about 24 children were served through the initiative, with a group of comparison children served in the traditional way.

Throughout Berkserve's period of operation, there was a great deal of frustration with and resistance to the initiative within both the county child welfare agency and the providers. The decision tree process was complex and often difficult for provider staff to follow, and many county workers were resistant both because they had not been involved in the design of the protocols and they were afraid privatization would end their jobs. Turnover was high on both sides.

Although Berkserve eventually ceased operations, it did move the county agency along toward a collaboration with private provider, and the providers got a better understanding of the public agency's work. Also, the initiative involved establishing a computer network of providers, and that computer network has evolved into "E-Home," an electronic referral system that will provide a quicker and more efficient method of locating foster homes for children. Providers will post available resources, and the county agency will reserve placements on-line. Within two years, 95 percent of referrals will be made through E-Home.

Tennessee: Continuum of Care

Like many other states in the late eighties, Tennessee experienced a surge in the number of children entering care. This large influx of children eventually overburdened the foster care system and the state budget for children’s services. In response to the state legislature’s concerns about the quality of care that children were receiving and the escalating costs of providing services for the growing number of children entering care, a series of reform efforts were initiated to both reduce costs and improve care. Of those reform efforts, Tennessee’s Continuum of Care (CoC) has been one of the most successful. Currently, the state has 40 CoC contracts with private providers who serve 4,400 (38 percent) of the 11,500 children in the state’s custody.

CoC targets children who have serious emotional and behavioral problems or problems that are more moderate but disruptive to family functioning and school life. Prior to the implementation of CoC in 1995, children with these problems were placed in residential treatment and often remained in those facilities until they were discharged from state custody. The intent of CoC is to divert children from placement in residential facilities or to step them down to non-residential settings with services as quickly as possible. In order to accomplish this goal, under the new CoC system, private agencies that were previously reimbursed only for residential care can use funds flexibly to provide a variety of services in a range of treatment settings including residential facilities, therapeutic foster homes, regular foster care, and the child’s home of origin.

CoC provider’s reimbursement rate is no longer determined by where services are delivered, but rather the level of care that the child needs. Thus, private agencies with CoC contracts are reimbursed a per diem rate based on the level of care that the state determines is needed when a child first enters care irrespective of the placement setting. CoC uses a staggered reimbursement structure in which a provider delivering services to children with greater needs are reimbursed at a higher rate. Once a reimbursement rate is established for a particular child, CoC providers continue to receive that rate regardless of the setting where services are delivered. Hence, although the CoC maintains a per diem rate system rather than shifting to a case rate as is common to other managed care arrangements, flexible use of the per diem rate provides a financial incentive to deliver services in the least costly setting. Providers receive the established per diem reimbursement until the child is either discharged from care or the annual maximum reimbursement that is stipulated in the contract is reached.

CoC contracts include provisions that are intended to ensure provider accountability. To prevent providers from avoiding more difficult and costly cases and to encourage the timely discharge of children from the most expensive care settings, providers are limited to a maximum number of rejected referrals and are required to accept a minimum number of admissions each month. Furthermore, to discourage providers from prematurely discharging children to lower levels of care, CoC contracts stipulate that 80 percent of children be successfully maintained in their own home or other family home for nine months after discharge.

In addition to using funds more flexibly, CoC providers also have greater responsibility for case decisions. Although the public agency retains responsibility for protective investigations, initial removals, and recommendations to the court for adjudication into foster care, once a child is referred to CoC, the private provider has primary responsibility for planning and delivering all services. The role of the public agency caseworker is limited to the approval of case plans, goal changes, reunification, case closure, preparation of court documents, and oversight of provider performance.

The state has put in place several mechanisms to monitor CoC providers. All providers are licensed by the state and their licenses are subject to annual review. State caseworkers monitor CoC contracts and conduct quarterly physical plant evaluations. To monitor provider performance, the state requires providers to submit monthly statistical reports. These reports include the number of children entering and exiting the provider’s care, the number of children returned to their own home or placed in another family home, the number of children who remain in a family home for nine months, as well as the number of placements that disrupt. The state also conducts an annual financial and program audit of 30 percent of contracted providers.

Texas: Project PACE

Project PACE (Permanency Achieved through Coordinated Efforts) was initiated by the Texas State Legislature, which wanted the Texas Department of Protective and Regulatory Services to examine ways to increase the competition in the procurement process for child welfare services. After a series of discussions with policy makers and service agencies, Project PACE was developed.

The initiative targeted children in Fort Worth and the surrounding ten counties (not including Dallas). The target population consisted of all children in the foster care system who required more than the most basic services and their siblings. This was determined using a six point level of care scale (1 meaning minimal care and 6 meaning extended in-patient psychiatric services). The average LOC for a child in the PACE program was approximately a 3.6.

Under the PACE program, the Lena Pope Homes, based in Fort Worth, was given a contract to serve children who met the program criteria. They handled all of the case decision-making process and were responsible for developing a provider network to meet all service provision needs. Originally, Lena Pope was expected to serve 217 children at any one time, at the peak of this program, in late 1999, they were serving almost 600 children. For the entire two and a half years of the project, they had a total number of 1,400 children. They were paid a flat case rate and bore the full risk for any case expenses that exceeded that rate. The rate began at $72 a day and eventually rose to $77 a day.

The advantage that the PACE project gave to the Lena Pope Homes was increased flexibility in devising their own service provider network. LPH pursued this part of the project vigorously, expending a good deal of resources training the organizations in their network. They hired an outside consulting firm, Praesidium Inc., to help staff members identify risk factors quickly. One of the problems facing the Texas child welfare system was the difficulty of finding foster and adoptive parents. Drawing on expertise within the Ft. Worth business community, LPH conducted a poll and had some focus groups to find out why people were not interested in foster parenting and adoption and how to overcome some of the barriers that prevented people from becoming involved. Through this process they developed an enlarged pool of foster and adoptive parents. This is considered by the state and by the provider as one of the enduring legacies of the PACE project.

Project PACE started operations on September 1, 1998 and closed down at the end of March, 2001. The project ceased operations because Lena Pope homes was losing too much money servicing the children in their care. The financing of private agencies in Texas is set up so that the state covers approximately 90% of the program costs and the private agency raises funds to make up the remaining 10%. Because the PACE program became so large, with expenditures of more than a millions dollars a year, the Lena Pope Homes was faced with the task of raising more than a $100,000 per year in addition to their other fundraising obligations.

This program is being evaluated by the University of Texas School of Social Work. As of this writing, the results have not been made public.

Washington: Title IV-E Waiver Demonstration Project

The Children’s Administration of the Washington State Department of Social and Health Services (CA) initiated this demonstration project in an attempt to provide more services to children in the state child welfare system at home (if possible) and within their communities. In order to achieve this goal, the CA made blended, flexible funds available to cover the cost of services, which included both basic and specialized services. The target population for this initiative were children in the child welfare system who were between the ages of eight and seventeen and in need of high-cost mental health and special education services.

The demonstration project was active in one region of the state (Spokane County). The CA had a contract with the Regional Support Network (RSN) to provide a wide range of residential, in-home, and follow-up services.

The RSN established a sub-contract with a local provider to manage the actual service provision. This provider then developed its own sub-contract with another agency to facilitate the Individualized and Tailored Care (ITC) teams. The ITC teams consisted of staff from the state, county, and providers where treatment plans were developed. These teams involved the foster parents and, where appropriate, family members in the discussions. The local provider had hoped that by sub-contracting out the facilitation of the ITC teams it could make case management more efficient and cost-effective. However, it was not able to achieve these efficiencies and its services sometimes overlapped with those of the sub-contractor.

This demonstration project was terminated after six months in November 2000. The local provider assumed the financial risk and during the course of the project it lost approximately $250,000. The RSN initially received $2,400 per child per month but this was quickly raised to $2,550 once it became clear that the sub-contractors were losing large sums of money. However, the local provider estimated that the services they were providing cost $4,800 per month. One problem that was highlighted in the final evaluation by the state was that the Children’s Administration believed that the RSN would free up some internal funds to complement the other flexible monies available for the children but the RSN denies that they ever agreed to do this. The evaluation notes that there was “insufficient communication” between state and county administrators and that this adversely affected the achievement of program goals.

Wisconsin: Bureau of Milwaukee Child Welfare

In 1998, the state of Wisconsin (previously entirely a county-administered child welfare system) took over child welfare in Milwaukee County. The purpose was to reform child welfare in the county to ensure the safety of children in care, achieve permanency as quickly as possible, and work cooperatively with the community to better serve children and families.

The county was divided into five service regions, with a contract in each region for a lead agency (called a “partner”). State workers do intake and investigative reports of child abuse and neglect, while contractors provide case management and develop and maintain networks of providers that provide services to children in care as well as families with children at-risk of placement. Contractors also track services used, authorize services, and arrange for payment for services.

Services are provided through the Safety Services Program to prevent placement (somewhat similar to family preservation services), and through Wraparound Milwaukee to support children in care. Overall about 6,000 children are in out-of-home care in Milwaukee County, and thus receiving case management from contractors, while about 500 families are receiving Safety Services and about 200 children in care are receiving wraparound services. Contractors receive a global payment (like a block grant) to provide case management for all children referred, and they receive a case rate of about $1200 per month for four months to provide Safety Services. Their funding is flexible – contractors are expected to use the dollars in ways that keep children safe. Wraparound Milwaukee services are paid for through mental health (Medicaid) and child welfare funding. Contractors can keep savings and re-invest them in services.

The Bureau of Milwaukee Child Welfare emphasizes community ownership and its accountability to the community. For example, when hiring new social workers, every interview has a community representative on the interview panel. It has focused on building capacity within neighborhoods in partnership with the community. The legislation establishing state responsibility also created a local partnership council that serves in an advisory role to the Bureau and ensure community input in service development, design, and delivery.

An evaluation is in process, but there is already a sense that the Safety Services component has reduced the rate of entries into care, that length-of-stay for new families has shortened, and that Milwaukee contributed to a drop in the state’s rate of child abuse and neglect.