Recurrence of maltreatment: an application of the National Child Abuse and Neglect Data System (NCANDS).

OBJECTIVE: The research describes and compares patterns of maltreatment recurrence across multiple states using large samples, confirms the patterns of recurrence found in the literature, and explores unreported patterns of recurrence.
METHOD: A recurrence data set for calendar years 1994 and 1995 was constructed from the multi-state case level data from the National Child Abuse and Neglect Data System. These data were available for 10 states and included a range from 2,419 to 99,288 substantiated or indicated report-child pairs per state. A common set of data constructs lent consistency to data construction and analysis, while preserving differences in policy. Event History Analysis (survival) techniques were used.
RESULTS: Single site studies were confirmed across the 10 states. These include the pattern where neglect is most likely to recur, followed by physical abuse and then sexual abuse. Similarly, younger children are more likely to recur. A finding of the analysis is that the likelihood of recurrence increases in a systematic and consistent fashion based upon the sequential ordering of recurrent maltreatment events. Also, the likelihood of recurrence is associated with the provision of postinvestigative services.
CONCLUSION: Highly consistent patterns of recurrence were observed across states. Children experiencing multiple recurrences compared to no recurrence or one recurrence may represent a special at risk population requiring additional research. Adequate baselines and an understanding of recurrence is needed when considering recurrence as an outcome indicator or in developing risk assessment tools. Important recurrence patterns may be difficult to detect reliably with relatively small samples.