Methods of Handling Missing Data in a Cluster-Randomized Community-Partnered Participatory Research Project to Reduce the Burden of Depression

• Activity Number: 256
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Mental Health Statistics Section
• Author(s): Lingqi Tang*+ and Thomas R. Belin and Susan Stockdale and Difan Zhao and Elizabeth Dixon and Jim Gilmore and Felica Jones and Klap Klap and Kenneth B. Wells and Loretta Jones
• Companies: UCLA-Center for Health Services & Society and UCLA Department of Biostatistics and VA Greater Los Angeles HCS, Sepulveda and UCLA Center for Health Services and Society and QueensCare Health and Faith Partnership and Behavioral Health Services and Healthy African American Families II and VA Greater Los Angeles HCS, Sepulveda and UCLA Center for Health Services and Society and Healthy African American Families II
• Keywords: weighting ; imputation ; nonresponse ; cluster-randomized trial

Abstract

Community Partners in Care (CPIC) study is a cluster randomized comparative effectiveness trial of community engagement and planning or program technical assistance to address depression disparities. Of 4440 clients screened from 93 programs, 1322 were eligible to the study; 1246 enrolled and 1018 completed baseline or 6 month follow-up. In this paper, we describe weighting and multiple imputation techniques used to handling missing data. We summarize the results of alternative modeling methods for CPIC data.