Study 1
Chaiyachati and colleagues (2018) used a quasi-experimental design with propensity score matching to assess the effect of the Nurturing Families Network (NFN) home-visiting program on child maltreatment, including neglect or physical abuse and out-of-home placements. The program was implemented in individual family homes across the State of Connecticut.
Socially high-risk families who had a child born between January 1, 2008, and December 31, 2011, were eligible for this study. During this period, NFN intake coordinators sought to screen every first-time mother in the state either prenatally or in the early postnatal period at obstetricians’ offices, community partner sites, and birthing hospitals to identify these families. The Revised Early Identification (REID) screening instrument, adapted from the Early Identification Instrument (Duggan et al. 2000) was used to identify socially high-risk births by assessing 17 factors associated with an increased risk of child maltreatment, including teen motherhood, single motherhood, social isolation, and housing instability. Families who had any 3 of the 17 risk factors, or who had any history of substance abuse, psychiatric care, depression, or marital/family problems, were considered to be high risk and therefore eligible for home visiting. NFN administrative records were then reviewed, and eligible families that participated in NFN were considered for participation in the study treatment group. The comparison group included families who either declined participation or were unable to enroll in NFN services. Eligible families were excluded from the treatment group if there was missing information regarding the mother’s or the child’s date of birth.
Of the eligible families, 4,560 were included in the final, matched sample analysis: 2,280 were in the home-visiting treatment group, and 2,280 were in the comparison group. Demographics and reported risk factors were used to match treatment and comparison groups using propensity-score matching. After matching, no statistically significant differences were found between families in the home-visiting treatment group and families in the comparison group on any measured risk factor. The average age of mothers was nearly the same in both groups: 22.3 years old in the home-visiting treatment group, 22.2 years in the comparison group. Nearly 90 percent of the treatment group families had mothers who were single, separated, or divorced, compared with 88.8 percent of the families in the comparison group. In the treatment group, 41.7 percent of mothers had less than or equal to 12 years of education, compared with 42.4 percent of mothers in the comparison group. Mothers in both the treatment and comparison groups had histories of psychiatric care (22.4 percent versus 21.2 percent, respectively); depression (37.1 percent versus 35.9 percent, respectively); and substance abuse (14.7 percent versus 14.9 percent, respectively). The study authors did not include race as a variable in their analysis, as this information was only available for families with a Child Protective Services (CPS) report.
CPS data were collected for the study sample from January 1, 2008, to December 31, 2013. Matches were identified between CPS records and NFN records by both the mothers’ and children’s birthdates. CPS records included the child’s age at the time of the CPS investigation, whether the report was substantiated, the type of suspected and substantiated maltreatment, and whether the child was placed in out-of-home custody. The length of follow up was considered the time between a child’s birth and the end of the CPS monitoring window (December 31, 2013).
Primary outcomes of interest included 1) investigated reports of maltreatment (categorized as physical, medical, emotional, educational, or moral neglect; and physical abuse); 2) substantiations of maltreatment; 3) out-of-home placements; 4) the time to each of these events; 5) types of maltreatment for first substantiation; 6) duration of out-of-home placement; and 7) percentage of out-of-home placements with reunification. The CrimeSolutions review of this study focused on the investigated reports of maltreatment; the substations of maltreatment, neglect, and physical abuse; and out-of-home placements. The other outcomes (time to substantiated maltreatment, time to out-of-home placement, duration of out-of-home placement, and family reunification following out-of-home placement) are considered subgroup analyses because these outcomes were not examined for the entire study sample. The time to substantiated maltreatment outcome was only examined for youth in the home-visiting treatment group and comparison group who had substantiated reports of maltreatment, and time to out-of-home placement, duration of out-of-home placement, and family reunification following out-of-home placement outcomes were examined only for treatment and control group youth with out-of-home placements.
To analyze the outcomes, bivariate associations between NFN participation and CPS outcomes in the matched sample were assessed using chi-squared tests. To account for the timing at which substantiations and out-of-home placements occurred, Cox proportional hazards regression was used to generate hazard ratios; the time to first substantiation and out-of-home placement were assessed using Kaplan-Meier survival analysis.
