National Study of Child Protective Services Systems and Reform Efforts: Literature Review. Demonstrations of New Approaches


The reviewed literature reveals that there is at least some aspect of CPS practice or policy undergoing change in almost all of the States. These changes are being undertaken at different levels. In some States, changes are occurring Statewide; in others, changes are limited to a specific locality; in still others changes are occurring at a number of demonstration sites. The literature clearly shows that change in the CPS system is being implemented throughout the Nation.

Some types of initiatives are found in several localities. These include:

  • Introduction of family assessment or differential response systems;
  • Creation or support of community-centered service delivery systems;
  • Clearer delineation of the relationship between CPS and law enforcement;
  • Increased collaboration between CPS agencies and alcohol and other drug (AOD) agencies; and,
  • Development of collaboration between CPS agencies and domestic violence (DV) agencies.

Exhibit 2 presents an overview of the current status of demonstrations and initiatives among the States. Notes explaining the criteria used to include a State in this Exhibit are found in appendix D. (This list may be incomplete due to lags in publication or other types of documentation. The Final Report of the national study of CPS systems and reform efforts will include an updated summary.)

Exhibit 2
Summary of New Approaches and Initiatives
  Family Assessments/
Differential Response
Community Centered Services CPS and Police CPS and AOD CPS and Domestic Violence
Alaska     X    
Arizona   X X    
Arkansas     X    
California   X   X X
Colorado   X      
Connecticut     X   X
Delaware     X X  
District of Columbia     X    
Florida X X X X X
Hawaii         X
Idaho   X X    
Illinois   X      
Iowa X X      
Kansas     X    
Kentucky   X      
Louisiana X        
Maine     X    
Maryland   X      
Massachusetts   X     *
Michigan     X X *
Minnesota X X X    
Mississippi     X    
Missouri X   X    
Nevada     X    
New Hampshire     X X  
New Jersey X     X  
New Mexico          
New York          
North Carolina          
North Dakota X   X    
Ohio     X X X
Oklahoma     X    
Oregon     X   *
Pennsylvania   X X    
Rhode Island   X X    
South Carolina          
South Dakota          
Tennessee     X    
Texas X   X    
Utah     X    
Vermont     X    
Virginia X        
Washington X     X  
West Virginia          
Wisconsin     X    
Total 10 13 27 8 8
Notes: See Appendix D for notes.

Introduction of Family Assessment/Differential Response Systems

One of the most widely discussed reform initiatives in the literature is the introduction of family assessment procedures in place of, or in addition to, other investigation procedures (AHA, 1996; Christian, 1997; English, Wingard, Marshall, Orme, & Orme, 2000).

The descriptions of the models being implemented indicate that there are many different approaches under this rubric. For example, Minnesota, Missouri, Virginia, and Washington are piloting models in which the family receives either an investigation or a family assessment. In Florida, localities are conducting assessments of all families with reported incidents of maltreatment. Additional investigation activities are being performed on only those cases involving criminal proceedings. In Iowa, family assessments and investigations are being conducted on all cases. In contrast, North Dakota is conducting only family assessments.

States differ in the implications of a family receiving a family assessment. In Missouri and Virginia, children in families on the family assessment track cannot be found to be a victims of child maltreatment, and parents cannot be found to be a perpetrators. Furthermore, information on the parent is not placed in a central registry. North Dakota no longer establishes findings of substantiation, but maintains a registry. All cases are entered into the registry if there has been a determination that services are required.

In addition to describing their differential response initiatives, some States have published evaluations of their initiatives. Evaluations in Florida, Iowa, Missouri, and Virginia found the initiatives to be generally successful. The criteria for success have included a decrease in number of families receiving CPS investigations, an increase in the level of services for families, and no increase in the number of children kept safe (Center for the Study of Social Policy [CSSP], n.d., 1996a, 1996b, 1997; Hernandez & Barrett, 1996; Siegel & Loman, 1998; Virginia Department of Social Services, 1999; Wakeling, 1995, 1996). Some of the findings are highlighted below:

  • The number of families investigated with resulting identifications on central registries decreased. In Missouri's and Virginia's pilot counties, only a minority of reports that would formerly have been investigated were investigated under their new systems, 33 percent and 27 percent respectively (Siegel & Loman, 1998; Virginia State Department of Social Services, 1999). In Iowa, only 16 percent of all reports in the assessment counties were considered significant and placed on the child abuse registry, compared to 32 percent of reports in the non-pilot counties (CSSP, n.d.).
  • The duration of family involvement with CPS also decreased. In Florida, case duration in the pilot counties was 56 days compared to 72 days in other counties, a 22-percent decline, (Hernandez & Barrett, 1996). In Missouri, families experienced a 15-percent decline in the number of days they were involved with the agencies between 1993 and 1997 (Siegel & Loman, 1998).
  • The use of community services increased for families in pilot projects. In Florida, the use of community-based services increased by 11 percent (Hernandez & Barrett, 1996). In Missouri, 25 percent of families in pilot counties received community services, as opposed to 20 percent in comparison counties (Siegel & Loman, 1998).
  • Child safety was not compromised in the pilot sites. Researchers in Iowa, Virginia, and Missouri all reported positive child safety outcomes. However, only the Missouri evaluation provided any quantitative outcome data. Researchers reported a 2-percent decrease in the frequency of repeated child abuse and neglect reports in the pilot counties compared to the non-pilot counties (CSSP, n.d.; Hernandez & Barrett, 1996; Siegel & Loman, 1998). English et al. (2000) offer a cautionary tale from Washington State. They found that the risk level and severity of some of the referrals to the family assessment track were inappropriately high. They also noted that the rates of re-referral were similar for families who did and did not engage in the assessment process.

While generally reporting positive results from the reform initiatives, the evaluations also qualify the level of benefits accruing from the reform initiatives. Hernandez and Barrett (1996) report that, in Florida, the better outcomes occurred in sites with higher implementation levels. Burrell (1995) suggests that reform efforts in Missouri are hindered by a lack of service resources and staff training in new methods of service delivery.

English et al. (2000) add that good initial risk assessments are crucial in a differential response model where the family's participation in an assessment is voluntary. This voluntary aspect is especially significant because research shows that high-risk families often resist acknowledging problems and, hence, would be less likely to participate in a voluntary family assessment and services delivery system. The authors suggest that the level of cooperation parents show should be a criterion in deciding whether a family is assigned an investigation or family assessment response.

The review of the above initiatives highlights the important realization that States using similar reform vocabularies may have different practices and that differences in vocabulary might actually obscure the similarities in how families are actually served. A common typology for analyzing such efforts would be useful.

Creation or Support of Community-Centered Services

Thirteen States were identified in the literature as having innovative community-centered services initiatives (AHA, 1996; Christian, 1997; Shirk, 1998). Geen and Tumlin (1999) and McCroskey and Meezan (1998) attribute the increase in community collaboration to the Family Preservation and Family Support Act of 1993. This legislation mandated that agencies convene a wide array of partners to help identify community needs for family support and family preservation services.

The implementation of community-centered services initiatives does not take the same form in all sites (AHA, 1996; Christian, 1997; Wakeling, 1995). One approach is to turn services over to non-traditional providers. For example, one district in Florida has turned over the assessments of substance-exposed and drug-dependent newborns to public health nurses. Jacksonville, Florida, has written up community- support agreements in which a community volunteer commits to help a family. When the agreement is signed, the case is closed, but the local agency remains available to the community helper for consultation. In Los Angeles, families are served by community-based networks, which have as their focal point non-child welfare agencies.

A second approach is to co-locate traditional service providers in a community-based site, convenient to clients (AHA 1997; Shirk, 1998; Wakeling, 1996). Iowa's Patch project, Louisville's Neighborhood Places, and Jacksonville's Full Service Schools are three examples of this approach. These programs bring together public agencies, such as CPS, income support, public health, and employment, with private agencies providing recreation programs, parent support groups, preschools, and community-building activities.

Several evaluations of community-centered services initiatives report that building collaboration is difficult (Apple et al., 1997; Burrell,1995; Hernandez & Barrett, 1996). Hernandez and Barrett (1996) found that the problems of community-based collaborations were different for communities that had few resources than for those that had many. In resource-poor communities, the difficulties for creating partnerships involved a lack of resources and a fear of risk. Community partners such as churches were already financially limited in what they could do. Non-traditional partners were also concerned about liability issues, fearing that they would be held responsible if a child died or was hurt. In resource-rich communities, problems were related to the complexity of coordinating the activities of a wide array of partners and developing data systems for sharing information. Apple et al. (1997) and McCroskey and Meezan (1998) also suggest that social workers have difficulty approaching community agencies and community members from a strengths perspective, which leads to communication and collaboration problems.

Based on experiences in a number of communities, the literature presents suggestions for overcoming such barriers:

  • Mulroy (1997) studied the Dorchester, Massachusetts, CARES project in which seven nonprofit health and human services organization collaborated to develop a community-based model to prevent child abuse and/or neglect. She found that the crucial elements for developing a dense network of interpersonal and inter-organizational relationship included developing mutual trust through the equal sharing of power over monetary resources; sharing decision-making among all partners; starting small and growing incrementally; sharing a common vision of child protection; and having a separate administrative infrastructure for the community network.
  • Farrow (1997) has made the following suggestions: include all of the major partners from the beginning; include parents as informed partners from the beginning; emphasize the mutual benefit of partnerships; define common outcomes; and use the new partnership to define the new approach politically.
  • A group of professional and natural helpers called People Helping People report on the challenges and advantages of developing natural helper/professional partnerships (Apple et al., 1997). In an appendix, they suggest concrete activities that natural helpers can conduct in neighborhoods to help build families' skill sets, provide emotional support, and encourage community leadership and networking. They also list detailed and concrete set of tools professionals could use in teaching natural helpers that would assist helpers in their work. These tools range from teaching different conceptual approaches to problem solving such as a systems approach or a cognitive approach, etc., to showing natural helpers what activities they might do with parents to motivate them toward positive change or to maintain changes parents have already made.
  • Scott and Bruner (1996) have written a guide to help State and local officials to develop local citizen review panels for CPS and form greater collaborative relationships between CPS, citizens, and consumers. The authors provide checklists for agencies on how to ensure the authority and independence of the review board, enlist citizen volunteers, staff the review, develop a case review process, and incorporate the citizen reviews into policy recommendations. They also provide checklists of activities an agency must conduct in order to incorporate consumer feedback into the CPS process, develop bridges to community institutions, support self-help networks, and broaden avenues for participation in policy making and service design. Scott and Bruner, with Veronika Kot, provide even more detailed citizen review panel guidelines and protocols in a later publication (Kot, Bruner, & Scott, 1998).

This review of the literature on community-centered services reveals a variety of initiatives States and localities are implementing in the name of community-centered services. Despite this variety, a unifying theme is found, which is that there is a need for a common vision and shared decision-making in order to achieve successful child protection community partnerships. What is not addressed clearly in the literature, however, is what changes, if any, these local community initiatives have made on the larger State child protection practices of which they are a part.

Clearer Delineation of the Relationship between CPS and Law Enforcement

The literature identified two types of involvement of the police in child protection issues. The first has been the inclusion of police in multi-disciplinary teams (MDT) and/or Child Advocacy Centers (CAC). These teams and centers, some of which have been in place for more than 10 years, have been formed in order to reduce the trauma to the child due to multiple interviewing and to improve the prosecution of cases, particularly in sexual abuse cases. They are an important example of the reliance that CPS agencies have had on community expertise.

In 1981, 16 States had MDTs, and the States used the teams primarily in a consultative role. However, there has been a significant increase in the use of these teams in the investigation of maltreatment and in the treatment of children. From a survey of State officials, Kolbo and Strong (1997) found that 33 States were using MDTs statewide for investigations and/or treatment planning. Twenty-six of the States included law enforcement as members of the team. In addition, a review of the National Children's Alliance's website showed that forty-four States have a CAC in at least one city in the State.

The second type of involvement has been the development of formal contracts to clarify the roles and responsibilities of the police and CPS in child maltreatment investigations. The main location in which these formal agreements are being attempted is in Florida. Christian (1997, p.20) notes:

It [a recent Florida law] also authorizes DCF to enter into agreements giving law enforcement agencies partial or full responsibility for conducting child protective investigations in cases involving a criminal investigation. These agreements are required to include protocols covering a number of items, including responses to abuse reports, conduct of investigations and performance of risk assessments.

A greater participation of law enforcement in CPS decision-making would clearly have an important impact upon the nature of CPS investigation activities and outcomes for children and families. Bollenbacher (2000) argues that the sharing of responsibility between the police and CPS is best for children. She suggests that either sole police or sole CPS responsibility for specific types of child maltreatment investigations is problematic for children because neither approach can secure a full range of rights for children. She argues that giving CPS sole responsibility over investigations leaves some children unable to secure their legal rights. In particular, it makes it difficult to prosecute abusers, torturers, or child killers in criminal court because the evidence gathered during a CPS investigation may not meet judicial standards. On the other hand, giving police sole responsibility makes it hard for children to secure their social and economic rights. In particular, it may make it hard for them to access needed services.

Increased Collaboration Between CPS and Alcohol and Other Drug Agencies (AOD)

Since 1997, five major reports have been issued on the alcohol and drug treatment needs of parents and children in the child welfare system (Gardner & Young, 1997; U.S. Government Accounting Office [GAO], 1998; National Center on Addiction and Substance Abuse at Columbia University [CASA], 1999; U.S. Department of Health and Human Services [DHHS], 1999; Young, Gardner, & Dennis, 1998). These reports describe innovative programs and explain some of the difficulties involved in creating closer collaboration between CPS and AOD. Several States have been identified as trying to build bridges between CPS and AOD providers.

Some of the variety in approaches is highlighted below (Young & Garner, 2000):

  • Sacramento, California, has trained large numbers of social workers in substance abuse issues. It has also changed its screening protocol and information systems to better capture and address substance abuse.
  • Delaware, Connecticut, New Jersey, and Ohio have hired substance abuse experts either in joint programs or through subcontracting to perform substance abuse assessments and treatment services. For example, Ohio began a treatment program for pregnant cocaine addicts staffed by teams with equal number of social workers and family advocates (recovering addicts).
  • San Diego, California, and Miami, Florida, have created special dependency courts. These courts refer addicted parents to treatment agencies affiliated with the court. These treatment agencies, in turn, agree to prioritize treatment for dependency cases and to provide regular treatment monitoring information to the courts. Special staff has been hired to do liaison work between the courts and the treatment agencies.

Research findings suggest that between one-third and two-thirds of substantiated child abuse and neglect reports involve substance abuse. Research also suggests that 30 percent of female problem-level drug users live with children, and 18 percent of male problem-level drug users live with children (DHHS, 1999). These overlaps in populations suggest the need for coordination and collaboration between CPS and AOD.

A general criticism in the literature is that the standard operating procedures of CPS and AOD agencies are not designed to facilitate collaborative relations between these two agencies. Examples of problems that are being examined include the following:

  • Standard AOD practice does not address the special needs of families in which child maltreatment is present.
  • AOD practice does not track whether or not its clients are parents. As a result, workers cannot evaluate the appropriateness of treatment programs for parents and non-parents.
  • Traditional AOD program design is based on confrontational tactics. This design may not work for women with low self-esteem and little past work experience.
  • AOD practice does not incorporate parenting education into its treatment programs.
  • AOD practice does not prioritize its limited treatment spaces for parents.
  • AOD agencies have traditionally not worried about client retention. If people do not show up, their spot goes to the next person on the waiting list. Child welfare clients, however, are likely to need more help getting started due to the possibly non-voluntary nature a court-ordered treatment plan.
  • Standard CPS practice is not oriented toward the special needs of parents and children with AOD addictions.
  • Many CPS risk assessment protocols do not screen for AOD, and even when they do, the linkages between substance abuse and actual risk of maltreatment is not well understood and is difficult to assess.
  • CPS agencies refer parents to treatment programs without knowing if vacancies exist or if the programs are useful. CPS workers may not know if parents can obtain services or have necessary child care. In other words, CPS referrals may take place when parents might be motivated to change, but not include sufficient follow-through to support the efforts of parents to enter and remain in treatment.

Gardner and Young (1997) suggest that progress in addressing AOD addiction and child maltreatment will not occur without better collaboration between CPS and AOD programs. In a forthcoming report, Young & Gardner (2000) suggest that while the development of pathways between these two agencies is critical to better supporting children and families with multiple challenges, the pathways must also be expanded to include the courts, mental health agencies, agencies that deal with family violence, the juvenile justice system, child development agencies, schools, and others.

Collaboration Between CPS Agencies and Domestic Violence Agencies

In 30 to 60 percent of families in which women are abused, the children are abused as well (National Council of Juvenile & Family Court Judges [NCJFCJ] 1999). Eight States have piloted innovative programs to improve their ability to deal with families in which both domestic violence and child maltreatment is present (Aron & Olson, 1997; NCJFCJ, 1998). These programs involve increased collaboration between CPS and domestic violence workers.

Several authors suggest that these collaborations are not easy to build, due to differences in mandates/orientations of the two agencies (Aron & Olson,1997; Beeman, Hagemeister, & Edleson,1999; Findlater & Kelly, 1999; Schechter & Edleson, 2000; Whitney & Davis, 1999). Aron and Olson (1997) identify a difficulty in building bridges between CPS and domestic violence workers in that there is a mandate for CPS to preserve families and reunify children with their parents, while domestic violence workers encourage women to leave batterers. Beeman et al. (1999) describe the principle barrier as the difference between CPS workers viewing their work as child-centered and domestic violence workers viewing their work as women-centered.

This tension results in contradictory approaches to similar problems. For example, in putting the safety of the child as the primary concern, CPS protocols may consider a mother accountable in domestic violence situations, even though the spouse or boyfriend is abusing the child (and her). Domestic violence workers, in contrast, argue that in such situations, holding the mother responsible is ineffective and results in blaming the victim. They suggest that if efforts are made to keep the mother safe, the children will also have a safer environment (Beeman et al., 1999).

The recent special issue on the co-occurrence of child abuse and domestic violence in Child Maltreatment (Edleson, 1999) profiles three such initiatives:

  • A program just beginning in Miami-Dade County in Florida is the Dependency Court Intervention Program. The goal of this program is to develop, implement, and evaluate a coordinated approach to the handling of child abuse cases in which domestic violence is also present (Lecklitner, Malik, Aaron, & Lederman, 1999).
  • In 1990, the Massachusetts Department of Social Services made domestic violence training mandatory for all of its new workers. It also hired battered women's advocates to provide consultation to its multidisciplinary assessment teams. In 1995, Massachusetts incorporated domestic violence protocols into its CPS investigations statewide (Whitney & Davis, 1999).
  • Michigan has developed and implemented statewide cross-training between domestic violence workers and family preservation and child protective services staff. One product of this cross-training was a national domestic violence curriculum for family preservation practitioners (Findlater & Kelly, 1999).

Underlying these change efforts are suggestions on ways to improve collaboration between these two domains. Beeman et al. (1999) suggest that collaboration can be improved in the following ways:

  • Cross training staff. Given that both types of agencies are concerned with issues of protection, cross training of staff would improve the ability to adopt mutually trusting working relationships. (Cross training manuals have been written by the Colorado Department of Human Services, 1995, Conroy & Magen, 1997, and Ganley & Schechter, 1996.)
  • Integrating and coordinating services through the development of specialized units that include both child protection and domestic violence experts. Central to the viability of these units would be the development of information systems that track male perpetrators and include men in service plans.
  • Fostering ongoing communication and consultation between child protection workers, domestic violence workers, and the courts to identify a common language around family violence. The authors recognize that the need for coordination of goals, activities, and services also extends beyond service providers to law enforcement agencies that may become involved in both types of cases.

In addition to the above resources, the National Council of Juvenile & Family Court Judges (1999) has published guidelines and recommendations for communities interested in improving their responses to families experiencing domestic violence and child abuse. The report offers community-level recommendations and describes specific recommendations for the three main partners in the process CPS agencies, domestic violence agencies, and the courts.