Pediatric Annals, Vol. 5, No. 3, March 1976, 140-144.
Knowledge and Epidemiology of Child Abuse: A Critical Review of Concepts
ELI H. NEWBERGER, M.D., and JESSICA H. DANIEL, Ph.D.
By identifying and clarifying concepts underlying present knowledge of child abuse, this article may provide the reader with critical tools for understanding it. We shall focus on the magnitude of the problem and probe the meaning of present data and practice.
There is reason to question the nature and quality of knowledge about child abuse. Formal recognition of an age-old phenomenon, demonstrated by an enormous increase in the number of official case reports annually since the mid-1960s, has created a difficult dilemma for medical professionals. Notwithstanding a century’s experience in the American child welfare movement and more recent medically based contributions from Kempe and others, we have a service system that, despite humane rhetoric, is unable to promote the safety and well-being of many children. This is in large part due to a paucity of such essential family supports as counseling, medical, homemaker, child-care, and nursing services and to a heavy reliance on foster-home care. A tightfisted social policy toward families and children ($2.5 billion for child welfare each year since 1972 in the face of two-digit inflation and expanding demands for service) means, simply, that when a professional person files a child-abuse case report, the services that follow may be incapable of dealing with the needs of family and child.
Inadequate or incomplete service is only part of the problem. Our basis for practice is flimsy. We have a commonly accepted humane philosophy (if not in reality programs that can translate that philosophy into humane action): to protect parents and children from repeated physical consequences of family crisis. But because we lack a solid theoretical and practical understanding of the origins of child abuse, our clinical work is at best intuitive and kind, at worst reflexive and mean. We read a literature in each of the professions characterized by homilies, bromides, and few scientific investigations of substance. And we look at child abuse as a phenomenon originating in the psychology of individuals, frequently ignoring the social and cultural realities that frustrate our treatment of particular families and impose formidable obstacles to the prevention of child abuse.
Because of the contradictions between philosophy and practice and our incomplete knowledge, we find ourselves wondering whether the following are unanswerable questions when cases of child abuse are identified: Is the child at risk? Can the family be helped? Are competent intervention resources available? Will I do more harm than good by reporting the case?
We do not mean to suggest that the clinician should throw up his hands in despair when the next case of child abuse is brought in. Within the framework of existing knowledge and resources, possible answers and helpful clinical guidelines can be drawn up, and these are the subject of a recent review.1 By presenting a glossary of concepts and pertinent current data in this article, we would like to help foster an informed and more logical response to child abuse.
“CHILD ABUSE,” “CHILD NEGLECT,” AND” ACCIDENTS”
Definitions of child abuse vary, from Kempe’s “battered child syndrome,”2 which identifies injuries inflicted by care givers; through Fontana’s “maltreatment syndrome,”3 which includes child neglect; to the current Office of Child Development model reporting statute,4 which embraces many physical and emotional symptoms attributable to parental failure; and to Gil’s concept of any force that compromises a child’s capacity to achieve his physical and psychologic potential.5 Virtually all definitions identify the child as victim, and most identify parent or family as perpetrator.
Important value concepts are built into the vocabulary, and in the words themselves are postulated etiologic mechanisms that logically imply diagnostic and intervention procedures. The names “battered child syndrome” and “maltreatment syndrome” have strong implications. They indicate that a child’s injuries were caused by his care giver, either actively or passively.
To make such “diagnoses” requires an investigation to determine whether or not thete is parental culpability. Inquisitions of parents to ferret out the facts have been characterized by some as clinically unhelpful, ethically absurd, and intellectually unsound.6,7 Faced with ambiguous data, conflicting accounts of how the child may have received his injuries, and a need to make a definitive diagnosis, the clinician may find himself playing a detective game for which he is professionally unprepared.
Stoked by the strong feelings that child-abuse cases promote in all of us, the diagnostic process may further alienate an isolated, frightened, and confused family and fulfill the preconception of parental failure: aggressive inquiry eliciting evasive response, angry affirmation of suspicion leading to confirmed’ diagnosis, and subsequent estrangement of family from clinician and separation of child from family.
Different professional people respond in different ways to the personal and ethical conflicts imposed by contact with troubled families. Some physicians find it difficult to believe that parents could injure children. Many characterize all children’s injuries as “accidents” (the term connotes an isolated, random event).
Although traumatic injury to children is the major cause of morbidity and mortality after the first year of life8 and is predictably associated with familial and child developmental crises,9-11 the nature and organization of child health practice do not usually permit exploring and acting on the causal antecedents of childhood “accidents.” Physicians and nurses may not have the time to interview parents or to make detailed child development observations, and such backup diagnostic services as social work and psychiatry are most often situated in separate institutions and practice settings. No treatment other than of the presenting symptom is implied by the diagnosis of an “accident.”
Further, because of the onerous significance of making a judgment that a particular family is “abusive” or “neglectful,” it is often easier to ignore these” diagnoses.” The finding that the great number of reported victims of child abuse are poor12 and disproportionately represent ethnic minority groups suggests that the more heavily value-laden diagnoses for childhood traumatic injuries (child abuse and neglect) are made more easily when the clinical setting is public and there is great social distance (social class or ethnic discrepancy) between clinician and family.
We clearly need a more scientific taxonomy of childhood “social illness,” one that would organize clinical data in such a way as to stimulate helpful and effective practice. Until we have it, however, we shall have to labor with the existing words.
The number of annual child-abuse case reports in the United States has jumped from about 7,000 in 196712 to over 200,000 in 1974.13 To understand the significance of this impressive increase for clinical practice requires – the explanation of a few simple concepts.
Incidence. This is the number of events in a defined interval or, for purposes of the present discussion, in a year.
Prevalence. This is the number of phenomena that exist at a particular time.
Duration. This is the length of the phenomenon.
A simple equation expresses the relationship of incidence, prevalence, and duration: Prevalence (P) varies as incidence (I) times duration (D) or, more simply, P = ID.14 (This assumes constant flow of information, or a “steady state,” into the system.) Although we are accustomed to thinking of the magnitude of the child-abuse problem in terms of the number of new case reports received each year, it is well to remember that these reports flow to human-service institutions (generally welfare departments) with finite capacities to deal with them. The prevalence of cases already “on the books” is a principal reason why the system bogs down.
Suppose, for the sake of discussion, that the duration of the typical child-abuse case, from diagnosis to resolution of the family problem, is 2.5 years. (This figure obscures the great variety of family situations that present as child abuse and, to be sure, minimizes the continuing, developmental impact of child abuse recently emphasized by Martin and his colleagues.15 Also, protective-service agencies follow few cases for this long. Take it as an arbitrarily selected number.) For 1967, then, an estimate of the prevalence of child abuse is 7,000 x 2.5 = 17,500. For 1974, the prevalence may be estimated as 200,000 x 2.5 = 500,000.
These estimates may bear little relation to reality, but in the unfortunate absence of reliable data on actual recorded prevalence (It is hoped that this situation will soon be remedied by the new National Center on Child Abuse and Neglect in the Office of Child Development in Washington, which has a specific Congressional mandate under Public Law 93-247 to report annually on the size of the child-abuse problem.), they provide an impression of the extraordinary burden that child welfare institutions are now shouldering. In the face of a rising number of case reports and the essentially fixed ability of the service structure to deal with them, there is an unavoidable risk that a case may not be acted on. Or, perhaps worse, precipitous, uninformed protective service action may separate a child from his family unnecessarily in the interest of his physical protection. So, when one sees a case of child abuse, one must assure that the case report is followed by sustained, helpful action.
Sampling bias. Some people are much more likely than others to be discovered as child abusers. Members of ethnic or economic minorities are particularly susceptible. This fact skews the findings of any surveyor study towards the attributes associated with the group to be studied. Poor families and ethnic minorities appear with special frequency in child-abuse case rosters and studies. This has in part to do with the dynamics of practice, as was pointed out above.
The pertinence of the findings of any investigation to the experience of the general population from which the study subjects are drawn depends on the investigator’s ability to account for whatever selective forces operate to include some and not other members of the population.
Our increasing comprehension of the personal and ethical dilemmas that go into making a diagnosis of child abuse inclines us to believe that as long as we persist in studying phenomena with such negative labels as “child abuse” and “neglect,” we will not be able to control the biases favoring the selection of poor and minority families in our samples. (The foregoing is not to deny the extent to which the exigencies of poverty are associated with physical and psychologic stress in families, which in turn may be expressed in violence or neglect toward children.17-19)
Confounding. Here, because of the investigator’s inability to control an intervening, perhaps unforeseen, variable, a spurious association is mistaken for a causal relationship.
Confounding can be reduced (although seldom completely avoided) by controlling studies with matched comparison groups. A striking example of confounding in the recent child-abuse literature is found in the British study by Smith and colleagues.20 The authors noted an impressive increase in the prevalence of psychologic problems in the parents
of abused children. This finding, however, may be an artifact of the different socioeconomic groups from which the cases and “controls” were drawn. Since social class and psychologic symptoms are associated,21 one cannot assert that there is a primary relationship between psychologic symptoms and child abuse. This finding is “confounded” by social class. The confounding would have been reduced by “matching” cases and controls on social class. This would “control” the confounding variable.
The concepts of sampling bias and confounding are of particular importance in regard to the postulated association between child abuse and later adult deviant behavior, which Schmitt and Kempe have underlined vividly in the most recent edition of Nelson’s Textbook of Pediatrics.22
If the child who has been physically abused is returned to his parents without intervention, 5 per cent are killed and 35 per cent are seriously reinjured. Moreover, the untreated families tend to produce children who grow up to be juvenile delinquents and murderers, as well as the child batterers of the next generation.
The truth of such assertions will have to be weighed by a careful consideration of the data from which they derive. Those socially marginal people who are most susceptible to having their children’s injuries characterized as having been “abusively” or “neglectfully” obtained are also those whose adult behavior is most likely to be labeled “criminal” This is not to minimize the importance of the developmental sequelae of child abuse. Rather, we urge critical and serious attention to the meaning of our professional language and knowledge, the better to develop a vocabulary of words and tools that will more adequately help children and families in distress.
1. Newberger, E. H., and Hyde, J. N. Child abuse: Principles and implications of current pediatric practice Pediatr. Clin. North Am. 22 (1975), 695.
2. Kempe, C H, et al. The battered child syndrome. J.A.M.A. 181 (1962), 17-24.
3. Fontana, V. J. The Maltreated Child: The Maltreatment Syndrome in Children, Second Edition. Springfield, III: Charles C Thomas, Publisher, 1971.
4. National Center on Child Abuse and Neglect, Office of Child Development, U.S. Department of Health, Education, and Welfare. Model Child Protective Services Act with Commentary, July 7, 1975.
5. Gil, D. G. Unraveling child abuse. Am. J Orthopsychiatry 45 (1975), 346.
6. Newberger, E. H. The myth of the battered child syndrome. Curr. Med Dialog 30 (1973), 327; reprinted in Chess, S., and Thomas, A (eds.). Annual Progress in Child Psychiatry and Child Development, 1974. New York: Brunner/Mazel, 1975, pp. 569-573.
7. Pollock, C. B., and Steele, B. F. A therapeutic approach to the parents. In Kempe, C. H., and Helfer, R. E. (eds.). Helping the Battered Child and His Family. Philadelphia: J. B. Lippincott Company, 1972, pp. 3-21.
8. Lessening Shadows. Evanston, III: American Academy of Pediatrics, 1970, p. 45.
9. Gregg, G. S., and Elmer, E. Infant injuries: Accidents or abuse? Pediatrics 44 (1969), 434.
10. Sobel, R. Psychiatric implications of accidental poisoning in childhood. Pediatr. Clin. North Am. 17 (1971), 653.
11. Holter, J. C, and Friedman, S. B. Child abuse: Early case finding in the emergency department. Pediatrics 42 (1968), 128.
12. Gil, D. G. Violence Against Children. Cambridge, Mass.: Harvard University Press, 1970.
13. De Francis, V. (Children’s Division, American Humane Association). Personal communication.
14. MacMahon, B., and Pugh, T. F. Epidemiology: Principles and Methods. Boston: Little, Brown and Company, 1970.
15. Martin, H. A, et al The development of abused children. In Schulman, I. (ed.). Advances in Pediatrics, Volume 21, Chicago: Year Book Medical Publishers, 1974, pp. 25-73.
16. Newberger, E. H Review of Violence Against Children, by D. G. Gil. Pediatrics 48 (1971), 688.
17. Light, R. Abused and neglected children in America: A study of alternative policies. Harvard Educ. Rev. 43 (1973), 556.
18. Lauer, B., Ten Broeck, E., and Grossman, M. Battered child syndrome: Review of 130 patients with controls. Pediatrics 54 (1974), 76
19. Newberger, E. H., et al. Toward an etiologic classification of pediatric social illness: A descriptive epidemiology of child abuse and neglect, failure to thrive, accidents, and poisonings in children under four years of age. Paper presented at the meeting of the Society for Research in Child Development, Denver, April, 1975.
20. Smith, S. M., Hanson, R., and Noble, S. Parents of battered babies: A controlled study. Br. Med. J 4 (1973), 388.
21. Hollingshead, A B., and Redlich, F. C. Social Class and Mental Illness. New York: John Wiley & Sons, 1958.
22. Schmitt, B. D., and Kempe, C. H. Neglect and abuse of children. In Vaughan, V C., and McKay, R. J. (eds.). Nelson Textbook of Pediatrics, Tenth Edition Philadelphia W. B. Saunders Company, 1975.
Dr. Newberger is director of the Family Development Study, Associate in Medicine in the Division of Child Development at Children’s Hospital Medical Center, Boston, and Instructor in Pediatrics, Harvard Medical School.
Dr. Daniel is Assistant Professor of Educational Psychology at Boston College and co-investigator of the Family Development Study, Children’s Hospital Medical Center, Boston.
The work reported in this article was supported in part by a grant from the Office of Child Development. Department of HEW (Project OCD-62-141).