A Finite Mixture Analysis of the Validity of the MAYSI-2 among Serious Juvenile Offenders

Background and Purpose: Epidemiological findings have shown that youth in the juvenile justice system are beset with a range of mental health and substance use problems. Therefore, it is critical to devise efficient screening measures that are reliable and valid in order to achieve the most effective service delivery for these youth. As such, this study investigated the validity of the Massachusetts Screening Inventory-Second Version (MAYSI-2) by using finite mixture modeling to form homogeneous subgroups and assess whether MAYSI-2 scores did indeed differentiate treatment needs based on external measures of behavior and prescribed medications.

Methods: Data was derived from 836 serious delinquents committed to the California Youth Authority. The racially diverse sample (46% Hispanic, 28% African American, 17% White, 9% other) is 81% male with a mean age of 16.9 years (SD=1.1; range 12.6- 20.4). Youth completed the MAYSI-2 screening measure,comprised of eight subscales and 52 items (alpha = 0.79), and then were followed to determine whether they subsequently were placed in mental health programs, were prescribed medications used to treat serious mental health problems, and/or were identified by staff as requiring these services. The assessment battery included four self-report questionnaires that were administered during the educational testing phase of the clinic process with 8 to 15 wards at a time by casework staff at the reception center. In addition to official records and psychiatric information, the data used for this study includes demographic variables commitment offense, prior offending, institutional misconduct, prescriptions, Global Assessment of Functioning, Weinberger Adjustment Inventory (WAI), Child Behavior Checklist (CBCL) suicide risk/suicide risk assessment, and treatment.

Results: The three-class solution provided the best statistical and conceptual fit to the data. The three classes consisted of 1) low distress (45.5%; n= 380), 2) moderate distress (39.1%; n = 327), and 3) high distress (15.4%; n = 129). Chi-square tests confirmed that the MAYSI-2 does not discriminate on the type of offense. Additionally, there is no significant difference in age between the different classes. However, there are differences in the number of different types of psychotropic drugs prescribed between the different classes. Between the classes, there are significant differences in anti-depressants, mood stabilizers, and anti-psychotic with increased MAYSI-2 scores paralleling increased medications. Further, all of the three groups were significantly different in each of the WAI and CBCL subscales.

Implications for policy and practice: Study findings provide further evidence of the validity of the MAYSI-2 as an effective screening tool of mental health problems for juvenile offenders. Furthermore, findings also have policy relevance in that they converge with other studies that show that juvenile offenders are a heterogeneous population. For example, low distress youth may be diverted while high distress youth may need more intensive follow-up care. Because the MAYSI-2 is low in cost, simple to administer, and can be employed on a large scale, juvenile justice facilities should utilize this measure as a means to enhance treatment delivery.