• The epidemiology of child maltreatment recurrences

      DePanfilis, Diane; Zuravin, Susan J., 1944- (1995)
      The primary aim of this dissertation was to elucidate the epidemiology of child maltreatment recurrences among families known to public child protective services (CPS). The specific objectives were: (1) To describe the pattern of recurrences over time; (2) To identify correlates of recurrence during and following CPS intervention; and (3) To describe the pattern of multiple recurrences. Subjects were a cohort of 1167 families who experienced a confirmed report of maltreatment and were followed for five years. Data on variables were collected from archival sources. To achieve objective 1, Life Tables were constructed at 30 day intervals to estimate the probability of recurrence during each interval. Survival functions of groups were then compared with the Wilcoxon (Gehan) statistic. To achieve objective 2, Kaplan Meier survival analyses were performed to compare the survival functions of potential variables and to test the Cox proportionality assumption of each variable. Models were then estimated with the Cox Proportional hazards model. To achieve objective 3, methods involved comparing the mean time until each recurrence between groups. Results suggest that risk of recurrence is greatest during the first thirty days following a report, that it is dependent on the type of maltreatment, and that it declines over time through the service period and remains relatively low for two years following the termination of services. Correlates of the time until first maltreatment recurrence while CPS was active were: child vulnerability, family stress, partner abuse, social support deficits and an interaction between family stress and social support deficits. Only one factor, age of the mother, predicted time until recurrence following services. Most recurrence families experienced only one recurrence. As the number of recurrences increased, the length of time between recurrences decreased. Future research should prospectively follow families over time to increase understanding about the specific treatment strategies and social supports most helpful to reducing risk of recurrence.
    • Intrafamilial child sexual abuse: Characteristics that predict maternal belief and protective action among non-offending mothers

      Pintello, Denise Anne; Zuravin, Susan J., 1944- (2000)
      The primary purpose of this study was to examine characteristics that predict maternal belief and protective action among non-offending mothers of sexually abused children. Secondary objectives included identifying the proportion of mothers who believed, protected and performed various combinations of both maternal responses, and examination of the impact of belief and protection on child sexual abuse recurrence. Data were collected from 435 biological, non-offending mothers through case record abstraction and computerized database review. Descriptive statistics measured proportions of maternal belief and/or protection. Logistic regression identified predictors of belief, protection, the four combinations of maternal responses and sexual abuse recurrence. Results indicated that approximately half of the mothers believed and/or protected. Four distinct combinations of maternal responses were documented, suggesting that mothers are a heterogeneous group concerning postdisclosure belief and protection. Findings supported prior research indicating that mothers who were not current sexual partners of offenders were more likely to believe and protect. The absence of maternal substance abuse was also found to predict maternal protection. Study results failed to support prior studies reporting victim age, gender, prior physical abuse history, sexual abuse severity and offender substance abuse as significant predictors of belief and/or protection. This study extended new knowledge by identifying 18 new predictors. Mothers were more likely to protect when they postponed their first birth until adulthood and their children did not exhibit sexualized behaviors. Mothers with no prior knowledge about the abuse before disclosure were more likely to believe and protect. Mothers who were unemployed, and had a prior trauma history of domestic violence and/or childhood sexual abuse were more likely to believe, yet not take protective action. Finally, this study contributed to new knowledge by identifying maternal non-protection as a predictor of sexual abuse recurrence. Further study is recommended to investigate the longitudinal impact of maternal, child and situational predictors on belief protection and their association to sexual abuse recurrence. It is hoped that the empirical data generated from this study will enhance child welfare interventions by fortifying maternal belief and protection, which may ultimately reduce out-of-home placements, maltreatment recurrence, and the psychological trauma endured by sexually abused children.
    • The neighborhood correlates of child maltreatment: Montgomery County, Maryland

      Ernst, Joy Swanson; Zuravin, Susan J., 1944- (1999)
      This study aimed to extend theoretical and empirical knowledge about the neighborhood-level determinants of child maltreatment in a suburban jurisdiction. The objectives were to: (1) compare the distribution by neighborhood of physical abuse, neglect, and sexual abuse; (2) identify and compare the neighborhood correlates of these three types of child maltreatment; (3) identify and compare racial- and ethnic-specific rates and correlates of child maltreatment; and (4) replicate a study of community-level factors and child maltreatment in Cleveland, Ohio, with Montgomery County, Maryland, data. Families investigated for child maltreatment in 1995 were matched to their U.S. census tract, the operational definition of neighborhood. Rates of maltreatment were calculated and mapped for each of the county's 159 tracts and used as dependent variables in regression models that revealed the neighborhood-level correlates of three, types of maltreatment and racial- and ethnic-specific rates of maltreatment. Predictor variables hypothesized to represent economic and social resources were derived from 1990 U.S. census data. To replicate the Cleveland study, variables representing levels of community social organization were subjected to a principal components analysis. Factors derived from that analysis were used as independent variables in a regression analysis of the child maltreatment rate.;Results for each objective extended knowledge of neighborhood-level factors associated with maltreatment. First, the distribution of child maltreatment varied by type of maltreatment and by neighborhood. Physical abuse was the most common; sexual abuse the least common. Second, variables representing economic and social resources accounted for a significant proportion of the variance in rates of physical abuse, neglect, and sexual abuse. Social resource variables contributed independently to the variance in the rates of physical abuse. Significant correlates varied by type of maltreatment. Third, the rates and distribution of maltreatment differed by racial and ethnic group, with the highest rates among Black families and the lowest among White families. The set of predictor variables accounted for a significant proportion of the variance in the rates of both White and Black maltreatment. Fourth, the replication revealed that neighborhood rates of child maltreatment varied with community-level factors associated with economic disadvantage and residential instability.
    • The parenting of women sexually abused as children: Abuse characteristics, cognitive responses, attachment styles and parenting adequacy

      McMillen, John Curtis; Zuravin, Susan J., 1944- (1993)
      Recently, researchers began exploring the parenting behaviors of women who were sexually abused as children. These studies suggest this group may be prone to parenting problems. Reviews recently admonished researchers to begin studying intervening factors in the development of long term consequences from sexual abuse (Cahill, Llewelyn & Pearson, 1991; Kendall-Tackett, Meyer Williams & Finklehor, 1993). Following the suggestions of these reviewers, three domains of intervening factors were examined in their relationship to parenting adequacy--characteristics of the sexual abuse, cognitive responses to the sexual abuse and adult attachment styles. Previous research identified 220 low income mothers in Baltimore who were sexually abused as children. These mothers came from two cohorts--women known to have a maltreated child, and AFDC mothers not known to have a maltreated child in 1989. One hundred sixty two (73.5%) were re-interviewed for this study. Eight interviews were unusable, leaving 154 subjects for data analyses. Parenting adequacy was operationalized in three ways, including an archival measure of Child Protective Service (CPS) Status, and two self report measures from the Conflict Tactics Scale (Straus, 1979)--severe physical aggression and verbal aggression. Five indicators of abuse characteristics were examined--the use of force, coitus, frequency, age at first abuse, relationship to perpetrator and perceived support. Several measures of cognitive responses to sexual abuse were created for this dissertation. They included attributions of blame and perceptions of harm and benefit. Adult attachment styles/internal models were assessed using the Relationship Questionnaire (Bartholomew & Horowitz, 1991) and standard measures of self esteem (Rosenberg, 1965) and views of the world (Janoff-Bulman, 1989). Few of these independent variables were related to parenting adequacy. No single independent variable was related to all three indicators of parenting adequacy. Those with more perceived support, and those who perceived a little benefit from the sexual abuse, were less likely to have a maltreated child. Contributions of this study include measurement development for assessing attributions of blame for child sexual abuse. The attribution scale appears to be composed of three factors--self-blame, family blame and perpetrator blame. Each factor appears to have adequate reliability and construct validity. In addition, the study provides new insight into the cognitive responses women make about sexual abuse.
    • Psychiatric readmission of adolescents in the public mental health system

      Fontanella, Cynthia A.; Zuravin, Susan J., 1944-; Burry, Caroline Long (2003)
      The growth of Medicaid managed care in the 1990s has led to substantial changes in the financing and delivery of behavioral health services for children and adolescents. Despite the rapid shift to managed care, few studies have investigated the effect of these changes on service provision to children with serious emotional disturbances. The primary aim of this study was to evaluate the effect of Maryland's Medicaid behavioral health plan on patterns of psychiatric readmission of adolescents. Specific objectives were: (1) to determine whether the rates and frequency of readmissions differed before (Fiscal year 1997) and after (Fiscal year 1998) the implementation of Maryland's Medicaid managed care program; (2) to identify factors that predicted readmission; and (3) to determine the relationship between neighborhood risk factors and readmission. To achieve study objectives, a non-concurrent prospective design was used. The sampling frame consisted of 881 Medicaid-eligible adolescents consecutively admitted to three private psychiatric hospitals between July 1, 1996 and June 30, 1998. Adolescents were followed up for a one year period past their index admission to determine whether they were readmitted to any psychiatric hospital in Maryland. Data was drawn from hospital case records, Medicaid claims data files, and the Area Resource File. While the study findings indicated that there were no significant differences in the overall rates of readmission for the two years, adolescents admitted after the implementation of the managed care program were far more likely to experience multiple readmissions. The cumulative one year rate of readmission was 33% for fiscal year 1997 and 38% for fiscal year 1998. The highest risk period for both years was within the first 30 days post-discharge (14% in 1997 and 13% in 1998). Rates of readmission also varied considerably across hospital providers. Adolescents were more likely to be readmitted if they were younger, had more severe emotional and behavioral disturbances and/or comorbid mental retardation, came from high risk families and had histories of childhood abuse. Type of aftercare services and living arrangements were also important determinants of readmission. Finally, the results indicated that two of the four neighborhood factors (family structure and residential mobility) examined were significantly associated with readmission for the Baltimore sample. Findings revealed that youths who came from areas characterized by high residential mobility, a predominantly African American population, female-headed households, and a high child/adult ratio were less likely to be readmitted. While these findings are counterintuitive, they may suggest racial disparities in access to health care.