Objective: The aim of the present study was to compare two versions of the Kessler 10-item scale (K10), as measures of population mental health status in New Zealand.
Method: A nationwide household survey of residents aged > or = 16 years was carried out between 2003 and 2004. The World Mental Health Composite International Diagnostic Interview (CIDI 3.0) was used to obtain DSM-IV diagnoses. Serious mental illness (SMI) was defined as for the World Mental Health Surveys Initiative and the USA National Comorbidity Survey Replication. Participants were randomly assigned to receive the 'past month' K10 or the 'worst month in the past 12 months' K10. There were 12 992 completed interviews; 7435 included the K10. The overall response rate was 73.3%. Receiver operator characteristic (ROC) curves were used to examine the ability of both K10 versions to discriminate between CIDI 3.0 cases and non-cases, and to predict SMI.
Results: Scores on both versions of the K10 were higher for female subjects, younger people, people with fewer educational qualifications, people with lower household income and people resident in more socioeconomically deprived areas. Both versions of the K10 were effective in discriminating between CIDI 3.0 cases and non-cases for anxiety disorder, mood disorders and any study disorder. The worst month in the past 12 months K10 is a more effective predictor than the past 1 month K10 of SMI (area under the curve: 0.89 vs 0.80).
Conclusions: Either version of the K10 could be used in repeated health surveys to monitor the mental health status of the New Zealand population and to derive proxy prevalence estimates for SMI. The worst month in the past 12 months K10 may be the preferred version in such surveys, because it is a better predictor of SMI than the past month K10 and also has a more logical relationship to 12 month disorder and 12 month service use.