Research That Matters (January 17 - 20, 2008)

Purpose: Patterns of mental health utilization among a community (non-clinical) population of youth is often difficult to examine because of the complexity in assessing symptomatology. Previous studies that looked at community samples and mental health utilization established psychopathology with such methods as cumulative scores of self-reported symptom measures or teacher and parent ratings of behaviors. Most of these measures were not psychometrically tested and for many, the symptom categorizations (e.g., depression) were often imposed by the researcher, creating a somewhat biased assessment of psychopathology. Additionally, the sample sizes of most of these studies were small, limited to either a region or a particular school system, especially if the study incorporated primary data collection (Oland & Shaw, 2005). The present study investigated patterns of psychopathology in a national representative sample of adolescents and examined how it affected mental health utilization. Unlike previous studies, the patterns of psychopathology was established through a data driven method (i.e., latent class analysis), allowing the data to inform the researcher on how these patterns fell and which indicators are most effective in measuring symptomatology. Specifically, using latent class analysis with a distal outcome, the study: -Assessed the prevalence of internalizing and externalizing symptoms in a non-clinical population of youth. -Investigated how these latent constructs affect the use of mental health services. Methods: Data from Wave I and Wave II of the National Longitudinal Study of Adolescent Health (Add Health) were used for this analysis. Although two Waves of data were examined, the study was cross-sectional. The only information derived from Wave II was whether youth utilized mental health services and this was gathered to establish continued mental health care. After merging databases and creating the necessary variables, the final analytical sample was 13,339. The data was then submitted to a latent class analysis with a distal outcome using MPlus, a multilevel modeling program. Results: Examining both class probabilities and conditional probabilities, the final and best fit model was that with four classes of symptomatology that were later named: internalizing symptoms, externalizing symptoms, a combination of both (i.e., co-occurring), and a normative class. A closer observation of variations by race and ethnicity, gender, and age found that Hispanic youth and boys were more likely to be in the co-occurring and the externalizing classes than white, non-Hispanic youth and that older youth were more likely to be in the internalizing class. Additionally, youth in the internalizing group had the highest probability of continuing care. Implications: The importance of appropriate diagnosing is paramount in this population because of the development aspect of adolescence, for example, is the behavior due to actual illness or a just a developmental stage? Additionally, differences were seen among youth of different racial and ethnic backgrounds and gender. Therefore, social workers must keep in mind not only the developmental stage of the individual but also the individual's culture and how that culture perceives illness.