Derivation of probable child maltreatment indicators using prospectively recorded information between 5 months and 17 years in a longitudinal cohort of Canadian children.

Background: Both prospective and retrospective measures of child maltreatment predict mental and physical health problems, despite their weak concordance. Research remains largely based on retrospective reports spanning the entire childhood due to a scarcity of prospectively completed measures targeting maltreatment specifically.

Objective: We developed a prospective index of child maltreatment in the Québec Longitudinal Study of Child Development (QLSCD) using prospective information collected from ages 5 months to 17 years and examined its concordance with retrospective maltreatment.

Participants And Setting: The QLSCD is an ongoing population-based cohort that includes 2,120 participants born from 1997-1998 in the Canadian Province of Quebec.

Methods: As the QLSCD did not have maltreatment as a focal variable, we screened 29,600 items completed by multiple informants (mothers, children, teachers, home observations) across 14 measurement points (5 months-17 years). Items that could reflect maltreatment were first extracted. Indicators were derived across preschool, school-age and adolescence periods and by the end of childhood and adolescence, including presence (yes/no), chronicity (re-occurrence), extent of exposure and cumulative maltreatment. Two maltreatment experts reviewed these items for inclusion and determined cut-offs for possible child maltreatment (n=251 items). Retrospective maltreatment was self-reported at 23 years.

Results: Across all developmental periods, the presence of maltreatment was as follows: physical abuse (16.3-21.8%), psychological abuse (3.3-21.9%), emotional neglect (20.4-21.6%), physical neglect (15.0-22.3%), supervisory neglect (25.8-44.9%), family violence (4.1-11.2%) and sexual abuse (9.5% in adolescence only). The degree of concordance between prospective and retrospective reports for each type of maltreatment was weak (.038-.110), yet significant (p<.01), except for emotional neglect (p=.148).

Conclusions: In addition to the many future research opportunities offered by these prospective indicators of maltreatment, this study offers a roadmap to researchers wishing to undertake a similar task.