Examining the Impact of an Integrated Early Head Start & Children and Youth Services Program

• In Event: Poster Session 04
  In Poster Session: PS 04 Topic - Prevention and Interventions

Abstract

In 2018, 4.3 million abuse referrals were made to Child Protective Services involving 7.8 million children throughout the US; 56% of these reports were screened in for additional investigation (US Department of Health and Human Services, 2018). Maltreatment is particularly damaging in the first three years of life as it may cause long term negative impacts on brain growth and development (Demers et al., 2019; Glaser, 2014). In 2002, the Office on Child Abuse and Neglect established the Early Head Start Child Welfare System Initiative. As a part of this initiative, SafeStart, an integrated Early Head Start and Children and Youth Services program, was established. This integrated program serves children between the ages of birth and 3 years.

Due to the needs of maltreated young children and families, outcomes of these types of programs need to be evaluated in terms of the timing of the services, child and parent maltreatment histories, and attrition; there are research gaps in regard to these contextual factors (Gedal, Douglass, Roche, Lavin, Ghazarian, & Perry, 2020; Love et al., 2002; Raikes et al., 2006; Roggman, Cook, Peterson, & Raikes, 2008). In order to address these gaps, this investigation evaluated the relative benefits of the SafeStart model for enrolled children.

Program data for 129 children were analyzed and included children’s age, date of enrollment into the program, length of enrollment in the program, and children’s scores on the E-LAP (Early Learning Accomplishment Profile; Glover, Preminger, & Sanford, 1995) in the areas of cognitive, language, and social emotional development upon exiting the program. Three independent ROC curve analyses (Pintea & Moldovan, 2009) were unable to determine the cut-off point for the age at which enrollment predicted optimal developmental skill attainment for the three E-LAP domains as the area under the curve (AUC) accuracy fell below or within in the low range (.39, .44, and, .56 respectively). These low AUC accuracies indicated that these three models were performing poorly and therefore, age of enrollment could not distinguish between those children who were proficient versus those who were developing on developmental milestone attainment as measured by the E-LAP.

Additionally, a series of three logistic regressions were conducted with age at enrollment, duration of enrollment, and the interaction of the two used as predictors. Results suggested that duration of enrollment significantly (b = .002, p = .017) predicted children’s attainment of language developmental milestones as measured by their final E-LAP, and that age at enrollment (b = .003, p = .041) and duration of enrollment (b = .004, p = .003) significantly predicted a child’s attainment of social emotional developmental milestones as measured by their final E-LAP. Taken together these results suggest that the longer a child is enrolled in this integrated program the more likely they are to attain expected developmental milestones. Discussion will focus on policy implications of these findings as well as the challenges associated with evaluating integrated program data.

Authors

• Laura Emily Wallace, University of Delaware
  Presenting Author

• Patricia Manz, Lehigh University
  Non-Presenting Author