Jud, A., Fallon, B., & Trocmé, N. (2012). Who gets services and who does not? Multi-level approach to the decision for ongoing child welfare or referral to specialized services. Children and Youth Services Review 34, 983-988.
This was a study of the services provided in child welfare cases after a report is investigated. Possible actions were ongoing agency service or referral to a specialized service, compared to no additional services beyond the investigation. A number of factors thought to be associated with the service decision were studied, including both case characteristics and agency characteristics. The multi-level strategy was based on a secondary analysis of the data from the Canadian Incidence Study of Reported Child Abuse and Neglect (Trocmé et al., 2008). The sample consisted of 15,980 cases from 111 agencies throughout Canada; in almost 60% of these cases some type of service was offered.
The following case characteristics were found to be significantly associated with the decision to provide services:
Any type of actual maltreatment either suspected or substantiated. The strongest association was with exposure to intimate partner violence.
Risk investigations involving caregiver and household concerns (e.g., few social supports, teen parenting, low socioeconomic status). These risks seemed to drive service decisions almost more than substantiation status.
Ethnic minority status was not found to be associated with the provision of services; this was surprising, since in many earlier studies ethnicity has been associated with increased placement in foster care. An unexpected finding was the marked effect of agency characteristics on referral rates; this was noted at both the provincial and agency levels. The proportion of cases referred ranged from 30% to 70% for provinces; from 15% to 77% for individual large agencies. Factors accounting for these differences remain largely unexplained; possible reasons might be differences in provincial legislation or funding, or in agency organization (e.g., supervisory support, working climate or staff retention). The authors argue this warrants further investigation.
This analysis was based on very sophisticated multi-level statistical procedures, including univariate and multivariate logistic regressions, using specialized statistical software (Mplus 5). Given the large sample size, the findings reported above were at a very high level of statistical significance. However, the authors pointed out a number of limitations. Longer-term service outcomes could not be measured and analysis was only of the referrals stage of the process. Very few agency-level variables were available in the analysis. In future studies it would be useful if more variables at both provincial and individual-agency levels could be included. In the latter case, this could potentially provide information to support improvements in practice.