Refining school mental health screening using mixture modeling : evaluation for use in applied settings

Abstract

Universal socioemotional and behavior screening procedures continue to gain momentum and adoption in schools; however, a major research-to-practice gap remains in translating these data into meaningful interventions for identified students. Despite advances in methodology in school psychology research, parallel efforts have not been made to translate screening data into highly individualized, data-based interventions. The current study sought to apply person-centered analytic strategies to an existing youth mental health screening tool (EIS-Student; Reinke et al., 2020) to determine the extent to which empirically-derived latent subgroups could reveal clinically relevant conclusions in universal screening data. The sample was drawn from existing universal screening data from fall 2018 EIS-Student participants across nine high schools in a single Midwestern county (n = 5,860). Results revealed a stable 7-factor structure for the EIS-Student as well as a 5 latent profiles. Profiles corresponded with meaningfully different scores on subscales of mental health risk from the EIS-Student, and significant relationships were observed between profile membership and academic and behavioral outcomes. Characteristics of each profile are discussed for possible clinical conceptualization and applications within a socioemotional screening program. Implications for universal screening efforts are discussed, including the use of mixture modeling to identify subgroups of student need, the value of person-centered analyses in school intervention decision-making, and considerations for key stakeholders in applied contexts.