Assessing anxiety among adolescents in Hong Kong: psychometric properties and validity of the Generalised Anxiety Disorder-7 (GAD-7) in an epidemiological community sample

• Published in: BMC psychiatry 2022-11, Vol.22 (1), p.703-11, Article 703
• Main Authors: Ip, Hang, Suen, Yi Nam, Hui, Christy Lai Ming, Wong, Stephanie Ming Yin, Chan, Sherry Kit Wa, Lee, Edwin Ho Ming, Wong, Michael Tak Hing, Chen, Eric Yu Hai
• Format: Article
• Language: English
• Subjects: Adolescent; Adolescents; Anxiety; Anxiety - diagnosis; Anxiety - epidemiology; Anxiety disorders; Anxiety Disorders - diagnosis; Anxiety Disorders - epidemiology; Anxiety in teenagers; Anxiety in youth; Comorbidity; Diagnosis; Drug dependence; Epidemiology; Evaluation; GAD-7; Hong Kong - epidemiology; Humans; Insomnia; Interviews; Likert scale; Medical personnel; Mental disorders; Mental health; Mental health screening; Methods; Patient Health Questionnaire; Patients; Psychiatry; Psychometrics; Quantitative psychology; Questionnaires; Rasch Model; Regression analysis; Reproducibility of Results; Surveys and Questionnaires; Teenagers; Validity; Young adults
• ISSN: 1471-244X; 1471-244X
• DOI: 10.1186/s12888-022-04329-9

The development of a valid and simple-to-use self-administered tool in Asian adolescents for clinical screening and intervention remains limited. The present study assessed the psychometric characteristics and validity of the Generalised Anxiety Disorder Scale-7 (GAD-7) among adolescents in Hong Kong. Epidemiological data from 3,261 Hong Kong adolescents aged 15 - 24 years were analysed for the construct validity, criterion validity, concurrent validity, and Rasch Model. All analyses were age- and gender-weighted according to the distributions of Hong Kong's general population. The GAD-7 showed high internal consistency and strong fit to the one-factor structure. The best cut-off value was set at 7 or more. Regression models found that the total scores of the scale were positively associated with symptoms of depression and hypomania, schizotypal personality and alcohol dependence. Rasch model analysis found that the separation index was 2.18 and 16.51 for the respondents and items, respectively and all residual pairs had small correlation coefficients (i.e., < 0.3). All psychometric findings presented in this study support the use of the GAD-7 as a legitimate measure of anxiety severity. A cut-off score of 7 should indicate a potential diagnosable condition in Asian adolescents, which requires our attention but should not be used as a formal diagnostic screening tool. The findings revealed the local dependence of the items of the GAD-7 and that the scale can separate respondents into at least two groups and items into numerous groups according to the separation index.