Screening for serious mental illness: methodological studies of the K6 screening scale.

The K6 scale is a shortened version of the K10, a 10-question scale originally developed to provide an efficient population-level screen for serious mental disorders (SMI) in the USA. Evidence that the six-item shortened version performed as well as the original 10-item version, coupled with strong psychometric properties, led to rapid dissemination and replicated validation of the K6 in a number of other countries around the world. Based on these results, the K6 is now often included in large general-purpose government health tracking surveys in a number of different countries. Until now, though, the scoring rules for the K6 in these surveys were inconsistent. The first paper in this special issue introduces the K6 scale and summarizes the results of a series of investigations to resolve these inconsistencies by providing optimal scoring rules for the K6 in 14 countries. Subsequent papers explore the usefulness of the K6 to screen for serious emotional disturbance among adolescents and report findings from validation efforts based on independent diagnostic assessments as well as of other measures of impairment and disability (World Health organization Disability Assessment Scale). Finally a highly innovative analysis using a Bayesian multilevel modeling approach is presented, designed to estimate the prevalence of SMI in small areas, such as cities, states, or schools, from surveys carried out in a larger population that includes only relatively small samples of respondents in each of the areas in which prevalence estimates are to be made. Taken together, these studies demonstrate that the K6 is an efficient and useful screening tool. The psychometric and methodological explorations will hopefully stimulate additional interest in the use of short screening scales in large-scale general health surveys to supplement the more in-depth information obtained in periodic psychiatric epidemiological surveys on the basis of diagnostic interviews.