This pilot study illustrates the modification of evaluation planning tools, such as documents models [Rutman, Evaluation research methods: A basic guide, 1984], logic models [Burt et al. Evaluation guidebook. Projects funded by S.T.O.P. Formula Grants under the Violence Against Women Act, 1997] and program theory models [Weiss, Evaluation. Methods for studying programs and policies, 1998] in planning the evaluation of the coordination of domestic violence (DV) services in the metropolitan area of Baltimore, Maryland. Coordinated community responses (CCR) are systematic approaches to intervention—in this case for DV intervention—that emphasize comprehensive, collaborative and integrated service delivery. Evaluation studies typically focus on assessing the performance of programs within various components of the system (judicial, crisis intervention, social services, health care). The present paper argues for a holistic, theory-based evaluation approach to examine the entire system, which involves modeling the structure of, and linkages among, the system components. This process reveals that the activities of the agencies in the CCR are guided by varying intervention models, leading to different priorities: batterer sanctioning, victim advocacy, service delivery, or family reunification. Any subsequent evaluation of the intervention system as a whole has to take these differences into account. Data collection methods and indicators for a system-wide process evaluation are suggested.
Since the 1980s increased public awareness of intimate partner violence, or domestic violence (DV), as a social and public health problem rather than as a private family issue, has led to improvements in the legal system and the availability of DV services, such as shelters, counseling and hotlines. In 1994 the Violence Against Women Act was passed, which led to various legal and policy changes relating to DV cases. Nevertheless, the health care costs of intimate partner violence in the US have been estimated at around $1.7 billion annually ( DeLahunta, 1996), and lost productivity increases the estimate to a staggering $5 to $10 billion ( Cromwell & Burgess, 1996). Recently DV has also been identified internationally as a human rights violation ( Walker, 1999) and as the number one public health risk to adult American women ( Dwyer, Smokowski, Bricout, & Wodarski, 1995).
This pilot study is an illustration of the utility of modeling in evaluation planning for CCR. The intent was to model the CCR in the metropolitan Baltimore area in Maryland in order to establish a framework that can guide future holistic theory-based evaluation efforts. The insights gained inform evaluation research design and encourage improvements in performance evaluations of comprehensive community initiatives. Ultimately it is hoped that theory-based evaluation facilitates the improvement of services within agencies, increases efforts to promote inter-agency communication, coordination, and cooperation, and helps formalize the cooperative relationship between regional and inter-regional agencies. Studies of local systems can form the basis of comparative studies in various communities. Both system accountability and functioning can be assessed when a clear theoretical framework guides the evaluation.
Although the legal system has developed ties to DV programs, the coordination with other agencies, such as the health care system and the social services system, and agencies outside of Baltimore, may be quite weak. Barriers related to jurisdiction, regional authority, procedural differences, and competition over limited funds, may inhibit well-developed inter-agency and inter-county coordination. This pilot study shows that only in few instances—such as the staff attorneys and paralegals in the DV programs—are there actual liaisons, who are partners in more than one system and thus can provide continual assistance for victims. Such ‘bridges’ based on specialized expertise are necessary to ensure ‘continuity of care’, a concept already well established in the health care and mental health field (e.g. the Berkeley Model). Ideally, the system should be ‘weaving a seamless web of care’ ( Smith & Daughtrey, 2000) after using any point of entry into the system.
Recent changes have increased the number of DV laws and policies as well as access to a variety of services. Nevertheless, if the policies are not efficiently enforced and inter-agency coordination is not implemented as envisioned, DV survivors fall through the cracks of the system. It is crucial to examine whether the case-management philosophy is appropriate for DV service provision and whether most agency cooperation indeed occurs on a case-by-case basis. Although formal ties may exist in theory among the agencies, they may not be sufficiently well-developed in practice, which in turn may lead to a failure to meet victim needs. These issues can only be examined when the entire system is considered. A successful well-coordinated DV response system needs the cooperation of all of its components. This involves cross-fertilization of expertise in the form of liaisons in the various subsystems, data quality monitoring and sharing, and establishing principles of ‘continuity of care’, where the victim is central.
One helpful development in solving the coordination problems would be a standardized information systems data base, which could be accessed by all DV service agencies. This would facilitate tracing victims and offenders through the various stages of the service provision and the judicial process. To increase system accountability the data collection efforts of the various parts of the CCR should be coordinated. Due to legal changes and increases in service delivery more data than ever are collected but the differing reporting requirements of funding agencies have created incompatible formats of record keeping. Standardization of the different data collection schemes and computerization of all data collection procedures and reporting systems should make shared data bases with client records and service delivery much more manageable. In Michigan, for example, the efficiency and reliability of such coordinated data base systems (MichDISCAR project; see Wilbert and Li (1998)) have been shown.
