Melissa Azur

Multiple imputation by chained equations: what is it and how does it work?

Multivariate imputation by chained equations (MICE) has emerged as a principled method of dealing with missing data. Despite properties that make MICE particularly useful for large imputation procedures and advances in software development that now make it accessible to many researchers, many psychiatric researchers have not been trained in these methods and few practical resources exist to guide researchers in the implementation of this technique. This paper provides an introduction to the MICE method with a focus on practical aspects and challenges in using this method. A brief review of software programs available to implement MICE and then analyze multiply imputed data is also provided. Copyright

Multiple Imputation With Large Data Sets: A Case Study of the Children's Mental Health Initiative

Multiple imputation is an effective method for dealing with missing data, and it is becoming increasingly common in many fields. However, the method is still relatively rarely used in epidemiology, perhaps in part because relatively few studies have looked at practical questions about how to implement multiple imputation in large data sets used for diverse purposes. This paper addresses this gap by focusing on the practicalities and diagnostics for multiple imputation in large data sets. It primarily discusses the method of multiple imputation by chained equations, which iterates through the data, imputing one variable at a time conditional on the others. Illustrative data were derived from 9,186 youths participating in the national evaluation of the Community Mental Health Services for Children and Their Families Program, a US federally funded program designed to develop and enhance community-based systems of care to meet the needs of children with serious emotional disturbances and their families. Multiple imputation was used to ensure that data analysis samples reflect the full population of youth participating in this program. This case study provides an illustration to assist researchers in implementing multiple imputation in their own data.

Differences in Mental Health Service Sector Utilization among African American and Caucasian Youth Entering Systems of Care Programs

Differences in unmet need and access to services between African American and Caucasian youth have been established; less is known about differences in specific mental health service sectors. This study examined differences in past year outpatient, school-based, day treatment and residential/inpatient service utilization among African American and Caucasian youth (n = 3,649) entering a federally funded system of care program. Random effect logistic regression models were implemented to examine the relationship between race and past year service utilization. Analyses revealed that African American youth were less likely than Caucasian youth to have utilized school-based and residential/inpatient mental health services in the past year. Findings suggest that racial disparities exist in service use for certain types of service sectors and highlight the importance of understanding and identifying individual, family, and community factors that contribute to disparities in service utilization.

The Potential to Reduce Mental Health Disparities Through the Comprehensive Community Mental Health Services for Children and Their Families Program

Few service systems are currently in place with the explicit purpose to reduce youth mental health disparities across socioeconomic status and race–ethnicity, despite substantial interest by the federal government and other institutions to redress health disparities. This study examines the potential for the Comprehensive Community Mental Health Services for Children and Their Families Program to address health disparities, even though this program was not explicitly designed for disparity reduction. Specifically, this study examines whether program sites disproportionately provide services within their catchment areas for youth who come from poor families, who are Black, and who are Hispanic. Data for this study come from 45 sites and 19,189 youth who were enrolled in program sites from 1997 to 2005. Meta-analysis was used to generate Forest plots and to obtain single, pooled estimates of risk ratios and their standard errors across all Children’s Mental Health Initiative communities. The results indicate that in comparison to the targeted catchment area (a) the percentage poor youth in the programs was almost three times higher, (b) the percentage Black in the programs was about twice as high, and (c) the percentage Hispanic in the programs was about the same. These results indicate that the program successfully reaches disadvantaged youth and can bring substantial infrastructure to address youth mental health disparities. In fact, to the extent that the program successfully improves mental health among enrollees it may be serving as one of the largest initiatives to redress health disparities, although its role in disparity reduction is not widely recognized.

Guided preparedness planning with lay communities: enhancing capacity of rural emergency response through a systems-based partnership

INTRODUCTION Community disaster preparedness plans, particularly those with content that would mitigate the effects of psychological trauma on vulnerable rural populations, are often nonexistent or underdeveloped. The purpose of the study was to develop and evaluate a model of disaster mental health preparedness planning involving a partnership among three, key stakeholders in the public health system. METHODS A one-group, post-test, quasi-experimental design was used to assess outcomes as a function of an intervention designated Guided Preparedness Planning (GPP). The setting was the eastern-, northern-, and mid-shore region of the state of Maryland. Partner participants were four local health departments (LHDs), 100 faith-based organizations (FBOs), and one academic health center (AHC)-the latter, collaborating entities of the Johns Hopkins University and the Johns Hopkins Health System. Individual participants were 178 community residents recruited from counties of the above-referenced geographic area. Effectiveness of GPP was based on post-intervention assessments of trainee knowledge, skills, and attitudes supportive of community disaster mental health planning. Inferences about the practicability (feasibility) of the model were drawn from pre-defined criteria for partner readiness, willingness, and ability to participate in the project. Additional aims of the study were to determine if LHD leaders would be willing and able to generate post-project strategies to perpetuate project-initiated government/faith planning alliances (sustainability), and to develop portable methods and materials to enhance model application and impact in other health jurisdictions (scalability). RESULTS The majority (95%) of the 178 lay citizens receiving the GPP intervention and submitting complete evaluations reported that planning-supportive objectives had been achieved. Moreover, all criteria for inferring model feasibility, sustainability, and scalability were met. CONCLUSIONS Within the span of a six-month period, LHDs, FBOs, and AHCs can work effectively to plan, implement, and evaluate what appears to be an effective, practical, and durable model of capacity building for public mental health emergency planning.

Behavioral and Emotional Strengths among Youth in Systems of Care and the Effect of Race/Ethnicity.

Behavioral and emotional strengths are important to consider when understanding youth mental health and treatment. This study examined the association between youth strengths and functional impairment and whether this association is modified by race/ethnicity. Multinomial logistic regression models were used to estimate the effects of strengths on impairment and examine whether race and ethnicity modified this relationship in 8,129 Caucasian, African American, Hispanic, and American Indian/Alaskan Native youth between 5 and 18 years of age. Results suggest that youth with average and above-average strengths were less likely to have impairment compared with youth with below-average strengths. Race and ethnicity modified this relationship in both expected and unexpected ways. Among youth with average and above-average strengths, racial and ethnic minority youth appear to have more impairment than Caucasian youth. However, among youth with below-average strengths, racial and ethnic minority youth have less impairment than Caucasian youth. Findings highlight the importance of incorporating strengths-based approaches in youth mental health treatment and the need for further research to understand the specific nature of strengths as it effects impairment across racial/ethnic groups. Implications and recommendations are discussed.

Gender Differences in Patterns of Child Risk across Programmatic Phases of the CMHI: A Multiple Group Latent Class Analysis (LCA)

Data from 18,437 children enrolled in the national evaluation of the Children’s Mental Health Initiative between 1994 and 2005 were used to examine the evolution of patterns of risk among boys and girls across funding phases using multigroup latent class analysis. Consistent with previous research, this study identified four subgroups of children with similar patterns of child risk. Membership to these risk subgroups varied as a function of age and was associated with differences in impairment levels. Changes in the distribution of boys and girls in the risk classes suggest that, over time, an increasing proportion of boys have entered the system of care program with complex histories of risk. Information on children’s exposure to child risk factors can aid policy makers, service providers, and clinicians in identifying children who may need more intensive services and tailoring services to their needs.

Violent and Nonviolent Delinquent Behavior Among Caucasian and Hispanic Youth in Mental Health Systems-of-Care Programs

The prevention and reduction of violent and nonviolent delinquent behavior is a public health priority. In recent years, attention has turned to ethnic disparities in delinquent behavior in efforts to target programs to youth at greatest risk. This study uses longitudinal data from a national evaluation of a federally funded children’s mental health systems-of-care program to examine self-reported delinquent and violent behavior among Caucasian and Hispanic youth who entered mental health services between 1997 and 2006. The results from hierarchical logistic regressions indicate that delinquent, but not violent behavior, decreased over time. There were no significant ethnic differences between these youth in either delinquent or violent behavior. Sociodemographic, clinical, and school characteristics were associated with delinquency and offer suggestions on where to intervene. Given that mental health problems and delinquent behavior frequently co-occur, juvenile justice and mental health agencies would benefit from close partnerships to help youth address these problems.