While mandated reporting was generated as a key prevention strategy for child abuse or neglect nearly 50 years ago, it is not widely recognized that reporting does not represent a primary prevention approach. In order to report maltreatment, maltreatment must be thought to have occurred. Secondary prevention (intervening in an occurrence) and tertiary prevention (rehabilitation) are the direct benefits of reporting. Mandatory reporting should have a strong benefit if there is a high likelihood of recurrence among maltreated children; if it increases the sensitivity of the diagnosis of maltreatment, finding more children who are maltreated; and if it leads to the improvement in the safety or outcomes of children. Mandated reporting of child maltreatment depends upon the recognition of the maltreatment by designated professionals in all states and by all citizens in some states. While state-mandated reporting has a more than 40-year history in the United States, child maltreatment remains a serious public health problem. Because of the limitations in science, the impossibility of any clinical trials of mandatory reporting, and the constant evolution of intervention approaches, clearly establishing which children are better off for having been reported is virtually impossible. However, in contrast to available evidence from other countries, the United States has seen a remarkable reduction in child abuse over 20 years (Finkelhor et al. 2014).

The harms resulting from child abuse and neglect and our understanding of the risks and benefits of intervention have become clearer in the years since the 1993 National Research Council report Understanding Child Abuse and Neglect (National Research Council 1993). The harms of death and disability described in the 1962 Battered Child Syndrome article in JAMA (Kempe et al. 1962) were described through cross-sectional studies. Better data on the harms of maltreatment came from cross-sectional and retrospective studies including the Adverse Childhood Experience or ACE studies (Anda et al. 2006). Only relatively recently have risk and protection from the harms of maltreatment been examined in longitudinal studies, with clear attention to temporal relationships and the proportion of outcome variance explained by the different type, chronicity, and severity of maltreatment experiences (National Research Council 2013; English et al. 2005; Barth et al. 2002). Child maltreatment has been incorporated into current thinking about toxic stress with associated brain changes¹ and directly linked to increased risky behaviors for sexually transmitted diseases (Jones et al. 2010), depression, aggression, substance abuse, and poor parenting.

In the LONGSCAN longitudinal study of children maltreated early in life or considered to be at risk for maltreatment, the majority of reported maltreatment occurred in early childhood (Proctor et al. 2012). This observation contrasts with parent self-reports of abusive or disciplinary behavior; children in the 5–9-year-old age group are more likely to actually experience physically abusive behaviors from their parents, whereas older children are more likely to be sexually abused (Theodore et al. 2005). Several studies have documented biases in child maltreatment reporting that appears to be associated with race and income (Laskey et al. 2012; Hampton and Newberger 1985) although child abuse and neglect are not classless or immune to income issues (Pelton 1978). Another important set of biases relate to the severity of child maltreatment reported by type of maltreatment. In the LONGSCAN studies, a set of severity criteria were applied that ranged from no visible injury to life-threatening injury or permanent disability for each form of maltreatment. Over 50 % of all maltreatment reports, except in the situation of sexual abuse, were classified as being characterized by the lowest forms of severity although 47 % of sexual abuse was classified as high severity as the definition of high severity included penetration (Litrownik et al. 2005).

Other systematic errors occur in the classification of type of maltreatment. Psychological or emotional maltreatment is rarely the official determination but is actually quite frequent. When child maltreatment allegations among children involved in the LONGSCAN study were reviewed and coded for type, severity, and chronicity of maltreatment by trained reviewers using specified research criteria for type of maltreatment, there is almost a tenfold increase in numbers of children who were classified as victims of psychological maltreatment from about 4 % of allegations to nearly 40 % of allegations (Runyan et al. 2005). Even the process of substantiation is an issue. Children with unsubstantiated allegations have outcomes that look more like the children with substantiated reports of child abuse than children who have not been reported (Husey et al. 2005). Unsubstantiation is clearly much more complicated than just that maltreatment did not occur.

In the past it may have been possible to dismiss societal intervention as there has been little evidence for effectiveness. However, intervention in child maltreatment has become more nuanced, and there have been major strides in evidence for effective mental health intervention. We have learned a great deal about foster care and other interventions. Rates of foster care placement are falling (Child Trends Databank 2014), the use of congregate care is declining (The Annie E. Casey Foundation), and changes in policy include adoption of multiple or differential response (Kyte et al. 2013) and family-group decision making (National Center on Family Group Decision Making) by social service agencies. Evidence-based treatments are being adopted to address the traumas that the children have experienced (Aarons et al. 2011) and to maximize the quality of interventions such as Fostering Health Futures