Construct validity of probable child maltreatment indicators using prospectively recorded information in a longitudinal cohort of Canadian children.

Background: Officially reported and self-reported measures of child maltreatment show poor agreement and may differentially predict psychosocial problems in adulthood. However, research remains primarily based on retrospective self-reports, warranting examination of the validity of prospective assessments of maltreatment.

Objective: To assess the construct validity of prospective indicators of child maltreatment using a longitudinal cohort of Canadian children.

Methods: The population-based cohort comprises 2120 participants born between 1997 and 1998 in Quebec, Canada. Methods: Maternal and familial risk factors (maternal age, depressive symptoms, and antisocial behaviors, socioeconomic status, and single-parent home) and early adulthood functioning difficulties (depression, anxiety, suicidality, alcohol misuse, and unemployment status) were assessed across various time points (0-23 years). Associations between factors and prospective and retrospective maltreatment indicators were appraised.

Results: Most maternal and familial risk factors (80 %) showed associations with indicators of prospective maltreatment (ΔM = +/-0.04 to 0.72; p < 0.05). Several early adulthood functioning difficulties (30 %) showed associations with physical (ΔM = 0.05 to 0.22; p < 0.05) and sexual abuse (ΔM = 0.33 to 0.34; p < 0.05), while emotional, supervisory, and physical neglect were only associated with educational/employment status (ΔM = 0.04 to 0.10; p < 0.05). Cumulatively assessed maltreatment also showed a dose-response relationship with maternal and familial risk factors/functioning difficulties.

Conclusions: The strong construct validity exhibited by our prospective indicators highlights the need to assess child maltreatment multi-modally. Our findings further contribute to the wider discussion surrounding the measurement of child maltreatment.