Keywords

The purpose of this chapter is twofold. We will first explore the underlying rationales for reporting and CPS response. We will then move forward to examine the evidence around both mandated reporting and CPS responses and services. We conclude with suggestions for policy and future research to provide a conceptual framework for understanding the issue of mandated reporting and to provide information relevant to understanding its appropriateness and utility relative to the broader child protective services system.

Why Have a Child Protection System?

The legal concepts underlying parental and child rights have undergone numerous transitions in the last two centuries and remain in a state of some contention. Parents have been variously seen as having a divine right to raise their children as they see fit, natural rights based on heredity, property rights over their children, and a societal responsibility to parent their child well, which implies to many the right to parent as they best see fit (Woodhouse 1992). The state is seen as having a right and a responsibility to intervene in supporting the child’s best interests under the doctrine of parens patriae, in which the state via the courts has a jurisdiction over the children which is even broader than parental power (Hart 2011). There is a general agreement that the innately powerless nature of childhood, best exemplified by the utter helplessness of the newborn, places children in a similar category to several other groups (e.g., severely cognitively incapacitated individuals) who are dependent on others to protect their rights.

We have therefore moved from a conceptualization of children as property wholly under parental authority to beings with autonomous and innate rights. These rights can be very specifically enumerated in principle (e.g., the Convention on the Rights of the Child (OHCHR 2015)) or can be operationally established in law (e.g., DHHS 2011). There can no longer be debate that the child’s autonomous rights, independent of the parent, are, at least in theory, generally recognized in developed western nations.

Missions, Statements, and Visions

While child protection laws and policies can vary radically from nation to nation and even state to state (e.g., Australia, the United States), the core general principles underlying child maltreatment policy and legislation are remarkable in that they tend to be quite consistent across nations and states.

Child Safety Is Framed as the Single Paramount Goal of Child Protective Systems

Supporting the child’s right to safety is unambiguously the paramount concern underlying virtually all international, national, and state child welfare legislation and policy. Clear examples abound, one representative example being: “The over-riding principle of the Act is that the safety, welfare and wellbeing of children or young people must be paramount” (Australasian Legal Information Institute 2014). One could easily modify the opening line of this paragraph in substituting the phrase “safety and well-being” for “safety” as a large number of documents do refer to child well-being as a paramount goal.

The Paramount Goal of Protecting the Child Is Best Pursued Through the Means of Working with the Family When Possible

There is also apparently a universal concordance around the value that if it can be done safely, a child should be maintained in his or her family setting, and the family should assist in making the child’s environment safe and supportive of well-being. Perhaps the most concise statement of this can be found guiding child welfare policy in the United States. The Child Abuse Prevention and Treatment Act (CAPTA) states, “The child protection system should be comprehensive, child centered, family-focused, and community-based” (DHHS 2011). The Convention on the Rights of the Child places a typical emphasis on the centrality of families, being “Convinced that the family, as the fundamental group of society and the natural environment for the growth and well-being of all its members and particularly children, should be afforded the necessary protection and assistance so that it can fully assume its responsibilities within the community” (OHCHR 2015). Here, we see shadows cast by long-standing ideas – both that the family has a standing to care for the child as they see fit rooted in natural or civil (DHHS 2011) law and that families have rights to make decisions about their children so that they can exercise their responsibility to society to produce the best possible next generation of citizens (Woodhouse 1992). Parental prerogative therefore must not be abridged frivolously, as this may both violate parental rights per se and may damage parents’ ability to discharge their obligation to society to contribute to a functional next generation.

Child Maltreatment Occurs in a Community Context

The word “community” is commonly present in legislation and guidance around child protection (OHCHR 2015; DHHS 2011). There is recognition that the community is the context in which families live, and abuse and neglect occur; that the community has a key role in identifying maltreatment; and that the community is where many preventative, supportive, or ameliorative programs exist and is both a venue and a means for furthering CPS work. For example, “The mission of The Texas Department of Family and Protective Services is to protect children, the elderly, and people with disabilities from abuse, neglect, and exploitation by involving clients, families, and communities” (Texas DFPS 2013). While “community” is commonly mentioned, it is an unavoidably nebulous term and one which tends to defy ascription of responsibility. For these reasons, policy language around community context and involvement appears frequently to be less actionable and well operationalized than other policy elements.

Responding to Maltreatment Is a Universal Responsibility

This is generally recognized. “All elements of American society have a shared responsibility in responding to child abuse and neglect” (DHHS 2011, p. 6). This recognition occurs even in jurisdictions lacking mandated reporting – “There are no mandatory reporting laws in England, but guidance issued by professional bodies and Local Safeguarding Children Boards emphasize the duty to make a referral where there is a reasonable belief that a child is at risk of significant harm” (NSPCC 2015).

Prevention Is Desirable

The word “prevention” is used, by our count, 54 times in CAPTA, including, obviously, in the title. “Prevention is the best hope for reducing child abuse and neglect and improving the lives of children and families” (DHHS 2011). Prevention is implicit in the Convention on the Rights of the Child, in which children are to be free from negative situations rather than having a right to “escape from” or “be treated for” such problems. Such language championing prevention is present in all legislation we have reviewed. Some locales are suggesting a fundamental realignment from what might be termed an “ambulance-like” service to a more general “public health-like” program: “Australia needs to move from seeing ‘protecting children’ merely as a response to abuse and neglect to one of promoting the safety and wellbeing of children. Leading researchers and practitioners – both in Australia and overseas – have suggested that applying a public health model to care and protection will deliver better outcomes for our children and young people and their families” (Council of Australian Governments 2009).

We are left, therefore, with a situation in which international, national, and local visions share much clear commonality at a general level. Child protection’s goals of safety, or perhaps “safety and well-being” as paramount, the preferred context of this being first the family and then community, the universality of responsibility, and the desirability of prevention, form a common vista.

Child Protective Services

In most countries, public child protective services agencies are tasked with identifying and serving children who need protection in accordance with the above policies. Child protective services vary markedly from place to place, but a series of core concepts do apply consistently across systems. It should be noted that relatively little attention has been paid to the following underlying concepts, and we hope to provide some clarification before moving forward.

Harm

Harm is a key concept in child maltreatment legislation and policy. It is not clear, however, that attention has been paid to thinking through what we mean by “harm” and whether CPS should be charged with providing services to protect from all forms of harm or more as an emergency response when there is a clear and critical safety issue. The most obvious exemplars of harm related to safety and threat of physical injury might include physical abuse or sexual abuse. These acute forms of maltreatment were the first popularly recognized (e.g., Kempe et al. 1962; Kempe 1978). By 1984, researchers were calling more attention to child neglect (Wolock and Horwitz 1984). Recent research indicates that similar levels of immediate harm can result from child neglect (Gilbert et al. 2009a, b). The need for a CPS role in regard to immediate safety or injury seems more generally accepted.

Harm, however, can also be understood in terms of denial of what is needed to establish healthy development or well-being (Davies et al. 2009; Noh and Talaat 2012). There are a number of ingredients that can be thought of as essential to healthy development, the absence of which potentially harms a child: (1) adequate food, shelter, and hygiene; (2) cognitive stimulation (particularly in the early years); (3) affection and nurturing; (4) supervision to help avoid hazards; (5) adequate health care; and (6) access to recreation for physical fitness and socialization with peers. The United Nations Convention on the Rights of the Child includes an even broader and more detailed set of needs. In terms of neglect and emotional harm, the idea of persistence is involved (Glaser 2011), meaning a single incident is unlikely to result in harm, but repetition or a pattern of incidents may have a cumulative impact. Various studies indicate that longer-term harm is equally likely from neglectful situations as it is from physical or sexual abuse (Gilbert et al. 2009a, b). It is unclear if this is the result of accumulated exposures to risk (Jonson-Reid et al. 2012). Harm can also be related to more indirect exposure such as domestic violence (Edleson 2004) or lack of adequate supervision of eating habits leading to obesity (Viner et al. 2010). There is much less consensus across states and countries about what constitutes the need for child protection when we expand the word harm to include the idea of future well-being.

Prevention

We have seen above how CPS is generally tasked with the prevention of harm. As we move toward a set of policies explicitly valuing well-being, it becomes important to not only prevent injury but to also prevent those circumstances which will interfere with healthy development. It is not clear that policy makers have taken this into account in designing child protective systems and preventative programs. We would therefore suggest that prevention can be thought of as an effort to reduce acute harm or an effort to interrupt ongoing circumstances likely to reduce well-being. Thankfully, a relatively small proportion of maltreatment reports involve serious immediate harm; many involve threats to well-being. Indeed, across most outcomes, relatively little difference is found in long-term outcomes according to system determination of harm in the United States (Drake et al. 2003; Jonson-Reid et al. 2009; Hussey et al. 2005; Kohl et al. 2009). Even when differences are found (e.g., Chiu et al. 2011), the relative number of unsubstantiated cases compared to substantiated cases makes this an important group. In the United States, among the three million unique children with screened-in reports in 2011, only about one in five (681,000) was “substantiated,” “indicated,” or classed as “alternative response victims” (U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012). These 681,000 children were proportionately more likely to receive post-response services, but due to their greater numbers, far more unsubstantiated children actually received such services. It seems clear that risk of harm, if one uses well-being as the metric, appears to apply to the majority of reported cases. The question becomes whether or not CPS is or should be tasked with providing more preventative services to this very large portion of cases. If protection from “harm” encompasses these broader long-term threats to development, this implies a response that will mitigate that harm.

How Do We Find Children Who Need Help?

If we move on with the assumption that children have a right to protection and we come to some form of consensus on what harm means, at least locally, then logically we need a response. This response would, at a minimum, assume there was a means of finding children who need help. We have already seen that virtually every society officially proclaims as a matter of policy that this is a priority. The method of identifying children who need help, however, differs.

There are essentially four potential means of identifying children who are experiencing some form of maltreatment, the first of which would be to avoid any mandatory reporting and formal state child welfare response (Melton 2005). While no rigorous test of such a model could be found in the literature, there is a trial of such a program in New Zealand (reported in Davies et al. 2009). Such a model requires the community to be knowledgeable enough to determine when a child needed protection and to be receptive to interacting with troubled families. Beyond this, there must be adequate resources and commitment to making these interactions effective. While the idea of a fully informal network has been identified among kin (Korbin 1994), there is some indication that this is not an insubstantial hurdle when it comes to identifying and caring for nonrelated children (Gaudin and Polansky 1986; Korbin 1994). Interestingly, in the section on detecting and reporting maltreatment in the World Health Organization’s Guide to Child Abuse Prevention (Butchart and Harvey 2006), we find the following statement (p. 71): “The usefulness of mandatory reporting is particularly questionable in situations where there is no functioning legal or child protection system to act on a report. At the same time, there is extensive evidence that the public as well as professionals are reluctant to act on knowledge or suspicions of maltreatment.” This would suggest that perhaps the early work on willingness to care for nonrelative children at risk is not without parallel in the international community.

In the current policy reality, there are three primary versions of reporting of abuse or neglect. In the first version, nobody is legislatively mandated to report (e.g., the United Kingdom). Typically, this involves a permissive or voluntary reporting system coupled with a state child protective services agency designed to respond to cases (Wallace and Bunting 2007). Permissive reporting relies on the adequate judgment of the reporter to know what and when to report. It also presumes that we are satisfied that enough of the children who need protection will become known to a permissive reporter. Such a system might, in theory, reduce the burden of having more reports than a system can handle. On the other hand, one might run a higher risk of missing children in need (Wallace and Bunting 2007). In the second version, some people, usually professionals whose work frequently brings them into contact with children, are mandated to report. This appears to be the most common form of mandatory reporting. In the third version, all people are required to report (e.g., Texas, Florida, and the Northern Territory in Australia). It is not clear how well this is operationalized for nonprofessionals. Other hybrid structures can be imagined and do exist in some places. For example, West Virginia would fit into the second version described above, except in the case of sexual abuse where all people are required to report.

While criticisms of mandated reporting abound, most of these are anchored in concerns that there is too much emphasis on investigation which seriously diminishes resources that could be provided for services (Melton 2005). As mentioned above, either having no formal policy or having only voluntary reporting could reduce the number of cases identified. It is not clear, however, whether voluntary or permissive systems might differentially place children unable or unlikely to disclose at greater risk, for example, infants or children with disabilities or victims of sexual abuse. Good empirical evidence, that a voluntary compared to mandatory system of detection is better or worse, is not yet available. So what do we know about mandated reporting? We will explore this partly by forwarding and replying to what we believe to be some popular misconceptions.

  • Myth: We have solved the problem of identifying maltreated children.

  • Reality: Most maltreatment still goes unreported.

One of the more difficult phenomena to accurately measure is the rate of actual maltreatment as compared to reported maltreatment. Some kinds of maltreatment, for example, physical abuse of preverbal children which leaves no visible injury, are impossible to identify absent self-reporting by the perpetrator. Many key child maltreatment studies (e.g., the National Survey of Child and Adolescent Well-Being (NSCAW)) do not even attempt to study maltreated children outside of the formal child welfare system. The largest studies to attempt to measure rates of actual maltreatment, as opposed to reported maltreatment, are the four waves of the National Incidence Studies in the United States (Sedlak et al. 2010). NIS-4 estimates that more than half of all maltreatment cases go unreported to CPS. There are several studies indicating that maltreatment is underreported by health-care professionals (Ben Natan et al. 2012; Flaherty et al. 2008; Markenson et al. 2007; Merrick et al. 2010). Studies of reporting in other countries have found similar problems among educators (Choo et al. 2013; Feng et al. 2010; Schols et al. 2013). It is difficult to know how this varies by type of maltreatment or victim profiles, at least in part due to deficits in current research. For example, a recent meta-analysis approach to estimating prevalence of sexual abuse found a dramatic difference between informant studies and self-report. This study suggests that underreporting was part of the reason for this difference (Stoltenborgh et al. 2011), but the same analysis could not be completed for neglect due to insufficient numbers of informant studies (Stoltenborgh et al. 2013). There is some indication that cultural beliefs may play a role in reporting in international studies (Ben Natan et al. 2012; Choo et al. 2013; Feng et al. 2010), but the magnitude of this impact in the United States appears to vary by type of reporter, region, and method of study (Ashton 2010; Ibanez et al. 2006; Krase 2013). Finally, there is some evidence to indicate that children with disabilities are underreported (Stalker and McArthur 2012). It is clear, however, that while some have asserted that “… whatever else one can say about child protection policy in the United States, it is clear that the primary problem is no longer case finding!” (Melton 2005, p. 10), this assertion is not supported by the evidence.

  • Myth: Mandated reporting laws cause large increases in report rates.

  • Reality: Evidence suggests reporting laws are not the main driver of report rates.

The United States offers a natural laboratory for understanding the degree to which mandated reporting laws cause increases in reports. One way to understand this phenomenon is to look at reports over time. The vast increases in reporting from the “discovery” of child abuse by Kempe until now happened well after mandated reporting laws were put into place. All states had mandated reporting by 1967 (AHA 1979), but the bulk of reporting increase happened more than 10 years after this. For example, between 1977 and 2003, reports per 10,000 children quadrupled (Drake and Jonson-Reid 2007; AHA 1979; U.S. Department of Health and Human Services, Administration on Children, Youth and Families 2005). Since 2003, the rate of child abuse reporting in the United States has only increased slightly moving from an estimated 2.9 million referrals in 2003 to 3.4 million estimated referrals in 2011 (U.S. Department of Health and Human Services, Administration on Children, Youth and Families 2005, 2012; CWIG 2012).

While reports from professional reporters increased faster than reports from nonprofessionals (4.88 times vs. 3.25 times), this could well be due to the proliferation of helping professionals (Finkelhor and Jones 2006) during this timeframe, rather than increased observance of mandated reporting laws which had already been on the books for a decade. Even if reports from professionals had escalated only at the (3.25 times) rate of nonprofessionals, total reports would only have been reduced by 19 % (Drake and Jonson-Reid 2007).

Another way to understand this problem is to look at the official report rates from those states in which everyone is a mandated reporter and compare them to the remaining states in which only designated professionals are mandated. Among the 18 states with universal mandated reporting, we found that the average report rate is 54.0 per 1,000, while the report rate for the remaining states is 52.8 per 1,000 – virtually identical numbers. These figures were calculated by averaging the report rates (U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012) from each of the 18 states with universal mandated reporting (CWIG 2012) and comparing them to the remaining states and the District of Columbia, while excluding West Virginia, which has universal mandated reporting for sexual abuse but only mandated reporting for professionals for other types. It should be remembered that in the United States as a whole with universal mandated reporting by professionals, almost half of all reports (42.4 %) were submitted by nonprofessional reporters.

The evidence presented by the historical data seems clear. The massive escalations in reporting happened more than a decade after mandated reporting laws were established, and expanding mandated reporting universally is not associated with any notable differences in reporting.

Another way we can look at these historical data is to understand how types of substantiated maltreatment have changed over time (data not being generally available for types of unsubstantiated reports). There are undoubtedly a number of factors which could influence changes, including different rates of actual maltreatment, different reporting tendencies, and different state policies regarding what kinds of reports are accepted. In comparing 1977 (AHA 1979, Figure 10) and 2011 data (U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012, Table 3.8), we can see that there were more multiple findings in 1977, with a total of 1.54 types of maltreatment recorded per case, compared to only 1.27 types of maltreatment reported per case in 2012. This is undoubtedly partly due to the fact that more subtypes of maltreatment were reported in 1977, with, for example, seven different types of physical abuse being broken out, as compared to simply “physical abuse” in the 2011 data.

Several types of maltreatment remained relatively stable across the years in terms of their proportion among substantiated reports. For example, physical abuse in 1977 was 23.7 % of substantiated reports and in 2011 it was 17.6 %. Neglect was present in 82.2 % of cases in 1977 and 78.5 % in 2011. Other/unknown cases stated stable at around 10 % (9.7 % in 1977, 10.6 % in 2011). The one type which rose dramatically was sexual abuse, being 5.8 % of substantiated reports in 1977 and 9.1 % of substantiated reports in 2011. Two kinds of maltreatment dropped, “emotional neglect” comprised 24 % of substantiated reports in 1977, but “psychological maltreatment” comprised only 9 % of all reports in 2011. It is unclear how similar these categories might actually be. Medical neglect also dropped from 9 % of all cases in 1977 to 2.2 % of all cases in 2011.

Interpretation of the above data requires awareness of several factors, such as the fact that neither timeframe includes 100 % of the states nor that the total number of reports has roughly quadrupled over time. However, it is interesting to note that the proportion of sexual abuse reports is now higher than in the past and that psychological maltreatment may be lower. This latter point is probably due to more restrictive agency guidelines regarding screening in or substantiating psychological maltreatment. No matter how these data are viewed, they do not seem to indicate a proliferation of lower risk or meaningless cases.

  • Myth: Child Protection is overwhelmed by investigative responsibilities.

  • Reality: Investigations are proportionately a very small burden on the system.

Based on available data, the evidence suggests that the cost of investigations to child protective service agencies is small, most likely below 10 % of total costs and possibly below 5 %. The most recent estimate of the total cost of child protective services in the United States dates from 2006 (DeVooght et al. 2008) and is 25.7 billion dollars per year. Adjusted for inflation, this would be 29.8 billion as of 2013. The Cost of Protecting Vulnerable Children IV (Scarcella et al. 2004) found that about half of the total child protection expenditures were for foster care and residential services. Investigative costs were so small as to not even comprise their own category, being included together with, among other things, “all prevention services, child protective services, family preservation services, reunification services, and in-home support services” (p. 5). These expenses together accounted for 14 % of all expenditures.

There have been several recent comprehensive workload studies done on caseload-carrying child protective service workers across the nation. In Washington (Washington State Department of Social and Health Services 2007), 18 % of caseworker effort was characterized as being spent on investigations and associated work (e.g., paperwork, travel), whereas the corresponding number in New York was 19 %. These numbers pertain to workers with caseloads only and exclude administration, management, support, research, and other staff. This 20 % figure therefore reflects time spent by only a portion of the workforce, and workforce costs are only a portion of the total child protective service expenses, with out-of-home care payments, and contracted services absorbing a higher level of resources. Even though about 20 % of state worker time is spent on investigations, the actual proportion of state child protective resources spent on investigations is clearly far lower.

We can use the data in these studies to approach this question from another perspective. For example, the New York workload study estimated that the average investigative case took about 5.2 h and that a reasonable monthly number of cases served per worker might be about 24, yielding an annual caseload of 288 cases. Given that there are about 2,000,000 cases investigated annually (DHHS 2011), this would imply a need for about 7,000 full-time employee equivalent investigative workers nationally. This number is useful to consider, even though investigative work is often parsed out as a “part-time” job for workers who have other responsibilities, especially in rural areas. A quick perusal of online employment and salary information websites suggests reimbursement commonly in the mid-$30,000 range for child protective workers, and, after adding fringe benefits to establish a cost-to-agency figure, we can confidently say that the average child protective services investigator costs the state something in the area of $50,000 per year. This implies a total salary cost associated directly with child abuse investigations of something like $350,000,000 dollars annually for the United States as a whole. Compared to the estimated total cost of child protective services nationally, we see that investigative worker salaries probably comprise slightly more than one percent of total expenditures. We must, of course, add costs of training and supporting services, such as supervision, but even then, this means of estimating the total percentage of resources spent on child abuse investigators can only yield an estimate in the single digits, probably the low single digits. These calculations are also confirmatory of the Urban Institute estimates. The conventional wisdom that child welfare agencies are being catastrophically drained of resources by overwhelming investigative responsibilities is a persistent one. In reality, the burden of investigations on the total child welfare system is probably proportionately less than the sales tax you pay on a cup of coffee at Starbucks.

  • Myth: Substantiated cases are “real” cases, and unsubstantiated cases are “bogus.”

  • Reality: Unsubstantiated cases feature almost the same risk as substantiated cases.

It is perhaps natural to instinctively understand “substantiation” as meaning “abuse or neglect happened” and to assume that “unsubstantiated” means “there was no abuse and there is no risk.” There may be a tendency for people to think of substantiated and unsubstantiated cases as being very different – polar opposites, in fact.

This perspective was challenged two decades ago (Leiter et al. 1994) by a study finding that children’s outcomes were not very different by substantiation status. Since that time, there have been a large number of studies showing that substantiated and unsubstantiated cases are at similar risk of negative future outcomes and re-reports of child maltreatment (Drake et al. 2003; Kohl et al. 2009; Hussey et al. 2005; English et al. 2002; Fakunmoju 2009). This may be due, in part, to the fact that substantiation requires both demonstration of harm or risk and clear evidence of the cause of the harm (Drake 1996), meaning many cases with low evidence cannot be substantiated. It is also the kind of result you would expect if the majority of reports involve persons who are at risk of child maltreatment. While it is clear that not all studies find equal risk of untoward outcomes across all domains (Chiu et al. 2011), there is good evidence to believe that the group of children who are assessed by child protective services are at higher risk than other children, even after controlling for other factors (Jonson-Reid et al. 2009).

Another way to understand the importance of unsubstantiated cases in prevention efforts is by noting that far more unsubstantiated than substantiated cases receive state post-response services (U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012; CWIG 2012). Even though such services are only provided to about 30 % of unsubstantiated children, compared to 60 % of substantiated children, the far larger number of unsubstantiated cases means that more unsubstantiated children than substantiated children get post-response services (747,369 compared to 358,838 based on 46 states reporting both statistics, U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012; CWIG 2012). Since child protective services are tasked with a preventative role, and given that unsubstantiated cases are at similar risk for re-report and other negative outcomes, it is not rational to simply classify all unsubstantiated cases as “misses” or as examples of inappropriate reporting. Quite the opposite is true – they provide a genuine opportunity for CPS to execute its primary mission – the prevention of harm to children at risk.

How Do We Protect Children Once Identified?

Depending on the means of identification, the response might vary. In a world without any reporting system, the response (if any) would have to be fully community driven. Communities would require the resources to be a positive support for families. One of the dilemmas of such a system is that it has been clearly established that those communities with the highest rates of child maltreatment are also those which are most disorganized and least efficacious (Coulton et al. 1999; Drake and Pandey 1996). While the nature of a causal relationship between community level factors and maltreatment rates remains a matter of debate (Coulton et al. 2007), many people suggest that structural inequalities over time (geographically concentrated, persistent poverty) result in highly disadvantaged communities which in turn place additional strain on struggling families (Blackstock and Trocme 2005; Jack 2004). In order to radically change this, there would need to be a societal willingness to fund what Sampson called “changing places not people” (2003). It is certainly a worthwhile idea to invest in communities to build the informal infrastructure for families (see Davies et al. 2009), but until sufficient community-based supports can be achieved to address maltreatment informally, some type of formal detection and response seems warranted for those cases that meet the standards of maltreatment in a given region, even strong advocates for community-based prevention call on both community responsibility and a strong formal service infrastructure (Daro and Dodge 2009).

Most countries opt for some level of formal agency response either operated through a governmental agency or contracted private agencies. This is true even for countries in the early stages of developing a formal response to child maltreatment (e.g., Almuneef and Fadia Al Buhairan 2012; Choo et al. 2013). Even in a qualitative study advocating for increased recognition of informal community supports, there was no call for eliminating the formal response to child protection (Holland 2012). Currently, several types of systems exist. Relatively, well-funded and established national child and family service systems exist in some nations (e.g., the United Kingdom, the Netherlands), while some nations have only a rudimentary or patchwork child welfare system. The United States provides a good example of the diversity which can exist in service systems.

Mandated Reporting: A Linkage to Services?

Current thinking about mandated reporting policies cannot be understood without reference to the systems and services that receive and act upon these reports. Early on in the history of detecting maltreatment, however, it is clear that a formal linkage between existing institutions and the potential reporter did not exist. For example, Kempe et al. (1962) exhorts each physician to “acquaint himself with the facilities available in private and public agencies that provide services for children. These include children’s humane societies, divisions of welfare departments and societies for the prevention of cruelty to children” (p. 111). While there has been a substantial change in infrastructure since that time, it is not clear that a fully evolved “system” has been achieved.

  • Myth: The US Public Child Welfare is a comprehensive child protection system.

  • Reality: The US child protection response is more like a patchwork quilt.

In the United States, while many systems contact children, the burden for assuring safety and well-being has fallen, by default, onto state and county departments of social services and the child protective system. This default assignment is now enshrined in law, for example, in CAPTA and in the Child and Family Service Reviews required by the US Government monitor states on how well their child protective service systems support child well-being. Child protective services are not currently charged with only protecting children from acute harm; they are also charged with supporting child well-being.

Perhaps because of these federal legislative efforts, child welfare systems are often conceptualized and evaluated as if they represent a comprehensive system of care. In most cases, they do not. Child welfare systems do have well-developed reporting protocols, established procedures for interfacing with the courts, and guidelines related to providing foster care. These are services virtually unique to public child welfare and are federally mandated and regulated. These services are remarkable for being reactive and represent tertiary rather than primary or secondary preventative roles. The more preventive service responses vary substantially by state and region.

When it comes to a preventative or early intervention role, the American public child welfare system is much more variable and tends to resemble a patchwork quilt rather than a comprehensive system. For example, the most common form of services is case management and referral, which are typically provided by public child welfare workers. The actual direct services or concrete supports, however, are delivered by other agencies (e.g., family counseling, housing assistance, addictions services, and parenting programs). This creates a unique dynamic where the outcomes of the “child protective services response” are actually dependent on agencies with whom they have little or no control and may or may not have a fiscal relationship. While this section is focused on the United States, it is worth noting that this natural dependence on other systems or services to produce positive outcomes for children is by no means limited to the United States (e.g., Almuneef and Fadia Al Buhairan 2012; Munro 2011). This is one of the key problems in holding the child welfare agency responsible for preventative functions and well-being outcomes (Barth and Jonson-Reid 2000). By way of comparison, other systems work very differently. For example, the 911 system – a general system in the United States of emergency telephone calls to police, ambulance, and fire services with many other countries using a 000 number – receives distress calls, clarifies the issue at hand, identifies the needed service, and responds to the problem. These intermediary responders (police, ambulance, fire) may then rely on further systems, courts and jails, hospitals, social service agencies, insurance companies, and the like to meet the more distal needs of the person served. How well this system works relative to the final outcome for an individual is, like child welfare, dependent on available resources and ability to triage to further intervention as needed. Unlike child welfare, however, the secondary response systems here are much more clearly aligned with the initial emergency responders. Some countries have child welfare systems that more closely align with this 911 approach with explicit protocols for government versus agency responsibility and data sharing to track cases (Angman and Gustafsson 2011; Wallace and Bunting 2007). There are a few regional examples of this type of close-knit collaboration in the United States (Daro and Dodge 2009). There are, however, no national or even consistent state-level models.

If we wish to have a system that is more like a 911-initiated system, then we must have adequately funded entities equivalent to a fire department, ambulance company, or police force and there must be an established secondary level (within or outside the public system) for taking over the case once the initial response is done. Likewise, the evaluation of such a system must be appropriately linked to the various roles and responsibilities so that improvements can be made. For example, if there is an increased rate of death from heart attacks, we would want to know if it was a training issue for responders, lack of quality care in hospital emergency departments, specialty cardiac services, or even something like traffic delaying the response time (Griffin and McGwin 2013). Calls for a comprehensive data system that allow for such tracking of services and outcomes have been made in both the United States and international literature (e.g., Jonson-Reid and Drake 2008; Munro 2011). This is critical as the remedy for the problem is very different based on where the problem lies. Arguably, we have a long way to go before we achieve this type of system.

  • Myth: Mandated reporters and parents receiving CPS intervention dislike CPS interventions and find them counterproductive.

  • Reality: They are generally satisfied with CPS interventions and think it helps assure child safety.

One of the strongest enduring myths in the mandated reporting debate is that mandated reporters are largely dissatisfied with the system as it exists. To the degree that dissatisfaction with the current reporting system does exist, the principal complaint is that child protective services do not take enough action or will not accept cases that the reporters consider serious, not that child protective services commonly overreach or are unnecessarily intrusive (Cocozza and Hort 2011; Drake and Jonson-Reid 2007). Various surveys of mandated reporters (Berlin et al. 1991; Compaan et al. 1997; Flaherty et al. 2006; Kalichman 1991; Kalichman and Craig 1991; Weinstein et al. 2000, 2001) have arrived at very similar findings, principally that reporting is helpful rather than harmful to families and that reporting aids in assuring child safety. While mental health professionals report some disruption to treatment in about a quarter of cases, “Several studies have found that making a report of suspected child abuse or maltreatment concerning a client in psychotherapy is more likely to have a positive outcome for the relationship or to effect no change, than to be damaging”. In one survey, mandated reporters were asked directly if they believed mandated reporting laws were necessary, and almost all (94 %) responded that they believed they were (Kalichman and Craig 1991).

Perhaps surprisingly, clients are also generally pleased with CPS contact, with one large study showing that more than three quarters of investigated parents were “satisfied” or “very satisfied” with the investigation, two out of three saying their family was doing better after the investigation, and less than one in ten saying they were doing worse (English et al. 2002). These general findings have been replicated in numerous other studies (e.g., Chapman et al. 2003; Fryer et al. 1990, 1988). Of course, satisfaction and effectiveness and even adequate coverage are different things.

Do CP Services Work?

This is a difficult question. Most children and families that are reported (whether by permissive or mandated reporters) do not receive any response other than an assessment or investigation (Jonson-Reid 2011; U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012). Unless the situation is clearly one involving very serious abuse or very high risk to the child – for example, sexual abuse or physical abuse of an infant – typically, more than one report (and often a few) is required before the priority to provide services is high enough to trigger a child welfare response. If we assume that a large proportion of children (and their families) reported a need for something but most will get nothing, then it is difficult to assess the overall system impact unless there was a significant deterrence effect or other benefit of the report itself. The idea of deterrence is based on the fact that you have been made aware that you have violated an established societal norm and that an authority figure both is aware of this and is able to become aware of future violations. This is similar to the argument of building community responsibility related to the control of youth behavior (Sampson 2003). It is predicated, of course, on the child’s parents or caregivers valuing the societal norm and the opinion or the response of the person who notices and having the capacity to alter their behavior toward the child. Second, there could be a value to the child. Perhaps there is a value for a child in knowing that what is happening is not acceptable or desirable separate from whether or not social agencies can effectively intervene (Wekerle 2013). Most would argue, however, that a positive outcome for a child protection response would more likely follow from some sort of service.

Current Models for In-Home Response

Differential Response

Sometimes termed “alternative response” or a “two-track system,” differential response (DR) represents a structural redesign of child abuse investigations and services. Prior to DR, agencies had a single “track” wherein cases of suspected maltreatment, whether reported by mandated or non-mandated reporters, could be investigated, after which the more serious cases might involve removal of the children or other court action, and less serious cases might receive services or be closed. Differential response is an attempt to split out cases earlier, often at the initial agency telephone intake, and provide either “investigations” – generally in the case of sexual abuse or other serious cases – or “assessments” in other less serious cases to provide the appropriate service response for respective cases. In this way, a kind of triage system operates at an earlier stage to more promptly and efficiently refer families to necessary helpful service providers in appropriate cases and to reserve more formal investigations only for cases where a child is or is more likely to be in need of protection.

Differential response is being configured differently in different states, and many open questions exist, including, for example, the degree of substantial practical differences between investigative and assessment tracks. Differential response was employed in 18 states as of 2011 (U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau 2012). This approach is also being explored internationally (Davies et al. 2009; Mathews and Bromfield 2012). The assessments are designed to trigger a case plan and connection to services. Early research (e.g., Siegel and Loman 2011) suggests that consumer satisfaction may be higher, while child safety appears not to be compromised.Footnote 1

Family-Centered or Case Management Services

Following an assessment or investigation, families may be offered voluntary in-home services that are essentially a form of case management. Research on lower level in-home services is mixed. In a study of families with a first report of maltreatment, those families receiving the lowest intensity of in-home services had lower recurrence rates than those receiving more intense (family preservation or intensive in-home services or foster care) or no services in data that also control for interactions with substantiation and other parent-level services received (Drake et al. 2006). In another study that grouped all post-investigative services together, there was no significant effect of services among unsubstantiated case and a moderate increase in risk among substantiated cases (Connell et al. 2007). Another study found a moderate increased risk of recurrence for served unsubstantiated cases that was mostly offset for substantiated cases, but non-child welfare services were not controlled (Fluke et al. 2008). A recent study found higher risk among children receiving some form of child welfare services (Fuller and Nieto 2013), but there were no controls for poverty or other services nor was the type of in-home services clear.

Why such a range of findings? First, as mentioned above, there are variations in measurement that may result in different outcomes. Additionally, some research indicates that different forms of maltreatment and intensity of initial investigation are associated with particular service recommendations which may confound understanding recurrence (Bae et al. 2009). Further in studies that have included assessments of parental engagement and readiness for change, this factor has been significant in improvements in family functioning and reduced maltreatment (DePanfilis and Zuravin 2002; Hindley et al. 2006; Littell and Girvin 2005). Most studies of services and recurrence have been unable to also capture such family level variables. Dosage of services (frequency of contact) is also at issue. In an analysis of a nationally representative group of families engaged in child welfare in-home services after investigation, only 66 % reported having seen their caseworker in the last month (Chapman et al. 2003). Finally, as mentioned earlier, the most common form of service is case management and referral, meaning that successful outcomes are frequently dependent on the other services offered and accessed.

Intensive In-home Services (IIHS)

Also known as family preservation services, this form of child protection response is designed for families that are at imminent risk of having a child removed from their care due to maltreatment but are diverted to these intensive services instead. Data regarding the effectiveness of IIHS are mixed. For example, the California Evidence-Based Clearinghouse on Child Welfare rates homebuilders, perhaps the best known IIHS program, as “supported by research evidence” but not “well supported by research evidence,” despite a very large set of studies on IIHS over the past several decades. The research has been confused and contentious to the point where articles have been written on how best to evaluate IIHS evaluations (Jacobs 2001). A key problem is that it appears that IIHS effectiveness varies markedly by client characteristics. One recent meta-analysis (Channa et al. 2012) found that these programs “were effective in preventing placement for multiproblem families, but not for families experiencing abuse or neglect.”

Thinking Ahead: How Do We Move to a Reporting and Response System Best Informed by Research Evidence?

More than 50 years after Kempe’s groundbreaking article in 1962, several forms of reporting and response have developed throughout the world. There remain, however, many gaps in our understanding of mandated reporting and child protection system response. Further, much of our data that does exist is limited to studies in the United States, Canada, the United Kingdom, and Australia. We know very little about the types of reporting and response systems in other countries. What follows are suggested areas of research or methodological issues that need to be addressed to improve our evidence base to protect children. We begin with issues related to reporting and then focus on services to intact families.

Consensus on “Maltreatment”

What is the appropriate metric for determining what is “abuse” and “neglect”? In the United States, responsibility for “defining” maltreatment is left to the individual states and changes over time. This makes it very challenging to understand the prevalence of different types of maltreatment and to understand effective response methods. A first step in improving measurement requires the setting of boundaries regarding the concept of harm and what constitutes risk (Davies et al. 2009). It is clear that some countries cast a very wide net in terms of what the government or agencies are responsible for in terms of supporting child development (Angman and Gustafsson 2011). It is not clear that other countries like the United States are accepting such a broad mandate. In cases where definition change precedes such acceptance of responsibility and preparedness of response, there have been poor outcomes (Edleson 2004). What is “bad” for a child in the context of their family may not necessarily be the same as what we decide is abuse or neglect in terms of a CPS response. Whatever is included in the definition of maltreatment related to reporting should be connected to an adequately resourced response.

What Form of Reporting Is Better?

Ideally, we would have a better understanding of permissive reporting versus mandated reporting versus fully community-based approaches that have no formal system at all. As discussed earlier, there is substantial evidence that maltreatment is underreported even among those mandated to report. This confounds the ability to assess this type of policy. Some research indicates that mandated reporting improves case finding (Al Eissa and Almuneef 2010; Mathews et al. 2010). Pritchard and Williams (2010) compared child abuse-related deaths (CARD) in infants with non-child abuse fatalities and found that CARD rates dropped relative to other causes of death in England and Wales, whereas this was not true in other countries like the United States. As aforementioned, the United Kingdom has more of a permissive reporting system. Comparing countries – or jurisdictions within countries – with different systems may help in this matter but has to be done with extremely careful attention to maltreatment definitions and the perceived and real value of the response system in place. In countries with mixed systems like the United States, we also need to understand the impact of reporting in the context of whether reporter type influences the system response.

When Reporters Choose Not to Report, What Happens?

It is clear that even mandated reporters do not always report (Sedlak et al. 2010). Questions have been asked about professionals’ comfort with reporting (Gilbert et al. 2009a, b), but this is not the same as asking what types of alternative actions were taken instead of reporting. There is also some indication that mental health-care professionals may not act on self-reported maltreatment because they do not believe it to be a priority issue in treatment (Read et al. 2007), but little work has been done in this area, nor is it clear whether or not most potential reporters who choose not to report have ongoing contact with the child and family and, if so, what (if anything) they did to help the child and the child’s family and whether this was successful.

Similarly, we do not really know much about how known cases that went unreported fare. We have retrospective recall studies like the famous Adverse Childhood Experiences Study (Donga et al. 2004), but we do not know who was reported and who was not and what happened. There is indication that future death and injury might have been prevented if abusive head trauma had been properly identified and reported in some studies (Jenny et al. 1999). However, severe physical abuse is relatively rare among all cases of maltreatment. We have studies that have used official reports such as the National Study of Child and Adolescent Well-Being (OPRE 2013), but these do not capture unreported cases. NIS asks questions about unreported cases, but we do not know what happens to them in the long term (Sedlak et al. 2010). Only two studies could be found that compared children with self-reported maltreatment (in adolescence) to officially reported cases. Those children identified by both official and self-report methods generally experienced more incidences of maltreatment and had worse mental health outcomes (Cohen et al. 2001; Shaffer et al. 2008). Cohen and colleagues found higher rates of poor mental health among children with official reports compared to self-report or no reports. It should be noted that the Shaffer and colleagues (2008) study appeared to have drawn their sample from a higher-risk population. It is also possible that some of the children who remain unreported are served in other ways or find other resources.

Services

What Services Are for Whom?

There is no current widespread triage system that provides for families with different levels of need, outside of determining whether a child can safely remain within the home. In other words, for the 95 % of children who remain at home, we have no systematic way of understanding which families will need just a brief connection to services and will be adequately treated and which families will need support, possibly for many years. As aforementioned, many surveys of families involved at some level with child welfare reporting want more services rather than less. But how much is enough? Several calls have been made for matching long-term intervention with apparent long-term or multiple problem families (Munro 2011). But appropriate targeting of resources at the time a family first comes to the attention of child protection is still elusive. Length of services tends to be arbitrarily set by policies related to funding or some other time constraint, rather than by when we anticipate seeing a sufficiently powerful positive effect.

It is also often unclear whether child protection should focus on the child’s needs, the parents’ needs, and the community’s capacity to support the child and the child’s family or some combination of these. If the services are tied to child well-being, then service determination is not dependent only on the parents’ needs or actions. So for example, a child who needs a developmental assessment would receive access to that even if the investigator did not see a risk of maltreatment because the goal is to promote child well-being. When there is a suspicion of maltreatment, the presumed assumption is that there is a need for parenting training and that this will improve the situation. While there are evidence-based and promising parenting programs, these are typically not delivered by child welfare, although some exceptions exist (e.g., Chaffin et al. 2012). It is also not clear that this is the best or sufficient response given the many other family level risk factors associated with maltreatment (Jonson-Reid and Drake 2008). Still others argue for a social capital or socioeconomic approach targeting poverty rather than maltreatment per se (Blackstock and Trocme 2005). Different foci dictate different forms of measuring both services and outcomes.

What Would Have Happened Without Services?

Since we do not randomly assign families to services within the CPS context, it can be difficult to assess and easy to draw conclusions based on inadequate information. We tend to study etiology without consideration of services (Jonson-Reid 2004), and since services are rare, we are unlikely to accidentally see impact (Jonson-Reid 2011). As illustrated in the figure below, there are several steps to think through in improving our understanding of CPS services. First, there must be appropriate assessment of need. Earlier, we reviewed the current status of the literature about substantiation and discussed the notion that this particular metric is not ideal in terms of understanding whether or not a child is at risk of harm. A substantial literature exists regarding risk assessment in child protection, and early doubts as to the effectiveness of risk assessment tools (Wald and Woolverton 1990) have yet to be satisfactorily answered. It may instead be more useful to move away from attempting to predict an act of maltreatment, per se, and consider instead what the threats to safety and well-being are and how they are to be best served (Drake and Jonson-Reid 1999; Davies et al. 2009). This, however, is an empirical question requiring that services be available to make the evaluation of such an assessment worthwhile. It also suggests the need for a certain level of training to be present among those making such assessments. Currently, the degree of training required and the amount of training actually provided vary substantially by region, and few studies control for the level of preparation of the child protection staff. This is a little researched area, but there are theoretical and some empirical reasons to believe this has an important effect (Munro 2005; Ryan et al. 2006; Strolin et al. 2006). In the future, it will be important to better understand the degree to which the expertise of child protective services personnel plays a role in both appropriate decision making and improved case outcomes.

Assuming an adequate assessment process is in place, understanding outcomes means understanding who is providing what, and how well. If the need is met by CPS, then the family need only complies with CPS and the outcome is directly linked to the CPS intervention. If CPS is the case manager, however, the family must first agree with the referrals provided, then access the services, and then complete their engagement with the services in order to anticipate a positive outcome. Further, the services outside the CPS system must be examined in terms of quality since complying with a service that is ineffective is unlikely to result in positive change.

figure a

The Promise of Big Data

The general issue of reporting should be understood in a broader context. “No single professional can have a full picture of the child’s needs and circumstances and, if children and families are to receive the right help at the right time, everyone who comes into contact with them has a role to play in identifying concerns, sharing information and taking prompt action” (Her Majesty’s Government 2013, p. 8). Linkage of agency data systems and administrative data to survey data may hold particular promise in advancing our understanding of how to improve our responses to and prevention of maltreatment. This requires careful attention not only to the construction of linked data across public and private agencies but also to creating an ongoing feedback loop between data and policy (Jonson-Reid and Drake 2008). This will require that data systems include variables that are relevant to the logic models developed specific to the form of detection and CPS response in a given country or region.

We have already arrived at a place where technology exists to have some child welfare contacts be initiated automatically, without a report at all. One recent study (Putnam-Hornstein and Needell 2011) found that for a very small subpopulation with a large number of risk factors noted on their birth record, the likelihood of a child maltreatment report within the first 5 years of life was 89.5 %. It is an open question as to whether child welfare should consider taking a truly proactive stance and offering voluntary preventative services on the basis of such data. Such action might seem highly desirable or even inevitable if child protection is viewed as a public health issue. On the other hand, those persons persisting (wrongly we believe) in regard to child protective services as a punitive system might see such preemptive action as a violation of parental rights. As data systems are cross-linked, our ability to predict who will be referred in advance will increase, and if we are serious about a preventative role for CPS, there should be a discussion as to if and when such data should be used for voluntary service provision. Indeed, some states are currently triggering child protective assessments based solely on administrative data. Under the “Birth Match” program (Shaw et al. 2013), newborn children born to parents with prior terminations of parental rights are assessed without a report ever being filed. It may well be that in the future, we are able to use existing data to provide a level of accurate targeting for voluntary preventative services which is simply not currently possible.

Conclusion

In our view, many debates over the advisability of mandated reporting laws are clouded by two key problems. One has to do with the failure of the empirical literature to adequately undergird the current debate, and the second has to do with a failure to place mandated reporting laws within the broader framework of child protective services.

In some quarters, there is a long-standing and established “conventional wisdom” about the adverse consequences of mandated reporting. Unfortunately, much of it is contradicted by empirical data and properly nuanced, more precise, and more thorough analysis of concepts and issues. These contradictions range from simple misunderstandings (e.g., that unsubstantiated cases are equivalent to unnecessary reports) to inaccurate characterizations of the current system (e.g., that child protective services are overwhelmed by their investigatory responsibilities) to broader misconceptions about the system in general. In order for the mandated reporting debate to move forward and for improvements to be made to child protection systems, all relevant discussions must be far more evidence based. We believe that most policy makers would prefer engaging in evidence-based policy making to the alternatives and mandated reporting policy is one area in which the evidence is strong enough to support such an approach.

Despite a long history of child protection services, we know surprisingly little about the outcomes for children and families who receive the most common and least intensive forms of services. We also know relatively little about children who receive investigations and no further services. This is undoubtedly partly true to methodological difficulties, including difficulties in establishing control groups. Clearly, the need for services as indicated by children reported far outstrips the current system’s ability to intervene. Without further evidence regarding what can be done within formal child protection, what can be done in conjunction with other agencies, and what can be done within the community, child protection practice will remain a disjointed system with a great deal of room for improvement.