Given the limited ability of informatics-based assessment technologies to reach individuals with serious mental health conditions, this study evaluated the feasibility and data quality of mail-out healthcare surveys in this population to assist in measure selection for a multi-site controlled implementation trial. Veterans were randomly selected from those who had received services at a mental health clinic in the Department of Veterans Affairs, and were randomly assigned to one of three questionnaire lengths. Survey length (48-127 items) was not associated with differences in response rate, percent of items missing, or data quality. However, internal consistency reliability was variable among scales and survey lengths. Additional analyses indicate the above measures of survey data quality may differ among respondents who are non-white and younger and have psychotic disorders. These results can inform survey protocols to ensure maximal representation of this vulnerable population in health planning and policy assessment.