Predicting Chronic Neglect: Understanding Risk and Protective Factors for CPS-Involved Families

Background:  In 2014, over 75% of child protection cases were referred for neglect, and children who were neglected had the highest rates of re-reporting (U.S. DHHS, 2016). Although neglect remains under-studied, research indicates the effects of chronic neglect may create an accumulation of harm for child well-being, including impact on early brain development, emotional regulation, and cognitive development (Hildeyard & Wolfe, 2002; Perry & Pollard, 1998).  Early identification of factors related to chronic neglect is needed.  Many child welfare jurisdictions have adopted standardized risk assessments, which may improve assessment of risk for future harm compared to clinical or caseworker judgement (Baird & Wagner, 2000).  However, critics suggest that these tools are not always implemented with fidelity and consistency (Gillingham & Humphreys, 2010). This current study aims to 1) identify risk and protective factors for families who experience chronic neglect, and 2) examine the use of the risk assessment tools in predicting subsequent chronic neglect.

Methods: Data were derived from administrative records in a large, diverse, Northeastern county based on screened-in allegations of neglect from 2009-2010, and included two sets of families: those who had at least four more reports of neglect before 2014 (Chronic Neglect cases, n=430) and those with no additional reports (n=1,644). Data included information from a risk assessment tool, similar to what is used in other jurisdictions, encompassing data on several psychosocial risk factors related to abuse and neglect, in addition to theoretically important protective factors.

Results: Bivariate analysis highlighted important risk factors associated with chronic neglect, including family size, number of allegations, prior substitute care, a child under one, and caregiver younger age, domestic violence, drug use, mental health problems, and cognitive impairment. Logistic regressions tested a multivariate model (Nagelkerke R2=0.18, classification=81.1%) to predict subsequent chronic neglect; all factors retained significance in the full model, except drug use. Theoretically important protective factors were not significantly related to chronic neglect in bi- or multivariate analyses. Risk scores (derived from built-in assessment scoring) were also not significantly related to chronic neglect.

Implications:  Multiple factors significantly predicted subsequent chronic neglect, bolstering prior research that showed these families exhibit substantial stressors, including domestic violence and caregiver behavioral health problems. Widely used risk assessment tools were not developed to predict chronic neglect, but theory and empirical evidence are accumulating to suggest that prediction is possible.  Findings from this current study also support critiques that risk assessment tools are not currently being implemented with fidelity, as indicated by the protective factors that were endorsed for almost every family and financial stressors endorsed for very few, although this is improbable based on the literature.  Prior research found staff resistance to using risk assessments, and found when workers are faced with significant time constraints, risk assessments may be only superficially completed (Hughes & Rycus, 2008).  Implications of this current study also suggest that child welfare jurisdictions should pay closer attention to the implementation of risk assessment tools, including the need for sufficient training, on-going coaching, and supervision of the use of the tool.