Mental Health Estimates Computed Directly from the Clinical Sample of the Mental Health Surveillance Study and Measures of their Standard Errors

• Activity Number: 83
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Phillip Kott*+ and Art Hughes and Jonaki Bose and Sarra Hedden
• Companies: RTI International and SAMHSA and SAMHSA and SAMHSA
• Keywords: Nearly pseudo-optimal ; Standard error ; Bounded weights ; WTADJX ; Mental health

Abstract

A clinical follow-up study to the National Survey on Drug Use and Health (NSDUH) collected information on specific mental disorders among adults that could be used to provide national and state estimates of serious mental illness. Specifically, a nationally representative subsample of adult respondents to the NSDUH was interviewed by trained clinicians over the telephone using a psychiatric diagnostic interview between 2008 and 2012. In order to estimate the prevalence of mental health disorders among adults in the U.S., weights were created for the clinical subsample. The weighting procedures included a nearly pseudo-optimal "poststratification" to non-mutually exclusive control totals from the NSDUH interview. This use of data from the entire NSDUH sample in weight creation resulted in estimates with increased accuracy. Both the nearly pseudo-optimal poststratification and improved standard error measures for the resulting estimates were completed using the WTADJX procedure in SUDAAN 11.