The findings in this report are subject to at least six limitations. First, AMI is an overall measure for DSM-IV disorders and cannot be separated into specific categories, whereas prevalence of smoking can differ among persons with various mental illness diagnoses (2). However, the estimate for the prevalence of AMI reported here is comparable to estimates from other national surveys (6). Second, estimates of smoking were self-reported and not validated by biochemical tests. Although studies of self-reported smoking might yield lower prevalence estimates than studies of serum cotinine (a breakdown product of nicotine) (19), it is unlikely that underreporting would substantially change the estimates reported. The estimates for current smoking in the population overall reported from NSDUH are higher than estimates from other national surveys, such as the National Health Interview Survey (NHIS) (19.0 in 2011), because of variations in the data collection methods and measures used to define current smoking. Nonetheless, both surveys have reported similar trends for current smoking among adults (7,8). Third, this report does not include persons residing in mental health residential communities, for whom smoking
