Date on Master's Thesis/Doctoral Dissertation


Document Type

Doctoral Dissertation

Degree Name

Ph. D.


Epidemiology and Population Health

Committee Chair

Hornung, Carlton


Mentally ill children--Diagnosis; Psychological tests for children; Child analysis


This dissertation is an applied research study to examine the performance of the K6 scale that measures serious emotional disturbance (SED) among adolescents. As the K6 was included in the 2012 administration of the Kentucky Incentives for Prevention Survey (KIP), three specific aims are included: 1) confirm the unidimensional structure of the K6 among adolescents using factor analysis; 2) define the prevalence and correlates of adolescent SED among Kentucky students; and 3) utilize latent class analysis (LCA) to empirically derive clinically relevant subtypes of adolescents with SED. Of the 122,718 students who completed the KIP in 2012, approximately 89% provided complete data for the K6 (n=108,736). Both principal axis and confirmatory factor analysis supported the unidimensional structure of the K6. Using the unweighted scoring algorithm (i.e., 13+) to screen for SED, the 30-day prevalence for Kentucky was 13.9% in 2012. Grade, gender, race/ethnicity, and family structure emerged as significant social and demographic predictors of SED. Among students with SED, the prevalence rates for substance abuse, antisocial behavior, role impairments, and peer victimization were significantly higher than those without SED. Four distinct subtypes of SED emerged from the LCA, varying by both symptom type and severity: mixed moderate risk, mixed high-risk, anxious moderate risk, and depressed high risk. Grade, gender, race/ethnicity, family structure, substance abuse, antisocial behavior, role impairments, and peer victimization were all significant predictors of class membership, although the magnitude of these effects were stronger for the two higher risk groups. The next steps include validation of the K6 on other state surveys that use school and community-based samples of adolescents, generation of cross-state comparisons, and the implementation of validated statistical approaches to generate more precise SED estimates, especially when gold standard diagnoses are not available. These results indicate the K6 is particularly useful for inclusion in large epidemiologic surveys like the KIP that have limited space and logistics that demand timely administration. Thus, this dissertation provides a foundation for increased epidemiologic infrastructure in Kentucky through the timely surveillance of SED.

Included in

Epidemiology Commons