Cross-type recidivism among child maltreatment victims and perpetrators

Abstract

Objective: This study investigated the extent to which child maltreatment victims and perpetrators were reported for different types of maltreatment over time (cross-type recidivism). Second, this study examined whether certain individual, community or child welfare service variables were associated with a tendency for the first recidivism event to be the same as the initial report among cases involving sexual abuse, physical abuse and neglect.

Method: Statewide administrative data on child abuse reporting at the child and perpetrator levels were linked to data on child welfare services and census information to examine cross-type recidivism prospectively for 4.5 years. Analyses include descriptive and logistic regression techniques.

Results: There was substantial cross-type recidivism at both the child and perpetrator levels, with neglect being the most common recidivism type. Among neglect cases, “lack of supervision” was the most frequent subtype at time of re-report regardless of the initial subtype of neglect reported. Predictors of remaining within-type varied by the type of maltreatment initially reported and by perpetrator compared to child-level analyses. For example, among physical abuse cases, older child victims were more likely to remain within-type, while the opposite was true for neglect.

Conclusions: Cross-type recidivism is common among re-reported cases of maltreatment. Non-neglect cases that re-reported to child welfare agencies are likely to return for neglect. Child welfare risk assessment, service provision, and research on children and families with a recidivism event should be focused neglect and on broad areas of need and risk rather than rely on typologies based on the index event. Researchers attempting to model recurrent maltreatment should consider recurrence at the child and perpetrator levels separately.

Résumé

Objectif: Dans un premier temps, cette étude a enquêté sur la mesure dans laquelle des victimes de mauvais traitements et des agresseurs furent signalés et re-signalés durant une certaine période de temps, dans la perspective de différents types de mauvais traitements (c.-à-d. la récidive où apparaissent plusieurs types de mauvais traitements). Deuxièmement, l’étude a voulu savoir s’il existe un lien entre certains individus, certaines communautés ou services de bien-être à l’enfance d’une part, et la tendance que le deuxième signalement porte sur le même type de mauvais traitement que le signalement initial. Les signalements comprenaient des cas d’agressions sexuelles et physiques et de négligence.

Méthode: L’étude prospective portant sur une période de 4,5 ans a analysé des signalements de mauvais traitements (victimes et agresseurs) dans le contexte des services de bien-être à l’enfance et des données tirées du recensement, afin d’étudier un récidivisme où apparaissent plusieurs types de mauvais traitements. On a utilisé des méthodes descriptives et de régression logistique.

Résultats: On a noté une prépondérance de récidives qui comprenaient plusieurs types de mauvais traitements, tant pour l’enfant que pour l’agresseur. La négligence est le type qui apparaı̂t le plus souvent, et parmi ces cas, le manque de supervision figure le plus souvent dans les re-signalements, peu importe le type de négligence antérieure. Les facteurs pouvant prédire de futures récidives du même type varient selon le type de mauvais traitement initial et selon l’agresseur. Par exemple, parmi les cas de mauvais traitements physiques, les victimes plus âgées sont plus aptes à redevenir victimes du même type de mauvais traitement tandis que le contraire existe pour les cas de négligence.

Conclusions: Parmi les victimes qui font l’objet de plusieurs signalements, on retrouve fréquemment plusieurs types de mauvais traitements. Les cas où il n’y a pas eu négligence dans le passé seront signalés pour cause de négligence. Les évaluations du risque, les services offerts et les recherches portant sur les familles qui récidivent devraient être attentifs à la négligence et ouverts à toutes les possibilités de mauvais traitements plutôt que de se baser sur des profils propres au type de mauvais traitements antérieurs. Les chercheurs qui élaborent des modèles d’intervention propres à la récidive devront prendre en considération séparément l’enfant et l’agresseur.

Resumen

Spanish-language abstract not available at time of publication.

Introduction

For both theoretical and analytic reasons, it is often desirable to create typologies. These typologies break down heterogeneous classifications, such as child maltreatment, into (presumably) more homogenous and understandable types, such as sexual abuse, physical abuse and neglect. It is attractive to consider each of these types as separate and distinct, and this trend has been evident in most of the published literature to date. Theory and research based on relatively “pure” typologies, however, may have difficulty accounting for children who may have several episodes of maltreatment over time. In order to address this concern, this paper asks the following question: “to what degree are children and perpetrators involved in multiple types of maltreatment over time?” The data presented describe cross-type maltreatment, which we define as recidivism events of child maltreatment that are not of the same type as the initial (index) event.

The need to differentiate between types of maltreatment has become a generally recognized standard in the literature (Drake, 1995; Eckenrode, Powers, Doris, Munsch, & Bolger, 1988; Giovannini, 1995, Sedlak & Broadhurst, 1996). Many studies have shown that statistical relationships obtaining in one type of maltreatment may differ from those found in other types. For example, Drake and Pandey (1996) found effects of community poverty upon rates of substantiated reports to be twelve times greater in neglect cases than in sexual abuse cases. Similarly, the NIS-3 (Sedlak & Broadhurst, 1996) shows dramatic differences on several variables (such as the association of poverty to child maltreatment) related to type of maltreatment. It is therefore clear that different types of abuse do distribute somewhat differently across the population and do have some different predictors and outcomes. What is not clear, however, is to what degree these different types of maltreatment overlap, both in individual maltreatment situations (mixed-type maltreatment), and over time (cross-type maltreatment).

A few studies have noted cross-type maltreatment trends in passing, but none has attended to the issue in a comprehensive or focused manner (DePanfilis & Zuravin, 1999a, DePanfilis & Zuravin, 1999b; English, Marshall, Brummel, & Orme, 1999; Fluke, Yuan, & Edward, 1999; Jonson-Reid, 2002). The present paper breaks new ground in four ways. Cross-type recidivism is described separately for five separate major forms of maltreatment. The study examines cross-type maltreatment as a function of multiple re-report events rather than as a function of the first re-report only. We explore predictors (e.g., gender, age, service use) of remaining within-type for cases with a recidivism event. Finally, the present study examines cross-type recidivism at both the child and perpetrator levels.

There is a growing empirical consensus that maltreatment subtypes often do not occur in isolation in a given case (Kinard, 1994). However, the estimates of the percentage of cases representing mixed or multiple types of maltreatment vary widely. At one end of the spectrum, Ney, Fung, and Wickett (1994) estimate only 5% of maltreatment allegations are “pure” types (e.g., sexual abuse only or physical abuse only). Daro (1988) describes 40% of maltreatment types in the National Clinical Evaluation Study as being pure types, while Levy, Markovic, Chaudhry, Ahart, and Torres (1995) found only 54% of cases to represent “pure” types of maltreatment. Among national studies, the NIS-3 shows approximately 15% of cases with multiple types of maltreatment data inferred from overlapping totals in Table 3-1 (Sedlak & Broadhurst, 1996, p. 3). The most recent NCANDS data (USDHHS, Administration on Children, Youth, and Families, 2001) unfortunately do not provide specific data on the incidence of mixed-type maltreatment.

A large number of studies have attempted to characterize risk factors for child maltreatment. Among these, Nelson, Saunders, and Landsman (1993) found that chronic neglect commonly co-occurs with a range of problems in the family, including parental mental health issues and parent/child conflict, which are also correlates of other types of maltreatment. In an unusual study that attempts to determine causes of maltreatment from the child’s point of view, many of the causes of physical abuse and neglect were found to be similar, including immature parents, marital problems and alcohol abuse (Ney, Fung, & Wickett, 1992). In a different article, these same authors suggest that neglect may be a precursor to sexual or physical abuse (Ney, Fung, & Wickett, 1993).

In summary, the rate of reported mixed-type maltreatment varies widely, due to different classification methods. It is likely that more comprehensive and detailed screening procedures (e.g., Levy et al., 1995) result in higher rates of mixed-type maltreatment than are commonly found through equally well structured but less detailed assessment procedures (e.g., NIS-3). These efforts are further hampered by the lack of agreed-upon and clear definitions of subtypes of maltreatment (Mash & Wolfe, 1991).

We define cross-type recidivism to be a recidivism event of a different type from the index event. There are at least three ways to think about cross-type maltreatment:

Sequential cross-type maltreatment may occur if a child is at risk of or suffers maltreatment of one form, and then later, after the initial form of risk goes away, is at risk of or suffers another kind of maltreatment (Ney et al., 1993). For example, a young child may be at risk of lack of supervision as an infant, but may be at greater risk of coming to the attention of a child welfare agency for physical or sexual abuse in his or her adolescent years. It is possible that when children move from one type of maltreatment to the next, that there is some contributory or evolutionary role of the first type of maltreatment. It is also possible that the presence of multiple types of maltreatment over time are due to other underlying factors such as poverty, environmental stressors, mental health, substance abuse or familial dynamics issues.

Hidden mixed-type maltreatment occurs when children are at risk of experiencing or do experience different types of maltreatment simultaneously, but only a single type of maltreatment is identified. This may come about due to the investigative agency not being able to find adequate evidence of all the various types of maltreatment which are present or because the reporting party is only aware of one aspect of the maltreatment the child is experiencing. When such children return to the system, they may be reported for another type of maltreatment, thus appearing to be examples of cross-type recidivism. In actuality, such cases are merely being reported at different times for different aspects of a set of continuing problems, making the particular type of abuse noted in any report a somewhat random classification.

Other forms of cross-type recidivism. Almost all the current research explores cross-type recidivism using the child as the unit of analysis. It is also possible to consider other types of mixed-type or cross-type recidivism. For example, a given family might include different children suffering different forms of maltreatment. If the perpetrator is used as the unit of analysis, one might investigate the degree to which the same individual has a tendency to commit the same (or different) types of maltreating behavior, possibly across several different children.

There is an emerging literature documenting the presence of cross-type (as opposed to mixed type) recidivism. In their study, Levy et al. (1995) state “although no particular initial maltreatment diagnosis was a statistically significant predictor of the likelihood of type of re-abuse, neglect was shown to be the most frequent type of re-abuse” (p. 1363). They highlight that “… of particular interest was the minimal relationship between the type(s) of abuse diagnosed at discharge from the hospital and the type of re-abuse.” English and colleagues (1999) also found that type of maltreatment at index does not predict re-referral type. DePanfilis & Zuravin, 1999a, DePanfilis & Zuravin, 1999b also found high rates of cross-type recidivism, with neglect the most common form of recidivism, but they were able to identify some within-type tendency in their data (e.g., a statistical tendency for sexual abuse to be more likely than other types of maltreatment to be re-referred as sexual abuse). Fluke and colleagues (1999) also found high rates of neglect in re-reports, regardless of index type. Using administrative data from 10 counties in California, Jonson-Reid (2002), found that over 40% of children with two reports or more were classified as having at least two types of maltreatment over time. Neglect was the most common second type of reported maltreatment, regardless of the type of maltreatment at the index event.

This report seeks to answer a series of questions about cross-type recidivism using a large statewide administrative sample. The basic descriptive questions addressed were: (1) What was the likelihood of cross-type maltreatment among a sample followed for 4.5 years? (2) How did cross-type maltreatment vary by index event type (sexual, physical, neglect, emotional, other), by number of recidivism events and by sample classification (victim, perpetrator)? The analytic question addressed by this study was “What community, family, victim, service and case characteristics predict the tendency of a case to remain within type?”