Title: Childhood sexual abuse: An evidence based perspective
Authors and Affiliations:
David M. Fergusson, Christchurch School of Medicine, New Zealand
Paul E. Mullen, Director, Victorian Institute of Forensic Mental Health, Australia
Citation: Fergusson, D.M. & Mullen, P.E., “Childhood sexual abuse: An evidence based perspective,” Thousand Oaks, California: Sage Publications, 1999.
Childhood sexual abuse (CSA) is generally treated as though it is a recognizable syndrome, similar to the measles, identifiable by certain invariant symptoms. It is not. Rather, its definition rests on normative judgments. Defining it results from first gathering information about childhood sexual experiences, then evaluating these experiences against a normative standard to determine whether they are acceptable or not. There are two consequences of this process.
First, there is great heterogeneity in those experiences classified as abuse and in those people classified as victims. Victims vary greatly in their characteristics and experiences, and are linked only by the fact that they have been exposed to sexual behavior deemed to be inappropriate and socially unacceptable. This fact results in major difficulties in interpreting estimates of the prevalence of CSA.
Secondly, there is no single universal definition of CSA, nor can there be one. There is no generally accepted set of normative and moral standards to classify sexual behavior as abusive. There is a range of opinion from definitions that classify any unseemly childhood sexual experiences as abusive, to those that argue that many incidents called CSA are welcomed by the child and beneficial.
The most straightforward approach to this situation is to describe different types of unwanted and potentially harmful sexual experiences, then to ask questions about the psychological effects of particular sexual behaviors on children, without debating whether they are abusive. In line with this approach, research into CSA has been moving away from definitions that combine a wide variety of behaviors into a general category of CSA, toward descriptions of the nature, extent, and intrusiveness of unwanted sexual experiences in children and how they may affect children and adults.
Knowledge and opinion regarding CSA comes from three sources: accounts by CSA survivors of their experiences and the effects of those experiences on their lives, similar accounts by clinicians, and population studies of prevalence and effects.
Each of these sources gives different data. Survivor accounts show that CSA occurs and may have a large impact on adjustment, but fail to tell how prevalent such experiences are and the extent to which their experiences are typical. Survivor accounts can also be influenced by opinions, attitudes, and values of others who attempt to give meaning to their experiences.
Clinician’s reports are more generalizable since they look for common features in cases of CSA. However, they may contain an overrepresentation of severe cases with more psychological effects. Population samples attempt to overcome these problems, but obscure the contexts and life experiences of individuals, and the particular influence of the experiences on them.
A review of about 30 studies of community and convenience samples found widely varying estimates of the prevalence of CSA. Estimates ranged from 3% to 30% for males and from 6% to 62% for females.
These variations were due to at least four factors. The first factor was differing definitions of CSA among various researchers. Because there is no agreed upon standard, each investigator must adopt arbitrary criteria regarding the extent of sexual intrusion, contact, and coercion that classifies activity as abusive. One approach to this problem would be to classify children by the nature of the unwanted sexual experience they encountered. This would allow more precise examination of the impact of CSA.
Another reason for the wide variation in estimates was the reliance on retrospective reports by adults on their memories of childhood events. Memory is subject to error, and adults may be reluctant to disclose sexual activities.
A third factor involves varying ways of framing questions about abuse. The wording of questions affects how subjects review their childhood history and decide whether their experiences constitute abuse. In addition, the medium of the survey—anonymous questionnaires, telephone contacts, or in-person interviews—affects the results.
Finally, the sample used in the survey influences the results. Clinical, child welfare, and criminal justice samples usually report high prevalence rates which are not representative of the general population. Bias can also arise due to non-response, method of sample selection, and geographical region from which the sample is selected.
The authors conclude:
...the popular rendition of the literature on CSA has frequently resulted in trite conclusions that are chanted like sacred mantras about the proportion of children who are sexually abused. However, underlying these trite and perhaps socially convenient claims, there is a complex body of evidence that is both highly variable and by no means easy to interpret. Reducing this evidence to claims that one in four (or whatever fraction of) children is subject to sexual abuse conceals the very real uncertainties, debates, and issues that surround this evidence...
...exposure to unwanted sexual attention in children is not uncommon and, more importantly, there is consistent evidence to suggest that 5% to 10% of children are exposed to severely abusive acts involving actual or attempted sexual penetration.
A review of over 30 studies found that children known to have been sexually abused are vulnerable to a wide range of behavior problems, mental health disorders, and adjustment difficulties, including anxiety, fear, depression, somatic complaints, aggressive behavior, inappropriate sexual behavior, and problems with learning and behavior.
It is probable, but by no means certain, that these problems can be attributed in part to CSA. However, research is likely to overestimate the impact of CSA due to biased sample selection and failure to control for confounding influences. Subjects for these studies are usually not selected randomly from the population, but rather from agency records and similar sources. Such studies may produce misleading or spurious associations between CSA and adjustment problems, the latter of which are the reason the subjects came to the attention of the agency.
CSA is often correlated with other adverse and disadvantageous situations such as family dysfunction and parental adjustment difficulties which may contribute to the children’s adjustment problems. Most studies do not attempt to account for these factors, which pose major threats to the validity of their conclusions that CSA causes adjustment problems.
Some studies do show that an increase in the severity and duration of CSA are correlated to an increase in symptoms, which would suggest some degree of causality. Biographical accounts and the fact that CSA-specific therapy reduces symptoms also support the hypothesis of causality.
There is little evidence that CSA causes a typical unique pattern of symptoms, although there is a pattern of symptoms that is similar to that found in children exposed to a wide variety of childhood and family adversities.
However, a substantial minority of children shows no symptoms at all. Whether or not symptoms occur seems to depend on the severity of the incident, the extent of family support and nurturance, and the child’s attitudes and coping skills.
Studies show that there has been rapid growth in therapies provided for victims of CSA , but the evidence for their effectiveness is weak. Treatment is complicated by the fact that CSA is not a disorder, but rather an assumed cause of disorder. This makes it difficult to determine treatment objectives, methods, and assessments, especially when the child exhibits no symptoms.
Studies have consistently documented correlations between CSA and increased depression, anxiety disorders, antisocial behavior, substance abuse, eating disorders, suicidal behavior, post-traumatic stress disorder, and sexual adjustment difficulties.
To investigate this further, the authors found 12 studies conducted since 1990 that met the following criteria: (1) they were based on community rather than clinical samples, (2) they involved at least 100 people, (3) they used a clearly stated definition of CSA, (4) they measured symptoms using standardized methods, and (5) they presented data so that odds ratios could be calculated. Odds ratios tell the reader how many times more likely a CSA victim is to experience a certain outcome than is someone who has not experienced CSA.
These odds ratios showed that CSA victims were more at risk for depression, anxiety, suicidal behavior (although it was rare among both CSA and comparison groups), substance abuse, eating disorders, and psychiatric disorders. There were pervasive associations between CSA and psychiatric disorders.
The odds ratios varied depending on the stringency of the definition of CSA; odds ratios were higher for more severe CSA. There was also a pervasive relationship between reporting CSA and sexual adjustment difficulties among women.
However, these negative outcomes were not necessarily caused by the CSA because of the confounding influence of disturbed and disadvantageous family backgrounds, including impaired parenting, parental marital problems, and parent maladjustment.
A growing number of studies accounting for these factors have found reduced but not absent associations between CSA and negative long-term outcomes, suggesting that family background is partly to blame. However, it is important to remember that not all relevant confounding factors may have been accounted for, such as other forms of abuse and deprivation.
The evidence did of this review did not support the belief that CSA was linked to specific types of dissociative disorders. That is, there was no CSA-specific syndrome that resulted. Although there has been much theorizing, there is no clear account of how CSA leads to psychiatric disorders or adjustment problems.
In addition, estimates suggest that up to 40% of those exposed to CSA may be symptom free. Strong advocacy has led to a rapid growth in therapies for adults who experienced CSA, but that growth has not been paralleled by research examining their effectiveness. In fact, evidence for the effectiveness of therapy is weak.