Toward a cumulative ecological risk model for the etiology of child maltreatment

The purpose of the current study was to further the integration of cumulative risk models with empirical research on the etiology of child maltreatment. Despite the well-established literature supporting the importance of the accumulation of ecological risk, this perspective has had difficulty infiltrating empirical maltreatment research and its tendency to focus on more limited risk factors. Utilizing a sample of 842 mother–infant dyads, we compared the capacity of individual risk factors and a cumulative index to predict maltreatment reports in a prospective longitudinal investigation over the first sixteen years of life. The total load of risk in early infancy was found to be related to maternal cognitions surrounding her new role, measures of social support and well-being, and indicators of child cognitive functioning. After controlling for total level of cumulative risk, most single factors failed to predict later maltreatment reports and no single variable provided odd-ratios as powerful as the predictive power of a cumulative index. Continuing the shift away from simplistic causal models toward an appreciation for the cumulative nature of risk would be an important step forward in the way we conceptualize intervention and support programs, concentrating them squarely on alleviating the substantial risk facing so many of society's families.

Research highlights
► The level of cumulative risk facing a family across the transition to parenthood is associated with markers of maternal caregiving cognitions, social support, well-being, and indicators of child cognitive functioning.
► After controlling for total level of cumulative risk, most single risk factors failed to predict later maltreatment, and the newborn cumulative risk index score provided more robust prediction of maltreatment through 1, 4, and 16 years-of-age.
► The cumulative nature of risk for maltreatment suggests that intervention and support programs focused on reducing risk in only one or a few domains are not likely to result in large shifts in the prevalence of maltreatment.