Background: One of the goals of the National Survey on Drug Use and Health (NSDUH) is to estimate serious mental illness (SMI) among adults aged 18 or older. Starting in 2001, NSDUH mental health estimates used a cut point methodology based on Kessler-6 (K6) developed by Kessler and colleagues, using a sample of 155 respondents from the Boston area. Subsequent Substance Abuse and Mental Health Services Administration (SAMHSA) research concluded that K6 alone was not adequate for the model-based estimation of SMI. SAMHSA undertook the Mental Health Surveillance Study (MHSS), which included a clinical diagnostic interview to assess the presence of mental disorders and functional impairment, and was a follow-up to NSDUH. SAMHSA developed a prediction model to estimate SMI using the 2008 MHSS data, and the model was reevaluated in 2012 based on 2008 to 2012 MHSS data. The purpose of this report is to summarize current and prior approaches and to document the research conducted to revise the 2012 strategy for estimating the prevalence of mental illness.
Method: This report compares the estimates of SMI based on the direct MHSS data, the 2008 prediction model, and the 2012 prediction model. There were 750 clinical interviews in 2008 and a combined 5,000 clinical interviews in the 2008 to 2012 MHSS.
Results: This report demonstrates that the revised 2012 prediction model produces SMI estimates that are more accurate because they are more consistent with the direct estimates from the clinical interview.
Conclusion: SAMHSA has made improvements to produce more accurate estimations of SMI.