1 Introduction

Research on the epidemiology of child maltreatment has primarily focused on establishing estimates of the problem’s prevalence in the general population through cross-sectional community-based surveys. While these studies provide critical information about the scope of maltreatment, and, through repeated surveys over time, about changes in rates of maltreatment (Clement and Chamberland 2007; Jones and Finkelhor 2003; Straus and Mathur 1996), information about the identification and response to victimization is needed to inform policy makers and service providers about the efficacy of child protection laws and programs. To create strategies for change, policymakers need information about which agenciesFootnote 1 in their jurisdictions have knowledge of child welfare problems, and what they are doing or not doing when they encounter them (Jud et al. 2015a). Although a few high-income countries such as, the United States, Canada, the Netherlands, New Zealand, South Korea, and the United Kingdom, can report on agency response using administrative data or data from agency surveys, the evidence on agency response to child maltreatment is still scarce (Krüger and Jud 2015). In low-income countries, data on the response of child welfare agencies is virtually non-existent. For example, in a 2009 report on child protection systems in Indonesia (Boothby et al. 2009), the government acknowledged that it does not currently possess a system capable of providing accurate and timely information of key child welfare and protection concerns, including the magnitude of these problems, causality analysis and pattern, and impact of programs and interventions.

This paper emerged from a “toolkit” prepared for the World Health Organization and designed to summarize a range of recommended practices and resources for researchers and decision-makers in child protection who are interested in conducting national or regional surveillance studies of agency response to child maltreatment through the collection of administrative data or through surveys of professionals (Jud et al. 2015b). In the following pages, we expand on an agency mapping framework that had been developed as part of the toolkit. The framework was designed to assist researchers in generating agency sampling frames as part of a multi-stage sampling strategy for conducting nationally representative studies of child protection investigations. It provides a conceptual framework for identifying the various levels of formal to less formal organizations or agencies that are tasked with responding to situations of child maltreatment. The framework contributes to Wulczyn et al.’s 2010 UNICEF working paper on a “systems approach to child protection”, by focusing on the specific concepts that need to be considered in developing a framework for sampling child protection agencies across a range of different systems.

2 Developing a Sampling Frame

The agency mapping framework presented in this paper serves as a guide for researchers wanting to sample child protection agencies, as part of a multi-stage sampling survey designed to identify a representative sample of child maltreatment reports. Because the structure of the systems tasked with responding to child maltreatment varies considerably from country to country, this endeavor can be quite complex. A broad range of agencies can be involved in protecting victimized children and supporting their families—government-run child protective services, child protection teams at hospitals, not-for-profit helplines, psychotherapists at private practices, and community-based child welfare organizations, to name a few examples. Agencies vary along several dimensions such as size, ages of children covered, and professional background of child protection workers. In low-income countries, informal structures such as councils of community leaders and non-governmental organizations’ consortiums have to be considered. In decentralized systems with a range of organizations tasked with responding to maltreatment, mapping all agencies and organizations would be an overwhelming task. In such instances an agency sampling frame would need to be developed in stages, first by sampling administrative or geographic areas, and then in constructing a list of potential agencies to sample within the selected areas.

It is usually fairly simple to obtain a list of agencies in countries with state-run specialized agencies supported by mandatory reporting laws and investigation regulations (Fallon et al. 2010; Sedlak et al. 2010). But this task can prove to be far more difficult in countries where there is no centralized government structure that mandates reporting and investigation or assessment responsibilities, both in federally organized countries or countries where there is a lack of effective government (Jud et al. 2013; Wessells 2009). A first example illustrating the complexity of obtaining lists may be Switzerland where the 26 cantons (provinces) are responsible for the government-run child protection authorities and child welfare services. These agencies are sometimes organized at the cantonal, regional or municipal level and therefore vary considerably in size and caseload. Some child welfare services are organized as generic social services departments, while others are specialized on children and youth or even on mandated custodianships only. Complexity is increased as they are usually complemented by a range of non-governmental organizations (Jud et al. 2013). At a less formal and even more decentralized level, particularly in countries where there is a lack of effective government infrastructure to address child protection concerns, Eynon and Lilley (2010) refer to community-based child protection mechanisms (CBCPMs), systems that protect children from maltreatment and promote children’s well-being. Even in countries with relatively centralized child protection investigation systems, some marginalized populations may operate their own indigenous and community-based response systems outside of the mainstream structure of child protection authorities (Sinha et al. 2011). While referrals to less formal CBCPMs may not be included in government counts of child maltreatment “reports”, these cases are nevertheless situations where maltreatment moves from being a secret to being a request for some level of authoritative helping, protection or justice seeking intervention.

Depending on the number and size, different sampling strategies may be appropriate for different types of agencies. The United States’ Fourth National Incidence Study on Child Abuse and Neglect (NIS-4), for example, applied a multi-stage strategy, starting with a random sample of municipalities, within which a universe inclusion strategy was applied to child protective services, public housing departments, public health departments, children’s hospitals, sheriff and county police departments and juvenile probation departments (Jud and Sedlak 2015). For licensed day care centers, shelters, social services and mental health agencies a simple random sampling was applied, while municipal police departments, short-stay general hospitals and public schools were sampled proportionate to size.

In contexts where a mix of formal and less formal structures both have a major role in responding to child maltreatment, mapping is often carried out in a way as to include all potential stakeholders in providing a “snapshot” of how the child protection system is functioning. One example of this is a 2012 mapping exercise of Ghana’s child protection system that used a multi-stage strategy (Casey 2012). Six districts were selected in collaboration with an advisory committee. From these districts one community was selected as being typical or representative of that district. Within these communities, interviews and focus groups were conducted with a variety of stakeholders such as child protection workers, government representatives, NGO workers, chiefs, community leaders, as well as children and parents. This comprehensive approach results in data on the entire child protection system including key institutions and structures, as well as services for children and families.

3 Defining the Child Protection Organization Sampling Unit

One of the challenges inherent in developing a list of child protection agencies is that the very concept of child protection varies considerably from one country to another. Even within countries, there can be differences from one region to another. In comparing child protection systems across several high-income countries, Gilbert (1997) identifies two broad approaches to child protection: a child safety approach (the primary model in Anglophone North America and Australia) and child and family welfare approach (the model in, for example, Germany, Sweden and New Zealand). Differences between the two approaches include: (1) the extent to which service providing organizations are limited to dealing with child-protection or whether service providers cover a broader range of child and family problems, (2) whether there are mandatory reporting laws in place, (3) the extent to which assessments focus more narrowly on risk of maltreatment or more broadly on child and family needs, and (4) the scope of services provided, with child safety-oriented systems tending to be focused on investigation and short-term services and child and family welfare systems tending to offer a broader array of services, often over a longer term basis. To address some of the challenges inherent in confronting child maltreatment, both approaches have been borrowing from each other, making the distinction between the child safety and child and family welfare approaches more difficult to establish (Gilbert et al. 2009).

Mapping child protection agencies is generally a simpler task in systems with a child-safety orientation. In these more closely regulated child protection systems, all reports of suspected child abuse or neglect are sent to mandated child protection organizations that process the reports in a standardized way (Sedlak et al. 2010; Trocmé et al. 2010). Even in instances where some components of the investigation may be carried out by specialized hospital-based units or in children’s advocacy centers, case management typically rests with the mandated child protection organization. In these instances it is fairly straightforward to obtain a list of all organizations that conduct child protection investigations, to the extent that cases fit within the mandate of these organizations. However, careful attention should be paid to understanding the scope of these organizations as set by legislated mandates. In Canada, for example, mandates vary considerably by province: some provinces limit investigations to children under the age of 16, but others extend the scope of investigations to youth up to 18 years of age; some provinces specifically include exposure to intimate partner violence, and others do not (Trocmé et al. 2010) (see also http://www.cwrp.ca). It should also be noted that in many systems, extra-familial cases of abuse where there are no concerns about parental supervision might only be investigated by the police.

In jurisdictions that have a less regulated and more flexible approach—an approach that may very well be better tailored to a range of child and family needs—mapping out the pathways from detection to reporting to investigation may prove to be more difficult (Jud et al. 2013). The very concept of a ‘child maltreatment investigation’ could be quite different in these jurisdictions. In more severe cases of maltreatment, for instance cases involving contact child sexual abuse or physical abuse with serious injuries, it may be relatively simple to identify which authorities are mandated to conduct an investigation—the police, for instance. However, the label of maltreatment may not be used as readily in jurisdictions using a broader child and family welfare approach, because this label is not necessarily an entry point to accessing services. In cases involving a mix of concerns about parenting, family dysfunction, extreme poverty and child difficulties, there is likely to be significant variation in the extent to which child maltreatment labels, such as neglect or emotional maltreatment, are used. In these instances, the very concept of a ‘report’ or an ‘investigation’ may be difficult to ascertain, when a ‘referral’ or a ‘request for services’ and an ‘assessment’ may in fact better describe the front end of these services. In Germany, for example, only around 8 % of in-home services provide by public child welfare agencies were initiated in 2102 based on a disclosure of child maltreatment (Fendrich et al. 2014). In fact, even in jurisdictions operating within a child safety framework, the distinction between a maltreatment investigation and a family assessment has become nuanced with the introduction of “differential” or “alternate response” models (Kyte et al. 2013; Shusterman et al. 2005; Waldegrave and Coy 2005). As a result of this broader approach, the task of mapping child protection organizations is complicated by the fact that many different child and family service organizations can be involved in assessing maltreatment, ranging from the police, services run by local government social services, and hospital-based child abuse and neglect teams, to charitable organizations and private clinics.

Although child protection services in Canada are generally considered to be fairly closely regulated and child-safety oriented, our experience in conducting the Canadian Incidence Studies of Reported Child Abuse and Neglect (Sinha et al. 2011; Trocmé et al. 2010) shows that mapping out the list of organizations responsible for investigating child maltreatment poses unexpected challenges. A growing number of jurisdictions have developed differential response models designed to provide a more flexible service response, which may involve streaming less urgent cases of reported maltreatment to specialized teams or community agencies (Kyte et al. 2013). Even when differential response models are not formally deployed, we have found that some Aboriginal agencies use more of a child and family welfare approach, where it can be difficult to distinguish between a report, a referral and an informal request for assistance (Sinha et al. 2011). Likewise, determining which organizations are mandated to investigate maltreatment has been a challenge, as mainstream government mandated child protection organizations delegate a growing range of child protection responsibilities to Aboriginal agencies.

The structure and delivery of child protection services varies considerably from one jurisdiction to another and, upon closer inspection, may vary in unexpected ways within jurisdictions. The task of constructing a sampling frame of child protection organizations in a particular jurisdiction requires a thorough understanding of the formal and informal structures and mandates governing a range of organizations that conduct child maltreatment investigations or assessments. Even in child-safety oriented systems with mandatory reporting laws and centrally organized child protection services, careful attention needs to be given to ‘exceptions to the rule’, including what types of maltreatment fall under which mandates, which sub-groups might have access to alternate service delivery systems, and which age-groups are covered by which systems.

4 Child Protection Mapping Framework

In their review of a systems approach to child protection, Wulczyn et al. (2010) identify the essential elements of a systems approach and offer a framework for initiating conversations about child protection decisions. With structures, functions, and capacities serving as the basic building blocks of the child protection system, their proposed framework operates at several levels (from formal to informal), involves nested systems, and relies on a range of different actors (e.g., children, family, community, state, etc.). Drawing upon Wulczyn et al.’s (2010) review, we propose that researchers consider three key dimensions—(1) level of authority, (2) function, and (3) reporting process—to facilitate the task of developing a sampling frame to map out child protection organizations. The relationship between these three dimensions is illustrated in Fig. 1. The level of authority varies from a centralized legislated and regulated child protection system, to decentralized but legally mandated systems, to national and community-based organizations that receive reports as part of their professional mandates, to traditional authorities such as village elders, tribal leaders, or community-based child protection committees. The function of the authorities receiving reports ranges from helping, to protecting, to prosecuting. Finally, the role of these levels of authorities and their function will vary depending on the point in the reporting process where sampling may occur: from receipt of a report, to screening, to investigating, to making a determination, to intervening.

Fig. 1
figure 1

Mapping agencies by function, authority and process

Full size image

5 Level of Authority

When serious concerns arise about parental care, societies have developed a range of more or less formal and structured mechanisms to address a child’s need for protection. The level of structure and formality defining child protection authority can be classified on a continuum ranging from traditional community-based structures, to professionally mandated authorities, to legally mandated ones, to government-run protection services. Regardless of the level of authority, all of these structures represent some type of external authority intervening in situations where a child is not being adequately protected by their immediate family.

5.1 Traditional Authorities

In the absence of more formal government mandated authorities, the head of a family clan, a council of elders or local political or religious leaders may serve as the local authorities for resolving serious family conflicts or intervening when the care of a child does not meet community standards. Many countries have developed local Child Welfare Committees (CWCs), consisting of parents, teachers, social workers, and other community stakeholders, who address violations against children and make demands on local government to improve child protection mechanisms. Likewise, in Uganda, communities have a right to create by-laws that apply only to them, as long as they do not conflict with national laws (Child Protection in Crisis (CPC) Network 2012). On the other hand, ethnographic research on community-based child protection mechanisms in Indonesia found that child protection concerns were primarily addressed within the family, which was viewed as having the responsibility and potential to intervene and improve the situation of the child (Bancroft 2011). Community leaders were rarely involved and referrals to governmental or outside agencies were extremely limited. Even though such traditional authorities are not formally sanctioned or recognized by government, it would be a mistake to overlook them if they are de facto serving a child protection function, and in some countries serving as the only mechanism of child protection.

5.2 Professional Mandates

Beyond legislated protection mandates and reporting laws, health, education and social welfare professionals operate with their own code of ethics or work in organizations that may have rules with respect to child protection, though these codes and rules vary from country to country. Even if their potential child protection role is not regulated through a legislated government mandate, professional or organizational codes of ethics may very well require professionals to take measures in response to suspected child maltreatment. This sometimes leads to specialized services such as interdisciplinary child protection teams at hospitals (e.g., Kempe and Schmitt 1978; Jud et al. 2010). For example, NGOs providing support for sexually abused children via low-threshold helplines or walk-in counselling play an important role in increasing rates of disclosure in Germany, Switzerland and many other nations (e.g., Jud and Fegert 2015). As with traditional authorities, professionals may also play an important protective function in jurisdictions where child protection legislation is minimal or there is limited implementation infrastructure.

It is also important to note national child rights legislation that offer frameworks and standards of external child protection systems, oftentimes mandating organizations and professionals to protect children. However, parents in countries such as Indonesia (Bancroft 2011) and Sierra Leone (Child Frontiers 2010) have expressed skepticism of national legislation and their related organizations, as they believe that these laws infringe on their perceived ability to raise their children properly. While these national systems are intended to protect children by mandating organizations to protect children, they represent a challenge to the autonomy and control of parents, resulting in a sense of animosity and alienation on the part of the individuals who are most often responsible for taking care of children in the community (Bancroft 2011; Wessells 2009).

5.3 Legally Mandated Non-governmental Organizations

In many jurisdictions the legislated mandate to respond to child protection concerns is delegated to community-based organizations. In Ontario, Canada, for example, the receipt of reports, mandate to investigate and powers to remove a child are delegated by the government to local, not-for-profit Children’s Aid Societies. These agencies are typically organized on a geographical basis but cover several populations defined through religious (Catholic or Jewish) or Aboriginal status. In Switzerland, the authority to investigate and the mandate to provide services is even further decentralized, with a mix of public and private agencies structured in different ways in each of the 26 cantons.

5.4 Government-Run Protection Services

In the case of government-run services, protection authorities are in principle easily identified through the government service delivery structure. This is the most straightforward when child protection functions rest with a dedicated department or service, although it is less so when government departments or services cover a range of functions, such as a generic social services department where staff carry mixed caseloads. This is further complicated in countries where national systems of child protection and government actors are not considered to be feasible sources of referral for some communities. In Uganda, approaching a government official or child protection agency was rarely the first step to address child protection concerns (Child Protection in Crisis (CPC) Network 2012); rather, these resources were usually reached through a process of referral. Similarly, in Indonesia, distance and cost were the main barriers to accessing courts or police to address child protection concerns (Bancroft 2011). In addition to distance and cost, Indonesian families also cited fear of revenge or violence as a result of reporting child protection concerns to higher levels of authority such as government agencies.

6 Function

Related to a certain extent to the source of authority, function will also affect what types of organizations are included in mapping out child protection activities. Three primary functions drive the structure of child protections systems: helping, protecting and dispensing justice. Most organizations serve more than one of these functions, but in mapping out the structure of protection systems it is important to clarify which types of functions a study is being designed to document.

6.1 Helping

We refer here to the broader term of helping rather than treatment in order to include non-professional help and support that may be provided through extended family and community. Although one might expect that all child protection services are designed to provide treatment services, many in fact are focused on investigation and case management activities, with treatment functions being contracted out to specialized agencies. Organizations that play a helping function, such as children’s mental health centers, therapeutic treatment programs, or some children’s advocacy centers, include specific individual family or group treatment services designed to mitigate the effects of maltreatment and minimize the risk of re-victimization. Some protection systems include investigation and treatment services within the same organization, and others access services from specialized treatment organizations.

In child safety-oriented systems, treatment services are often tertiary level services accessed through post-investigation referrals, and a finding of maltreatment may be a condition of receiving services. In contrast, in child and family welfare-oriented systems, a request for treatment may be the first point of contact, and, if there are no mandatory reporting laws and no concerns that may require judicial or police intervention, the provision of treatment services may be the only protection activity involved. In Germany, for example, “parents have a legal right to child and youth welfare services (e.g., counselling, family preservation or placement services) even when child maltreatment is not an issue” (Jud et al. 2013). If parents are not willing or unable to address the child’s needs, child welfare services can opt for intensified protective activities by referring the case to family court.

6.2 Protecting

The protection function is central to the organization of services in child safety-oriented systems. The focus of these services is first and foremost on preventing recurrence of maltreatment through protective activities. In situations where a child can be kept at home safely, protection takes the form of supervising the home situation to ensure that parents have made the changes required to ensure the child’s safety. The alternative is removal of the child from the home, either on a temporary basis until the home is deemed safe again or on a permanent basis. The process of protecting a child from maltreatment through removal of that child from the home obviously varies from country to country, depending on child protection systems in place as well as cultural norms. Treatment services may also be provided, but the primary focus is to monitor the child’s safety. For example, the formal child protection system in Ghana has been oriented towards rescue and removal of children rather than working with families to address concerns at home (Casey 2012). Child welfare officers have limited capacity to provide follow-up and counselling support to children and families after the initial crisis-stage. Some NGOs try to fill this gap by providing counselling, psychosocial support, and legal aid, but their geographical coverage is limited.

6.3 Dispensing Justice

Whereas helping and protecting are generally thought of as being the central components of a child protection system, the police and the courts, as well as traditional leaders, serve a justice function through criminal proceedings or some type of restitution process. For example, in Liberia, chiefs are recognized by the national government as part of the judicial branch and are networked into regional chiefdoms. Chiefs govern child protection cases in the community and report cases to county authorities or magisterial courts, thus creating a community-based child protection mechanism that dispenses justice and also interacts at the national level (Lanning 2011). Nevertheless, there still remains a lack of judicial support for children in Liberia, with the only child-focused court located in the capital of Monrovia (Lanning 2011). Similarly in Uganda, community elders settle child maltreatment cases in locally relevant ways such as punishing perpetrators by requiring them to pay the child’s family, rather than focusing on sending that person to jail (Child Protection in Crisis (CPC) Network 2012). In a mapping of the child protection system in Ghana, the majority of community stakeholders felt that informal ways of addressing child maltreatment was preferable to the formal system, while a significant minority felt that it was preferable for the police to intervene so that there would be justice (Casey 2012).

Even in centralized child safety focused protection systems, some forms of child victimization, such as extra-familial sexual abuse or physical assault by non-family members, may be dealt with solely through police investigations. In many of these situations, if there are no concerns about the family’s ability to support and protect the child, child protection authorities will not be notified. As a result, surveys that only include data from child protection authorities, such as the Canadian Incidence Study of Reported Child Abuse and Neglect, may undercount reported cases of extra-familiar abuse and cannot be used to derive justice-specific estimates (see Figure 1-1 in Trocmé et al. 2010). Given that in some jurisdictions extra-familiar abuse may only investigated for the purpose of dispensing justice, a complete list of organizations responsible for conducting child maltreatment investigations may need to include the police.

7 Process

In mapping out child protection organizations one also has to consider at what stage of the process services need to be tracked. Counts of child protection case activities range from numbers of reports, to numbers of investigations, to numbers of substantiated reports, to numbers of children placed in out-of-home care. Because some organizations are only involved in parts of this process, one must clearly delineate which types of activities need to be mapped. The response to child protection concerns can be classified in terms of three key processes: receiving a report/referral, investigating/assessing, and intervening.

7.1 Report/Refer

In more structured child safety-oriented systems with mandatory reporting laws, specific organizations are mandated to receive reports. In jurisdictions with a child and family welfare orientation, the concept of a ‘report’ may be limited to the most severe forms of maltreatment, whereas situations involving a range of child and family problems may instead be referred for support or counselling services without going through a report.

7.2 Investigate/Assess

In a similar fashion, the assessment process will have more of an investigative form in child safety-oriented systems, whereas it may be more of a needs assessment in a child and family welfare-oriented system. Different types of investigation may involve different types of organizations, with criminal investigations conducted by the police, some forensic investigations by specialized medical units, and safety and protection investigations conducted by protection authorities. As a result, mapping out which organizations are involved in investigating or assessing child protection concerns requires a good understanding of these different potential processes. In child and family welfare-oriented systems the distinction between a protection investigation and a child and family assessment may be more difficult to establish.

7.3 Intervention/Placement

As noted earlier, the organizations that receive reports and conduct investigations are not necessarily the same as the organizations that provide services. In Switzerland, for example, reports of alleged cases of child maltreatment are received by the child protection authorities. They will then mandate an in-house or external service to assess the child’s and its family’s needs before deciding upon child protection orders. The mandate to implement the order is then transferred to a local or regional generic social service or a specialized children and youth service. In some instances services might be contracted out to other community organizations or community professionals; in others, the same organization is responsible for all three levels. Finally, some organizations, like shelters for street youth, may provide support and counselling directly to victimized youth without being part of the reporting and investigation sequence.

8 Conclusions

Constructing a sampling frame of child protection organizations is a key step in conducting any child protection agency survey. The framework that we propose in this paper is designed around a broad understanding of the full range of formal to informal systems that are designed to respond to situations where serious concerns about parental care, including abuse and neglect, may require some type of intervention. In many instances this mapping process should be viewed as a study in and of itself requiring dedicated resources and timeframe to ensure that the full spectrum of the diverse child protection agencies and related organizations are accurately represented. Mapping agencies is only a first step in implementing the agency selection stage of an agency survey. Developing a sampling design, applying weights and establishing a reciprocal practice-research relationship to commit agencies to participation are further steps detailed by a newly published international toolkit on mapping legal, health and social services responses to child maltreatment (Jud et al. 2015b). Written by researchers who have conducted a range of agency surveys, the toolkit includes chapters on all the key stages in the life cycle of an agency survey: selecting cases, developing the definitions, variables and codes of a questionnaire, disseminating the findings and implications for practice.

In the absence of a full national agency data collection infrastructure, child protection agency surveys are a key tool in assessing efforts to respond to child maltreatment. Recognizing the diversity of approaches that have emerged to assist victims of maltreatment, the framework presented in this paper provides a frameworkfor mapping the agencies and organizations tasked with coming to the aid of maltreated children and youth.