A six-year-old girl and her older sister are brought by their father to be evaluated for a history of cough, runny nose, and low-grade fever. In addition to signs of a cold, the physician notes that the girl’s nasal bridge is quite swollen and bruised. When asked what happened, she innocently shrugs her shoulders, and her father’s only conjecture is that since she sleepwalks she might have bumped into something. The father sits impatiently, but as questioning progresses becomes increasingly defensive, at one point angrily declaring “we don’t beat our kids, if that’s what you’re asking.” Further complicating the situation is information from several nurses that this family is “on the brink” both socially and financially, and that additional stress is likely “to blow this family apart.”
The term child abuse encompasses physical abuse, sexual abuse, psychological abuse, and neglect –though the phrase “child abuse and neglect” is also common parlance. In its typical usage, child abuse refers to actions (or failures to act) by a parent or caregiver that result in serious physical or emotional harm, sexual abuse or exploitation, or imminent risk of serious harm.
What could be more simple than the ethics of child abuse? For those who commit it, don’t. For everyone else, do what you can to protect children from it. As one looks more deeply, though, definitions, interpretations, conflicting responsibilities, and, most prominently, uncertainty (on a range of issues) raise difficult questions. What exactly counts as abuse? How should we understand reasonable suspicion (which serves as the trigger for mandated reporting)? How sure must we be that abuse has occurred before initiating a child abuse investigation? Knowing that biases are inevitable in our assessments of risk and probability, how can we treat families fairly in our efforts to protect children from abuse? Finally, what should we as mandated reporters do when we do not think that reporting abuse is in a given child’s best interest?
In addressing these questions, there are several important things to understand about child abuse. Firstly, it is prevalent. In one Canadian study of over 10 000 households, 21% of women and 31% of men reported having been physically abused as children (MacMillan et al., 1997). In the USA in 2004 there were over 3.4 million investigations for suspected child abuse, with 872 088 confirmed cases, and at least 1490 deaths (Gaudiosi, 2006). Moreover, there is reason to believe that these numbers significantly underestimate the true incidence of abuse (Finkelhor, 1990; Herman-Giddens et al., 1999; Crume et al., 2002). Secondly, it is ubiquitous. One finds child abuse occurring in every community and at all levels of society (Finkelhor, 1994; Wyatt et al., 1999; Lampe, 2002; Lalor, 2004; Daro, 2006; Gaudiosi, 2006), though a variety of risk factors do make abuse more likely. Those at increased risk include children who are younger, acutely ill, have chronic medical conditions and/or behavioral disorders, or have low intelligence (Warner and Hansen, 1994; Levitzky and Cooper, 2000; Gaudiosi, 2006). Family characteristics that predispose to abuse include increased stress, marital conflict, a young and/or single parent, unwanted pregnancy, and poverty (Warner and Hansen, 1994; Kotch et al., 1995; Brown et al., 1998; Drake and Zuravin, 1998; Overpeck et al., 1998). Thirdly, it has significant sequelae: bruises, lacerations, sexually transmitted diseases, pregnancy, post-traumatic stress disorder, chronic somatic disorders, serious brain injury, plus acute and chronic medical care (Irazuzta et al., 1997; Emery and Laumann-Billings, 1998; Widom, 1999; Discala et al., 2000; MacMillan et al., 2001; Diaz et al., 2002; Fein et al., 2002; Lansford et al., 2002; Scheid, 2003). Recent estimates for the USA alone put the total direct costs related to child abuse at more than $24 billion annually, with indirect costs exceeding $64 billion (Fromm, 2001).
There are significant interpersonal and cultural variations in terms of what counts as child abuse (Hansen, 1997; Dubowitz et al., 1998; FitzSimmons et al., 1998; Daro, 2006). An illustration of this is recounted by Dr. Catherine DeAngelis, editor of the Journal of the American Medical Association, from her years teaching police officers about child abuse: “I started each new class by asking how many had ever spanked a child; almost all hands were raised. I then asked how many had ever beaten a child; no hand was ever raised. I then asked them to explain the difference, and the fun began” (Fargason et al., 1996).
At root, DeAngelis’ question is about harm and proportionality. In meting out punishment, how much harm is acceptable before the threshold into abuse has been crossed? Research tells us that our answers are heavily influenced by individual attitudes regarding discipline, as well as our personal experiences with corporal punishment (Howe et al., 1988; Hansen, 1997; Bonardi, 2000; Jankowski and Martin, 2003; Tirosh et al., 2003). So, too, significant cultural norms come into play (Daro, 2006). Some communities hold that beating children is a sign of love and commitment (Visser and Miller, 2002; Rakundo, 2006) whereas others (e.g., almost half of European countries) have outlawed even routine spanking (Daro, 2006). At times, it will be difficult to identify precise cutoffs, and inevitably reasonable people will disagree whether a given practice qualifies as legitimate discipline or abuse. But no reasonable person would dispute the notion that an adult who non-accidentally inflicts serious harm on a child commits an act of abuse (Chen, 2004; Maiter and Alaggia, 2004; Daro, 2006).
Though less central than harm, intention also figures into what we consider abuse. A parent can bruise their child by accident, or out of anger, or by pulling them out of harm’s way, or even by administering non-traditional medical therapy. In each case, the parent’s intention influences not only the nature of our response but also our judgement whether their action qualifies as abuse. Again, there are cultural components to this. Certain practices (e.g., coining, circumcision, scarification, and other body manipulations) are accepted because, within their cultural context, they are not intended to harm. But good intentions do not render any practice immune from being judged abuse. Children have the right to an open future (Feinberg, 1980) as well as the right to protection from serious harm (Archard, 2002). “Well-intended” culturally bound practices that violate these rights, and in so doing cause significant harm, do constitute abuse. Examples include severe shaking of an infant to raise a sunken fontanelle (i.e., caida de Mollera) or cutting off a girl’s clitoris and labia, then sewing her vagina shut (i.e., clitoridectomy and infibulation). The intention of these acts may mitigate our reactions to the adults who carry them out. But the lasting injuries and impairment and psychological damage that befall children who are subjected to such acts are testament to the abusive nature of these practices (Barstow, 1999; Chalmers and Hashi, 2000; el-Defrawi et al., 2001; Refaat et al., 2001; Whitehorn et al., 2002; Nour, 2003; K. M. Yount and D. L. Balk, unpublished data). The growing criticism from within cultures that engage in these practices simply reinforces the judgement that they are indeed abusive (Adinma, 1999; Eke and Nkanginieme, 1999; Nkwo and Onah, 2001; Msuya et al., 2002; Nour, 2003).
Separate from concerns about intention or even thresholds for harm are questions about what should count as cause to suspect child abuse and when is it appropriate to actually report suspected abuse to child protective services (CPS) agencies. According to a recent report, 82% of countries sanction voluntary reporting of suspected child abuse, while 65% require reporting from at least certain individuals (Daro, 2006). In the USA, CPS agencies screen approximately 60 000 reports of alleged abuse per week, investigating two-thirds of them (Gaudiosi, 2006). Most reports are initiated by individuals who qualify as mandated reporters in that their professional work brings them into routine contact with children. At a minimum, this includes teachers, law-enforcement personnel, firefighters, healthcare professionals, social services providers, and child care workers – though in 18 US states any competent adult qualifies as a mandated reporter (National Clearinghouse on Child Abuse and Neglect Information, 2001, 2003). Of note, countries that have mandatory reporting have a significantly lower mortality rate for children under five years of age (−0.46; p < 0.0001; Daro, 2006).
As mentioned above, there are known risk factors for abuse. So, too, there is an extensive medical and social science literature that documents clinical indicators of abuse (Kempe et al., 1962; Warner and Hansen, 1994; Reece and Ludwig, 2001; Giardino and Giardino, 2002; Vasquez and Pitts, 2006). At times, these resources combined with good interviewing skills and clinical acumen will cause mandated reporters to be certain in their judgement that abuse has occurred. Still, we will be less than certain in most instances. Hence our judgement and eventual decision to report will of necessity be grounded in a calculus, whose cofactors will include physical findings, observed behavior, risk factors, and so on. Though such a calculus is itself unavoidable, several ethical issues arise in how it is utilized.
The first has to do with bias, in terms of whom we suspect and report, and why. We know from research that mandated reporters are more likely to suspect and report children whose ethnic and socioeconomic profiles do not resemble their own (Hampton and Newberger, 1985; Brosig and Kalichman, 1992; Bonardi, 2000). Perhaps we presume that people like ourselves are not likely to have abused a child. But it is interesting that ethnic and socioeconomic resemblance seem to make us more prone to identify with a child’s caregiver, exonerating them from suspicions of abuse, rather than make us more protective of the child who has been harmed (Adler, 1995). However this bias is to be explained, it is a problem of justice that minorities are several times more likely to be reported and investigated for suspected abuse (Sinal et al., 1997; Lane et al., 2002).
Were there little downside to being reported and investigated, this might be of minor concern. However, reports and investigations of child abuse are events that can destroy families and careers (Renke 1999). The ideal, of course, is that CPS agencies will carefully and sensitively investigate reports of child abuse, coordinate with social services and law enforcement to rehabilitate or remove offenders, and work with at-risk families to stabilize the home environment. But the reality is that families are stressed and disrupted – and sometimes blown apart – by the very process of state intervention into their lives (Thompson-Cooper et al., 1993; Beck and Coloff, 1995; Richman, 2000). Moreover, because the CPS system is fundamentally underfunded and overburdened, support services frequently do not materialize, leaving families no better off than before (Murphy-Berman, 1994; Melton et al., 1995; Melton, 2005; Gaudiosi, 2006). For these reasons, significant numbers of mandated reporters who have had experience with CPS agencies are wary of reporting further cases of suspected abuse (Applebaum, 1999; Flaherty et al., 2002; Melton, 2002; Flaherty et al., 2004).
None of this is to say that reporting suspected child abuse is, on balance, wrong to do. Rather it is to acknowledge that as currently configured the system is not without risks – which must be entered into our calculus of whether to report.
A second and related ethical issue has to do with the threshold that has been set to trigger mandated reporting, which varies from country to country, and even region to region. Across the USA, for example, there are 11 distinct thresholds; five use some variant of belief, while six use some variant of suspicion (National Clearinghouse on Child Abuse and Neglect Information, 2003). From a conceptual standpoint, there are important differences between believing and suspecting (White, 1993; Levi and Loeben, 2004), and various statutes have been changed in recognition of this (National Clearinghouse on Child Abuse and Neglect Information, 2001). In an nutshell, the problem with “belief” is that it involves holding an idea to be true, whereas in the context of mandated reporting one is seldom sure that abuse occurred but instead concerned that it might have occurred.
Collectively, the various statutory thresholds are referred to by the legal penumbra of reasonable suspicion. As such, mandated reporters typically are instructed that reporting is required any time they have reasonable suspicion that a child has been abused (Myers, 2001). The problem, however, is that there is no consensus what reasonable suspicion means (Deisz et al., 1996; Kalichman, 1999; Levi and Loeben, 2004; Levi and Brown, 2005; Levi et al., 2006). For some, it means whenever the thought of child abuse goes through your head, even if it goes right out (Deisz et al., 1996). For others, it requires a substantial likelihood that abuse actually occurred. For example, one survey of over 2000 pediatricians found that 15% of respondents indicated that abuse would need to be over 75% likely to qualify as reasonable suspicion; while a quarter of respondents set the threshold at a 60–70% likelihood; another quarter set it at 40–50%, and 35% of respondents set the threshold as low as 10–35% probability (Levi and Brown, 2005).
Surprisingly, the professional literature on child abuse provides no substantive clarification on this (Deisz et al., 1996; Kalichman, 1999; Levi and Loeben, 2004), nor does the law (Myers, 2001; Levi and Loeben, 2004). Mandated reporters are “left to define their own personal standards for what constitutes a reasonable suspicion of child abuse” (Kalichman, 1999): the result being an ad hoc system that ensures neither equal protection for children nor justice for those who are reported. In addition, this absence of a standard can create significant burdens for conscientious mandated reporters trying to determine the right thing to do (Flaherty et al., 2004). So, too, it can foster conceptual confusion, as was shown in one study where physicians interpreted reasonable suspicion differently for severe versus minor injuries (Levi et al., 2006). Though perhaps clinically understandable, this is conceptually problematic because the question is not whether physicians should have a heightened level of suspicion when the stakes are higher, it is whether they are clear on the concept itself. By way of analogy, it may be more important to look for fever in an obtunded patient than in someone with an itchy rash; but what constitutes “fever” (i.e., the properties that make up one’s conceptual understanding of fever) should not vary with the clinical situation. Moreover, because the vast majority of child abuse cases do not involve severe injury (and prior severity of injury is not predictive of subsequent severity; Levy et al., 1995), to safeguard children it is important that mandated reporters regard all instances of harm with the same level of careful consideration.
From a systems standpoint, the absence of a standard is equally troubling. Imagine that police were directed to write speeding tickets for motorists driving “too fast” but given no clear guidelines for judging what should count as “too fast.” In the case of mandated reporting, the result is a system of indiscriminate reporting (Blacker, 1998) that not only disrupts families in which no abuse has occurred, and misses cases that warrant investigation, but also diminishes the effectiveness of CPS by dispersing already scarce resources and by eroding confidence in the legitimacy of child abuse investigations (Applebaum, 1999; Flaherty et al., 2000; Richman, 2000; Melton, 2005).
Preliminary evidence suggests that statutory wording can significantly influence how individuals interpret and apply the threshold for mandated reporting. Two studies in particular have shown that mandated reporters express significantly greater willingness to report abuse when the threshold is defined as there being a 25% chance or greater that abuse occurred (Blacker, 1998; Flieger, 1998). This is not to say that “25% probability” is the appropriate threshold to endorse. But it does demonstrate that specifying the threshold in more concrete terms makes a significant difference.
A third ethical issue in our calculus of when to report abuse is where society should set the threshold for mandated reporting. Setting the threshold too low – say, 10% estimated probability – not only would bring about considerable disruption to families where no abuse has occurred but also would greatly stretch already limited CPS resources. Moreover, it would potentially overload the legal system, strain relationships between parents and mandated reporters, and (if reporting requirements are seen as unreasonable) increase disrespect for the law (Applebaum, 1999; Melton, 2005). These implications are further compounded when mandated reporters have immunity from criminal and civil prosecution, as occurs in the USA (State of Minnesota v. Curtis Lowell Grover, 1989). In point of fact, mandated reporting systems often provide little check or balance of mandated reporters’ power, little recourse for non-malicious reporting injustices, and no ready mechanism for constructive feedback to educate mandated reporters (Thompson-Cooper et al., 1993; Kalichman, 1999). So, setting an extremely low threshold risks indiscriminate reporting with little prospect for amelioration.
By contrast, setting the threshold for mandated reporting too high – say, 75% estimated probability – risks overlooking children who have been abused, since the signs of child abuse are often ambiguous (Giardino and Giardino, 2002). Consequently, some balancing is in order to identify an appropriate threshold. This, in turn, will require deciding how much we are willing to invest to protect children from abuse, as well as how much harm and what kinds of harm we are willing to tolerate. For this, we will need public dialogue and debate, as well as a much better understanding of the costs and benefits of various cutoff points: 25% estimated probability versus 35%, 50%, and so on.
A temptation to be resisted is the construction of different thresholds for different kinds of abuse, depending on the severity of risks in play. The reason is that there is simply too great a variability in education and expertise among the millions of mandated reporters to expect individuals to discern accurately which threshold ought to apply for one kind of abuse versus another.
Even if there was a well-defined threshold, however, there is a fourth ethical issue in our calculus of when to report abuse. What should we do when we have reasonable suspicion that a child has been abused, but do not think that their interests are best served by reporting the abuse? From a legal standpoint, the answer is clear: mandated reporters are required to report whenever they have reasonable suspicion of abuse and the suspected abuser is either a parent or a person responsible for the child’s welfare. In many jurisdictions (e.g., throughout the USA), failure to report makes one guilty of a misdemeanor (punishable by a fine and up to several months in jail) and civilly liable for damages if the abused child (or another child) is further victimized because of the failure to report.
Despite this, large numbers of mandated reporters regularly do not report suspected abuse (Singley 1998; Kalichman 1999; Delaronde et al., 2000; Flaherty and Sege, 2005), though prosecutions for this are rare (Singley 1998; State of Missouri v. Leslie A. Brown, 2004). Reasons that mandated reporters do not report suspected abuse include the many ambiguities and uncertainties discussed above, as well as competing interests that mandated reporters often experience – such as worries about their relationship with a family, costs (financial, social, professional), and so on. In addition, the decision whether to report suspected abuse weighs heavily on many mandated reporters precisely because reporting does not always benefit the child (Johnson, 1999; Flaherty et al., 2000, 2004). Relatedly, some competing interests can be intricately intertwined with the interests of a given child. For example, we know that child abuse is present in 50–80% of families in which domestic violence occurs (Garbarino et al., 1991; Appel and Holden, 1998; Edleson, 1999), but it is not at all clear that it is in the interests of all (or even most) children exposed to domestic violence to be reported for suspected child abuse just because of the known association between the two. Another, and perhaps more problematic, example of intertwined interests involves instances where a parent who has committed child abuse is in therapy. Here, the concern is that reporting parents’ abusive behavior could impede not only their own rehabilitation, but also (if such reporting were standard practice) prevent other child abusers from coming forward for help (Berlin et al., 1991; Budai, 1996).
Putting such twists aside, the question remains whether a mandated reporter should follow the law and report suspected abuse when doing so does not appear to be in a child’s interest. At root, it is a matter of conflicting obligations: obligation to follow the law versus obligation to protect children from harm. What makes the matter particularly difficult is that, unlike many laws or countervailing ethical principles such as patient confidentiality (see Ch. 7), mandated reporting laws were constructed with the protection and well-being of children specifically in mind. Hence, the tension is between following a rule specifically designed to protect children from abuse and following one’s own judgement about how best to ensure a child’s safety and well-being.
I think it is possible ethically to justify such acts of conscientious refusal (Rawls, 1971) that are grounded in one’s professional principles and responsibilities. But to do so certain conditions must be met: (i) you genuinely believe that reporting the suspected abuse will result in a net harm for this child; (ii) you are confident that the child is not at risk for subsequent injury, and you are willing to take responsibility for their safety; (iii) all other law-abiding alternatives would (in your estimation) also conduce to significant harm; and (iv) you are prepared to defend your decision publicly, and if need be accept the legal penalties for not carrying out your responsibilities as a mandated reporter.
What these conditions reflect is the strength of conviction necessary for true conscientious refusal of mandated reporting. In weighing one’s resolve, however, one must be careful of overconfidence in predicting either a child’s safety or one’s ability to intervene on their behalf (Adler, 1995). However imperfect, CPS agencies provide the only systematic approach for investigating and safeguarding a child’s well-being.
In the case presented at the outset, a careful physical examination revealed no other injuries, and a thorough review disclosed no evidence of prior suspicious injuries, frequent visits to the emergency room, or bleeding abnormalities. While the evidence does not point to abuse as the most likely explanation for the injury, it is not ruled out. In this case, the default decision must be to report suspected abuse, unless the physician has a strong relationship with the family and can meet the four criteria for conscientious refusal mentioned above.