Measuring psychological distress using the 12‐item general health questionnaire and the six‐item Kessler psychological distress scale. Psychometric comparison and equipercentile equating of the two scales

• Published in: International journal of methods in psychiatric research Vol. 33; no. 3; pp. e2033 - n/a
• Main Authors: Lundin, Andreas, Muwonge, Joseph Junior, Lalouni, Maria, Åhlén, Johan
• Format: Journal Article
• Language: English
• Published: United States John Wiley & Sons, Inc 01.09.2024; John Wiley and Sons Inc; Wiley
• Subjects: Adolescent; Adult; Aged; agreement; Anxiety; crosswalks; equating; Female; Health surveys; Humans; Male; Mental depression; Mental disorders; mental distress; Mental health; Middle Aged; Original; Psychiatric Status Rating Scales - standards; Psychological Distress; Psychometrics - standards; Public health; Quantitative psychology; Questionnaires; Surveys and Questionnaires - standards; Young Adult; mental distress; agreement; crosswalks; equating
• ISSN: 1049-8931; 1557-0657; 1557-0657
• DOI: 10.1002/mpr.2033

Objectives This study aimed to examine if the General Health Questionnaire (GHQ)‐12 and Kessler 6 (K6) assess the same underlying construct and to develop a score conversion table for the two scales. Methods A random sample of 4303 people who completed both the GHQ‐12 and K6 in 2021 were analyzed. Exploratory bifactor analysis evaluated if both scales measured the same construct, and Rasch analysis assessed item severities. The scales were transformed using Equipercentile equivalence for comparability and score conversion. Agreement was estimated with Cohen's Kappa coefficient, along with raw positive and negative agreement. Results We found that the two scales measure the same phenomenon to the extent that they can be made equivalent. Conversion tables between GHQ‐12 and K6 are presented. Applying the commonly used cut‐off of ≥3 on the GHQ‐12 bi‐modal scoring, we found that the best corresponding cut‐off on the K6 would be ≥8. The prevalence of psychological distress was then 22% with GHQ‐12% and 21% with K6. Conclusions The GHQ‐12 and K6 measure the same construct and corresponding cut‐off scores on one scale were found for the other scale. This is valuable for longitudinal studies or time series where one scale has replaced the other scale.