State level classification of serious mental illness: a case for a more uniform standard

This study reports a national survey of U.S. states that was conducted from July of 1999 through March of 2001. The lack of consistent data on serious mental illness (SMI) provided the impetus for this study. Data was collected through a survey on states' definitions of SMI, on demographic information for patients with SMI, and on total annual per capita expenditures for SMI. Based on a 100% response rate, we found considerable variation among states in the definition used for SMI and the records kept on patients with SMI. This paper also involves a state-level statistical analysis of factors that may influence rates of per capita expenditures for SMI. The main finding using regression analysis was that per capita income and state definitions of mental illness that included DSM-III, DSM-IV, and ICD-9-CM diagnoses are significant and positively associated with a state's per capita expenditures for SMI. An additional finding is that accounting for all of the above factors, there still remains significant differences across major census divisions in per capita expenditures for the seriously mentally ill. Another major finding is that more consistent data collection is needed to take an epidemiological approach toward understanding the social conditions that contribute to SMI.