A School-Based Intervention to Reduce Stigma & Promote Mental-Health Service Use

Overview

This is a school-based field experiment conducted in sixth grade classrooms to evaluate a multifaceted intervention designed to change attitudes and behaviors regarding mental illnesses. The research tests hypotheses as to whether alone or in combination interventions that are 1) a curriculum-based in-class presentations, 2) contact-based with a person who has experienced a mental illness, or 3) or based on educational materials distributed in classes improve knowledge/attitudes and encourage help seeking for mental health problems in a follow up study lasting two years.

Purpose. Research has documented that stigma and discrimination are painfully present in the lives of people with mental illnesses and their families, blocking opportunities, compromising self-esteem and keeping people from accessing helpful treatments The proposed project seeks a rigorous test, with long term follow-up, of an intervention in sixth grade classrooms that is designed to 1) improve knowledge, attitudes and beliefs about mental illnesses, 2) change behaviors toward people with mental illnesses and 3) facilitate appropriate help seeking for mental health problems. This study evaluates the short- and long- term effectiveness, both individually and in combinations, of 1) the Eliminating the Stigma of Differences curriculum, 2) contact with a young adult with a mental illness, and 3) the use of supplemental materials, including posters and other media that reinforce an anti-stigma message. Design The study involves two phases. Phase I is a pre-post design to assess changes in knowledge, attitudes, and behaviors in sixth-grade students from before the intervention to 3 weeks after its conclusion. Phase II is a longitudinal assessment of Phase I participants who were willing to participate in home-visit assessments at 6, 12, 18 and 24 months post intervention. The study was designed as a fully crossed two by two by two factorial cluster randomized behavioral field experiment. Sixteen schools with separate zip codes from an urban school district in Texas were ranked according to performance on the statewide standardized assessment of math, English, and science. We randomly assigned the top eight schools to one of eight cells; the bottom-ranked eight schools were then assigned to a cell in the reverse order so that, for example, the top- and bottom- ranked schools were paired. Each cell (two schools in each) was randomly assigned to a study condition. Before the study began, two schools dropped out for non-study-related reasons. For this reason the study was repeated during a second academic year with a new set of sixth-grade students in five of the original schools chosen because they had demographic characteristics similar to the lost schools. Thus a total of 19 sixth-grade classes from 14 schools were included in the study. Recruitment Procedures. In each school a member of the research team met with students to explain both phases of the study and provide students with information to take home to their guardians. The information included a detailed description of the study, contact numbers for Principal Investigator's, and a request for their children's participation in the study (both phases). In addition, consent/assent forms and a demographic questionnaire were included for guardians who wished to participate to complete. A form declining their child's participation was also included. A self-addressed stamped envelope was provided for return of either the questionnaire and consent/assent forms or the declination form. Youths were not included in the study without signed consent/assent forms. Phase I pre- and post- test -instruments were self-administered in physical education classes on laptop computers from February to May and September to December 2012. Each class received its assigned combination of interventions within one week of pretesting. Post-test instruments were administered within a week after the intervention all students in the classroom were exposed to the assigned intervention(s); only the students who provided consent completed the assessment instruments. In the longitudinal Phase II component of the study, members of the research team arranged to meet participants in their homes where participants completed surveys on laptop computers. Interventions Curriculum. Eliminating the Stigma of Differences is a three-module, three-hour curriculum delivered by teachers over a three- to six-day period. Module 1 of the intervention addresses the bases on which we judge others to be different; the definition, causes, and consequences of stigma, including for students themselves; ways to end stigma; a definition and description of mental illness; causes of mental illness; treatment for mental illness; barriers to help seeking; how stigma applies to mental illness; and sharing personal experiences with people who have mental illness. Modules 2 and 3 address attention deficit hyperactivity disorder, anxiety disorders, depression, schizophrenia, and bipolar disorder and include descriptions of the disorders, discussion of causes and treatments, and content that stimulates empathy. Fidelity to the curriculum content, quality of delivery, and level of student engagement were assessed by two independent observers in each classroom with a 60-item measure based on two existing tools with good psychometric properties . Possible scores range from 60 to 240, with scores of 148 to 192 considered "good" and scores of 193 to 240 "high fidelity." Contact. Two young adults with a history of mental illness each made a ten-minute in-class presentation (20 minutes total) describing onset and course of their symptoms, hospitalizations and treatments, their feelings about the illness, coping strategies, and impact of the illness on social relationships and functioning at school, college, and work. Based on previous research, the talks were constructed to moderately disconfirm stereotypes of mental illness. Printed materials. Teachers prominently displayed posters in the classroom for two weeks and provided students with bookmarks. The materials focused on individuals' personal traits and abilities as opposed to language that labels a person as "mentally ill." Samples Phase I Sample. A total of 751 students (60% of those invited) agreed to participate in Phase I. Based on calculations conducted before the study was implemented, the minimum sample size to conduct analysis was deemed to be 242, with alpha = .05, power = .95, and a moderate effect size. The achieved Phase I sample size of 721 is therefore more than adequate for analysis. Phase II Sample. Of the 751 consenting in Phase I, 479 (64%) agreed to participate in Phase II. and eventually 416 (87%) of those who consented were interviewed at least once during follow up. Of the 416 who participated in any follow up, 99% (N=412) participated in the 6-month interview, 89% (370) at 12 months, 81% (338) at 18 months and 75% (312) at 24 months. Based on pre-study calculations, the minimum sample size to conduct the analysis is 273, with alpha =.05, power = .95, and a moderate effect size . Outcome measures Described in Section 9 Covariates. A self-reported mental health symptoms checklist based on screening questions from the National Institute of Mental Health Diagnostic Interview Schedule for Children, Version IV, was administered to youths pre-intervention, 3 weeks post-intervention and at 18 and 24 months follow up. This compact symptom screen was used to identify mental health need so that we could assess whether interventions led to help seeking among youths with high need. Other covariates assessed included sociodemographic variables a 7-point scale assessing the closeness of contact with a person with mental illness and a reliable 6-item measure of social desirability bias for children. Statistical Analysis. Phase I Short Term Effects. We use analysis of covariance to test the effects of each intervention on the knowledge/attitudes and social distance outcomes. We included pretest values of the corresponding outcome measures in the main analyses to control for any pre-intervention group differences that randomization msy not have accounted for. We also conduct analyses including any personal characteristics that may have differed at baseline to determine whether adding these additional controls alters conclusions. Phase II Longitudinal Effects. The cluster randomized longitudinal design involves 1) clustering of youths within classrooms and 2) assessments at multiple follow-ups within youths. In order to assess the extent of clustering at each level we will calculate intra-class correlation coefficients for each of the dependent variables. We will use Generalized Estimating Equations (GEE's) to account for the nesting of occasions within individuals. We examine whether there are significant differences by pre-intervention characteristics in the 8 cells of our intervention so that any differences might be accounted for by controlling these variables in our analysis. We begin analysis by testing a fully saturated model that allows for interaction between the interventions in our fully crossed design. We further test whether intervention effects persist across time by examining interaction between interventions and time periods. Finally, we explore whether the interventions appear to be more or less effective for groups of individuals as defined by gender, age, race/ethnicity and the educational level of the child's caregiver. Missing data. To address missing data, multiple imputation by chained equations in Stata 15.1 is used to impute missing values. We will impute twenty five data sets, conduct regression analyses on these twenty five imputations and then recombine them using Rubin's rules. Analyses will present results that impute all missing values except when the missing value is the dependent variable. Sensitivity analyses include 1) analyses without imputation (complete case analysis) and 2) analyses with imputation of all variables including dependent variables. The impact of missing data on study conclusions is assessed in these sensitivity analyses by determining whether conclusions vary according to these different specifications. Strengths and Limitations. Beyond addressing a critically important problem for people with mental illnesses, the strengths of this study include: 1) the multi-component intervention that allows not one but three different chances for success; 2) the development of the intervention within the ecology of the school environment that both helps assure its acceptability in that context and the likelihood of its broad dissemination in other contexts like it; 3) the implementation of the intervention during the important developmental period of early adolescence; 4) the assessment of not only knowledge and attitudes but help seeking behaviors; 5) the relatively long follow-up period of two years, 6) the inclusion of measures of youths' mental health, and 7) an ethnically and socioeconomically diverse sample. As with any experimental study, recruitment bias can be problematic. This is important because nonparticipating students or their families may hold particularly negative views about mental illness, possibly limiting the generalizability of our findings. Our study is limited in its use of self-report data for both attitudes and help-seeking behaviors. Our inclusion of a measure of social desirability bias and our ability to control for pre-intervention values of dependent variables (which would include any tendency toward biased reporting) somewhat mitigates our concerns about reporting bias. Loss to follow up, a common problem for longitudinal studies, is another limitation. To gauge how large of a problem it is we will assess whether differential drop out occurs by assigned intervention and other measured variables. Further we will use sensitivity analyses to impute and missing values for dependent variables so that all individuals who completed at least one wave are included in the analysis. Despite matching schools on test scores, our cluster randomization of classrooms could result in pre-intervention differences between groups on some of the baseline characteristics we measured. As a result, we adjust for covariates and pre-intervention measures of dependent variables.