Covert Video Recordings of Life-threatening Child Abuse: Lessons for Child Protection

Findings of Surveillance

We have shown that CVS can readily identify parents who intentionally suffocate their children and thereby induce or fabricate illness, a behavior called Munchausen syndrome by proxy (MSBP).²⁴ We have also shown that parents may abuse their children in other ways. Thus we avoid the term MSBP because it does not describe adequately the range of abuse that occurred; we prefer to describe the actual abuse identified. Although attention-seeking through induced illness (ALTE, poisoning, or failure to thrive) was always a feature of the abuse in our patients, there were frequent additional abuses. CVS identified other serious physical and emotional harm, reflecting a severely dysfunctional relationship between parent and child (Tables 3 and4) that would not otherwise have been brought to the attention of health or social services professionals.

Abuse was inflicted without provocation and with premeditation and, in some instances, involved elaborate and plausible lies to explain consequences. We wish to emphasise the plausibility of many explanations and histories given by such parents. For example, one mother (case 13) claimed that she had suffocated her son because of stress related to his crying and continually waking her from sleep. It is possible to empathize with this kind of stress. However, under surveillance, the mother was seen, with premeditated planning, to suffocate her infant when he was deeply asleep. The majority of other cases showed attempted suffocation when the child was asleep or lying passively on the bed. Children did not appear to provoke their parents into abusing them.

In some children, abuse was inflicted frequently, if not almost continuously (see “Logs of Selected Observations,” below). That it was detected in the caring and child-centered environment of a hospital raises the possibility that even more severe abuse may have been inflicted at home. Defense witnesses have testified in court that the stress of being in a hospital may have led some parents to abuse their children. We do not accept this argument. Moreover, the median duration of CVS was only 29 hours. In our opinion, all of the abuse detected would have occurred regardless of whether CVS was being implemented.

The sadistic nature of the abuse inflicted in a number of our patients is reminiscent of other child abuse in which brutal injuries, such as cigarette burns or multiple fractures of different ages, are pathognomonic. It is difficult, for example, to accept that placing a burning cigarette on a child's body could represent a legitimate form of punishment or an impulsive act of anger. Among possible explanations for this behavior and that seen during CVS is that it gives some form of gratification to the perpetrator or that it is a learned behavior from the parent's own childhood. Although the origins of these abusive behaviors may be genetic or may result from earlier traumatic experiences during the abuser's childhood, or both, the mechanisms whereby such factors are translated into abusive behaviors in the next generation remain poorly understood.

Another issue raised by these observations is the difficulty many health professionals, relatives, and others may have in accepting the possibility that parents would deliberately abuse a child in ways revealed under surveillance. Professionals and members of the judiciary, who have not seen abuse in action as shown by CVS, may be almost unable to acknowledge that such acts can be and are committed by apparently caring mothers, fathers, and stepparents.

How Can Abuse Be Identified as a Cause of ALTE?

In 1986, National Institutes of Health defined ALTE as “an episode that is frightening to the observer and that is characterized by some combination of apnea (central or occasionally obstructive), color change (usually cyanotic or pallid), marked change in muscle tone, choking or gagging.”¹² ALTE is a combination of symtoms and signs arising from a number of mechanisms.

Most of our patients presented with recurrent near-death events requiring CPR; as a group, these patients are known to have a very high risk of sudden death.²⁵ In our experience,¹⁴such events are rare in otherwise healthy infants and children and are attributable to three primary mechanisms: 1) a respiratory disorder, 2) epilepsy,²⁶ and 3) child abuse. Respiratory disorders have occurred predominately in the following settings: 1) in preterm babies in whom some ALTE may be considered a form of apnea of prematurity; 2) during respiratory infections²⁷ such as bronchiolitis attributable to respiratory syncytial virus; 3) as a severe form of cyanotic breath-holding²⁸; and 4) in response to severe gastroesophageal reflux.²⁹ All of these may involve sudden intrapulmonary shunting.³⁰

Child abuse as a cause of ALTE is the most difficult to diagnose. In our patients, bleeding from the nose or mouth in association with an ALTE distinguished intentional suffocation from ALTE attributable to natural causes. Having a sibling die suddenly and unexpectedly also occurred significantly more frequently in the abused group. Petechial hemorrhages of the face were identified in three patients (as described previously by Meadow in 5 of 27 cases of intentional suffocation),³¹ but were not found in the comparison group. Preterm birth was more common in infants suffering ALTE resulting from a natural respiratory mechanism, and the onset of ALTE resulting from natural causes occurred at an earlier age. Finally, the increased white blood cell counts, usually with a lymphocytosis, determined after some suffocatory events may mislead doctors into suspecting infection as the cause.

The background of parents who abused their children revealed important indicators of abuse as potential causes. Our earlier reports^9,10 showed that all 14 parents who had abused their children had personality disorders. No systematic psychiatric assessment was performed on the latter 25 patients in this report, and only 9 of these were described as having a personality disorder after individual psychiatric consultations. Two parents refused assessment. There was variability in the quality of background information available, and some psychiatrists were not willing to provide us with their reports.

Although there is a growing recognition of the comorbidity of psychiatric disorders (especially depression, anxiety, and substance abuse) among parents who abuse children,^32-34 these were not found in our patients. No parent had a psychotic mental illness. It is possible that in some patients, the diagnosis of personality disorder lead professionals to give less attention to associated psychiatric symptoms, given the known difficulties of successful therapeutic engagement. It is also possible that some of these parents were themselves resistant to psychological intervention and focused instead on physical symptoms in themselves (somatization) or on inducing or fabricating illnesses in their children.³⁵Sometimes overt Munchausen's syndrome was present in the parent (Tables 2 and 4). Many parents spent much of their time speaking about how badly life had been for them, using fabrication to invoke sympathy from listeners.

It is beyond the scope of this report to review the conceptual and therapeutic controversies about personality disorder or the difficulties in defining it. However, despite valid concerns about labeling and measuring this heterogeneous group of disorders, personality disorder has been described repeatedly in anecdotal, clinical, and psychodiagnostic studies of parents who abuse children.^35-39 In a controlled study involving a structured interview of 54 mothers who abused their children, Famularo et al³³ found an increased incidence of personality disorder. Furthermore, personality disorder has been reported to have a predictive value for emotional and behavioral problems in children.³⁴

Psychiatrists and psychologists have researched ways of identifying and classifying personality disorders. Techniques, including validated questionnaires, are available, although the more intrusive the initial investigation, the more elusive the response of the abusing parent may be. There are a number of different personality disorders, and additional research is needed to define more effectively those most likely to lead to the abuse of children, particularly the sadistic and pervasive types described here.

To predict the risk of child abuse, lists of risk factors have been developed. Although they may be of limited predictive value,⁴⁰ they can form a basis to develop risk-assessment strategies. The data in Table 2 present some of the features we identified along with their frequency of occurrence. Refined techniques to identify persons who may inflict severe abuse may also be useful to screen others who have close contact with children, including foster parents, babysitters and carers, nurses, doctors, social workers, humanitarian aid workers, and teachers.

Finally, in a proportion of families, the mother reported major difficulties in her relationship with her partner/husband. Some partners were passive and alleged to be relatively uninvolved with their families, giving little support to the mother. Others allegedly were abusive to the mother. In some families, the father/partner was insecure in his commitment to the mother; we considered that the mother may have harmed the child as a way of maintaining the relationship. For example, on the evening that her partner walked out of the home claiming he was leaving her, one mother (of patient 36, and already under investigation after 14 apneic/cyanotic episodes), presented her child to the hospital with a series of three ALTEs, one of which had features on a physiologic event recording suggestive of upper airway obstruction. The mother of patient 37 alleged that her partner had threatened to leave her if she continued to care for her child (Table4).

In addition to sudden death (as described above), abusive incidents occurred in 15 siblings (Tables 2 and5). Such events should also highlight abuse as a possible/probable cause for ALTE.

Additional features of risk identified by others and seen in some of our cases include 1) the dramatic nature of the presentation, with parents implying or claiming to have saved the infant's life; 2) a tendency for events to occur when the parent is awake and when they are stressed; 3) absence of an independent observer at the onset of events. In this feature, the key words are independent andonset. A proportion of parents who suffocated their children provided stories of others witnessing some part of the events. Usually, observers had not seen the events begin; they were often called to help resuscitate the infant. Perhaps not surprisingly, some witnesses were friends or relatives who had been persuaded by the parent that the event was typical. However, almost all healthy babies stop breathing when they are asleep,⁴¹ and a number appear to turn blue around the mouth when crying.⁴² Such normal phenomena can be used by abusing parents as alibis for abuse; 4) cessation of events when the child is separated from the parent. However, in three patients not included in this study on whom we were unable to carry out CVS and in whom we used separation as a diagnostic test, expert witnesses argued that the children had suddenly grown out of their problem.

These findings support our earlier results^9,10,14 and suggest new strategies for managing infants presenting with ALTE. The presence of bleeding from the nose and/or mouth and a family history of sudden death in childhood should dictate a full and forensic analysis of the family history, including information from social services and other child protection agencies, the police, accident and emergency departments, and the family doctor. There should be a low threshold for performing a skeletal survey, retinal examination, and brain imaging (to identify possible coincidental subdural hemorrhage). Siblings should also be examined and their records analyzed. The Child Protection and National Police Registers (in the United Kingdom) should also be reviewed. Such investigations should be conducted with discretion if CVS is a possible subsequent option.

The high proportion of surveillance circumstances that revealed abuse (33 of 39) stresses the importance of protecting patients in whom there is a similar index of suspicion (with or without the use of surveillance). CVS was initiated because members of a strategy planning group strongly suspected abuse, although in 10 of the 33 positive results, the children had been referred for medical investigations with no suggestion from the referring pediatrician that the ALTE may have been attributable to abuse. Based on our experience with the child protection system in the United Kingdom, the specificity of the evidence available in the majority of patients is insufficient, without CVS, to ensure that child care proceedings are successful in protecting the child, particularly in the longer term. Only three children (cases 14, 32, and 35) were old enough to have developed sufficient verbal reasoning to recount what had happened to them. Interviews of such young children in a court are difficult, particularly for the children. During the period that we used CVS, we were also involved in child protection court proceedings in which there was no surveillance evidence, but in which we had the same index of suspicion of abuse. We were confronted by the plausibility of deceitful parents as well as by pediatricians acting as expert witnesses who were unable to accept what we considered reasonable evidence of abuse. The result was that a proportion of these children and their siblings were returned to parents despite our concerns that abuse of the type described here might continue. The absence of definitive evidence in these latter cases also resulted in lengthy court processes that delayed decision-making about the child's future. Although CVS could never be justified for financial reasons alone, an additional benefit was that shorter court proceedings enabled a limited social services budget to fund more resources to support other families in need. As a consequence, it is our view that CVS should become standard practice and used whenever possible to provide direct (noncircumstantial) evidence that will ensure effective and long-term protection. CVS also may be appropriate in investigating other abusive activity, particularly severe emotional abuse, involving parental deceit. This type of abuse should be regarded as extremely threatening to the child's mental and emotional health, and possibly fatal should it lead in later years to self-injury. However, the question arises as to what should be done if CVS cannot be implemented, as is currently the case in our own unit because of adverse media activity, or in other countries where such surveillance may not be permitted by law. Our suggestion is that all child protection personnel be educated to recognize the type of abuse that involves deceitful and sociopathic parents, and that nonadversarial strategies, including the use of experienced, impartial professionals, be implemented by the court to determine the facts in these cases.

Strategies for Identifying and Protecting Children From Abuse Inflicted by Parents Who Intentionally Deceive

Many of the parents reported here were severely disturbed people who not only torment, but sometimes kill, their children or stepchildren. In most respects, they are not similar to parents who inflict more common forms of abuse in situations in which poverty, poor education, isolation, parental depression, alcohol abuse, unemployment, and lack of social and family support may have been present for generations. Such families have been described as “families without hope”' by Tonge et al.⁴³ Injuries incurred by children in these families without hope may still be serious. However, our CVS group demonstrated abuse that was deliberate, sometimes planned, and associated with unusual injuries or medical signs.

Child abuse has complex origins, but abuse that involves long-standing, sadistic, premeditated harm is potentially most damaging. In this severe form, there is a complex interaction of constitutional factors between parent and child, including personality traits, and broader family and other socioenvironmental influences that in some cases and at certain times result in abuse.³⁶Child protection programs are rudimentary in many countries, and even nonexistent in others. Given the contents of the United Nations Convention on the Rights of the Child,⁴⁴ programs to address child abuse, particularly that involving parents with this form of psychological and personality disorder, must become a priority for all member states.

In most families in which CVS was implemented, there were previous substantial concerns based on family and social histories, often known only to individual professionals. This illustrates the importance of basic clinical skills, such as the ability to collect accurate data and to work well with colleagues in different disciplines (interagency collaboration), as the key to diagnosis and protection.¹³It is often difficult to collect and analyze all relevant information;³⁶ there is a need for objective, critical, forensic methods of data collection from different sources. Taking a detailed history from parents (including a search for some of the factors outlined in Tables 2 and 4) is within the scope of responsibilities of all health and social service professionals. In our experience, a multiagency strategy meeting (professionals working together), initially without parental involvement, can be valuable in collating information and developing an effective strategy. It is less time-consuming than a case conference in which parents are present, and professionals tend to be less inhibited in their assessments. Also, if parents become alert to professional suspicion of the cause of the child's medical condition, they may attempt increasingly to disguise abusive behavior.

Family doctors and their health visitors (HVs) often have access to the information needed to identify parents with personality disorders who are potentially likely to inflict the types of abuse described in this report,⁴⁵ particularly because some will have attention-seeking behavior about their own health (Tables 2 and 4) as well as about the health of their children. Unfortunately, in our experience, most family doctors did not attend case conferences or play a significant role in child protection. This may reflect divided loyalties^46-48 or the fear of legal repercussions, aggression from the patient's family, or media involvement. Systems should be established to make it easier, perhaps obligatory, for family doctors to participate more actively in the child protection process (as recommended by the UK Department of Health).¹³Similarly, it is important that hospital and community-based doctors work together to identify high-risk families.

Perhaps the most critical observation, and the most difficult to address, is the ability of the abusing parents to deceive (see the account of Mary Bryk “My mother is an incredible liar and a powerfully deceptive actress”).⁴⁹ Covert surveillance has revealed that many such parents appeared caring and kind in the presence of professionals, yet within seconds of being alone with the child, became cruel and sadistic. They were frequently charming and attractive people who revealed an ability to evade and deceive professionals from all agencies⁴⁷ as well as the media (see below). They are aptly described in the book People of the Lie.⁵⁰ Partnership with such parents may be simply ineffective or frankly dangerous in regard to the type of abuse described here.

A recurring experience encountered during our study was the tendency of some professionals to deny the possibility of life-threatening abuse. The perspective of these professionals can be dangerous. The attitude of optimism among social workers and others described by Dingwall⁵¹ still exists, and excessive reliance on partnership when the parent is determined to deceive has been documented previously. Sometimes inappropriate and dangerous alliances have developed between the abusing parent and health professionals involved in treating the child.⁵² On occasion, pediatricians were made to feel insensitive for addressing the potential role of abuse in the cause of the problem, particularly in the presence of life-threatening symptoms and signs. We agree with Donald and Jureidini,⁵² who stated that severe abuse is more likely to occur when there is “a medical system that is specialised, investigation orientated, fascinated by rare conditions, often ignorant of abusive behaviours and too accepting of reported histories.” This indicates the hazards of large specialist hospitals and investigative pathways that ignore psychosocial background.

It must be accepted that when working with parents who intend to deceive, agencies representing the interests of the child may have to work without parental collaboration. This results in a process geared more to child protection than to family support. A central issue that should be addressed at the onset of and during any child protection inquiry is determining how much trust can be placed in the parents. If they are already known to be deceitful and have features of a sociopathic personality disorder as described previously and outlined in Table 4, the identification of abuse and the protection of children may require a highly skilled investigative approach to the many deliberately placed obstacles to diagnosis.^48,51-55 Such an approach is difficult in the care professions. Staff involved in such investigations need to cooperate effectively in a forensic setting to produce evidence that can withstand rigorous examination in the courts. Highly professional, well-trained, and suitably experienced staff and sufficient government resources are essential to ensure early identification of abuse by deceptive parents and to provide adequate protection for children. In addition, the efforts of effective staff and resources may eliminate criticism of social workers for failing to protect children from this kind of abuse.^3,4 In our experience, and according to studies conducted in both the United States and Australia,^56-58 close cooperation with police and child protection professionals has provided important safeguards. We agree with the concepts raised by Boros et al⁵³ who have also implemented CVS to help detect this type of abuse. They classified it as a form of child torture and stated that “children must be protected and mentally competent adults must be held responsible for the violence they inflict.” Further, they viewed “this entity more as a profile of criminal abusive behavior and less as a medical syndrome.” This suggested approach is similar to that required to address child sexual abuse.

This analysis brings us to question child protection practices in the United Kingdom when dealing with such severe abuse. The current approach is to work with parents to overcome problems.¹³The UK Department of Health guidelines on the Children Act emphasize “the importance of professionals working in partnership with parents and other family members [and] … fully involved from the outset in all stages of the child protection process and [with] … as much openness and honesty as possible between families and professionals.” Inherently, this approach assumes that parents will collaborate honestly with the child protection team through registration, in case conferences, with care plans, and by accepting home meetings from social workers or HVs (even when unplanned) designed to monitor and thereby provide some protection for the child. Parents who inflict the severe abuse described here may provide attractive homes and satisfactory basic care for their children, thereby deceiving some professionals during home visits. Knowing that they are under close observation may also lead some parents to greater levels of concealment. Partnership can be effective only when genuine.⁵

Relevance to Sudden Unexplained Deaths and Neurologic Disorders

Despite reservations about the selection of our sam-ple, it is clear from these and other results^{31,32,35-3659-64}that a number of sudden unexplained infant and early childhood deaths, including some labeled SIDS, is attributable to intentional suffocation. To date, there is no marker to identify intentional suffocation at postmortem examination. The proportion of SIDS attributable to intentional suffocation therefore remains unknown. However, child abuse is not uncommon. In 1995, there were 10.6 million children <16 years of age in England, 34 954 of whom were on Child Protection Registers (3/1000).⁶⁵ In 1995, 996 000 children (15/1000) in the United States were confirmed victims of abuse and neglect.⁶⁶ Although intercountry comparisons of incidence present difficulties,⁶⁷ in part because of differences in definition, these figures suggest that it is probably appropriate to conduct a careful, sensitive forensic investigation of all sudden unexpected infant and early child deaths where no explanation is found during examinations after death, including the autopsy. Skeletal surveys should always be considered. Multidisciplinary confidential inquiries after such deaths may also help to identify possible intentional suffocation or other abuse.⁶⁸ Such identification will not help the victim, but will alert professionals to the risk to subsequent children, siblings, and even children outside the family, and address treatment for parents, if they acknowledge their behavior.

The high number of deaths in previous siblings of our index cases, many of which were confirmed to result from abuse, highlights the intractable and long-term risk to children within these dysfunctional families.³⁶

A significant proportion of cerebral palsy and learning difficulties in children remains unexplained. In four patients undergoing CVS, hypoxic cerebral injury was likely to have been the direct result of intentional suffocation. As with sudden unexplained deaths, the precise contribution of suffocation to these clinical problems remains difficult to quantify. The potential contribution of abuse as a cause of permanent neurologic deficit was first described by Nixon and Pearn,⁶⁹ and has been championed by Oliver^70,71in his epidemiologic data from NE Wiltshire. The analysis of 140 children in two hospitals for children with learning difficulties⁷² suggested that violent abuse is a cause of between 3% and 11% of learning disabilities. However, because cerebral injuries may arise from subdural hemorrhage and suffocatory abuse, both of which may occur without immediate clinically diagnostic findings, it was not possible to know the precise contribution made by each mechanism in Oliver's studies.

It is accepted that abuse is one cause of severe behavioral disorder in children.^36,73-75 Our follow-up of severe behavioral disorder was neither systematic nor comprehensive and we almost certainly underrepresent the frequency of this problem.

How Common is Severe Life-threatening Child Abuse?

Our data are selected primarily by referral from a large number of district hospitals. A recent survey by the British Paediatric Association Surveillance Unit⁷⁶ found 128 examples of MSBP that had reached a child protection case conference in the United Kingdom and Ireland during a 2-year period. Although this may appear relatively rare (0.5/100 000 children <16 years of age), there are undoubtedly many cases in which suspicions are not followed up. Carefully tracking a cohort of families in NE Wiltshire, Oliver³⁶ revealed at least 39 cases of intentional suffocation during a 20-year period in a population of 200 000 people (∼1 in 25 000 children).

Practical Difficulties in Implementing CVS

Although guidelines had been made about the timing of intervention during suffocation, it was always difficult to specify with precision when to interrupt abusive behavior by parents. Our opinion, supported by experience in the court system, was that the demonstration of abuse that did not involve suffocation, poisoning, or injury, however unpleasant and distressing for the child, would not ensure protection. Therefore, we adopted an approach that we felt would result in the best outcome for each child. The log of case 24 illustrates the dilemma for observers faced with abusive behavior that did eventually result in an injury to the infant. During observation of the events leading to the injury, we had decided to intervene only if the mother's behavior was considered sufficiently violent to produce or be about to produce an injury. Without seeing the videorecordings for the logs, it is difficult for the reader to know precisely the degree of violence that accompanied, for example, a slap to the infant's head or an episode of shaking. We might be criticized for risking a serious injury such as a subdural hemorrhage and, in retrospect, we might have avoided the incident in which the child's arm was fractured. We acknowledge that the risk of an unexpected violent act is a potential problem during CVS and that more discussion and experience are needed to minimize the risk of such adverse outcomes. However, we would argue that CVS per se cannot be deemed responsible for these injuries as evidenced by the suffocation of previous infants at home in a proportion of these families. Furthermore, we reiterate our views expressed above, that actions on suspicion of abuse before implementing CVS may not ensure protection.

Helping Parents Who Inflict Severe Forms of Abuse

There are many child abuse situations where the context of family dysfunction lends itself to successful intervention through a partnership between parents and professionals.⁷⁷ There is also evidence of effective treatment in some cases of the kind of abuse described here, but only when there is parental acceptance of the problem⁷⁸ (which is rare in our experience) as well as statutory intervention. Outcomes appear less favorable when there is short-term or less intensive intervention.^{53,54,74,75,79-82}

In parents with problems with their partners (see above), resolution of these difficulties might also contribute to the prevention of additional abuse.

Unfortunately, much of the literature on adults with personality disorders suggests that it is difficult to change behavior within the time limits required to make suitable decisions for the long-term care of young children. Parents may not wish to change and may not continue to attend psychotherapy or counseling sessions. Given the devastating effects they may have on children, and the usually intractable nature of the abusers' difficulties and their poor prognoses, children who are subject to severe abuse may have to be removed from the parents' care unless and until their safety and well-being can be ensured.

Role of the Media

There is no doubt that the media may be instrumental in raising awareness of child abuse, as exemplified in 1996 in the campaign byThe Independent (United Kingdom) publicizing abuse in childrens' homes. However, attention-seeking is usually an important part of the lifestyle of these parents. Media attention is thus attractive to them, and their assistance may be sought after a disclosure of potential abuse. Convincing evidence acquired by CVS has usually inhibited parents from approaching the media. However, questions about abusive parenting from less stringent evidence lead to efforts by abusers to counter accusations by manipulating public opinion through the media. A trial, usually resulting in an innocent verdict for parents, is seemingly readily achieved by the media, although reporters are unable to hear evidence held by professionals because of appropriate confidentiality considerations for protection of the child. By collusion with abusive parents and the need for sensational stories, the media have made CVS difficult to implement in our hospital, thereby eliminating an effective form of child protection. The media would also appear to have been a major contributor to the crisis in diagnosis and management of child sexual abuse that resulted in a public inquiry into child abuse in Cleveland, United Kingdom, in 1987.^83,84 Others have criticized the media in the United Kingdom^85,86 for its approaches to child abuse reporting and the resulting negative impact on child protection practices.