Validation of the World Health Organization Well-Being Index (WHO-5) among medical educators in Hong Kong: a confirmatory factor analysis

Abstract

Background The COVID-19 pandemic has exacerbated the pre-existing global crisis of physician burnout. Physician and particularly medical educator well-being, has come into focus as educators can influence their own and learners’ well-being. Measuring this construct is one important step towards promoting well-being in the work and learning environments. The 5-item World Health Organization Well-Being Index (WHO-5) has been validated in different populations worldwide for assessing well-being. Yet, its psychometric acceptability remains unexplored among medical educators in Asia including Hong Kong (HK). This study evaluates the validity of the WHO-5 when used among HK medical educators. Method Using data from 435 medical educators, we employed combined within-network (confirmatory factor analysis; CFA) and between-network approaches (correlation and regression) to scale validation. Results CFA results indicated that our data fit the a priori WHO-5 model, suggesting structural validity. Results of comparison of means indicated no gender differences, but there were significant differences when participants were compared by age and professional backgrounds. Resilience predicted well-being as measured by the WHO-5, suggesting construct criterion validity. Conclusions Our findings extend the validity evidence for the WHO-5 to HK medical educators examined in this study. This enables their well-being to be assessed when evaluating the impact of future well-being programmes.

Psychometric Properties of the WHO-5 Well-Being Index among Nurses during the COVID-19 Pandemic: A Cross-Sectional Study in Three Countries

“…These combined findings emphasize that nurses from Norway reported better well-being than those in Spain and Chile during the pandemic. These results appear consistent with studies reporting that the WHO-5 Well-Being Index can differentiate between populations [ 31 , 43 , 45 ]. However, further research, using several measurement points, is needed to investigate whether the nurses’ mental well-being changes over time and whether the WHO-5 Well-Being Index can capture these changes.…”

“…Contrary to previous studies showing low Cronbach’s alpha [ 42 ], the WHO-5 Well-Being Index showed high internal consistency in the three countries, with Cronbach’s alpha varying from 0.810 to 0.903 (Norway to Chile). These findings are within the range of Cronbach’s alpha reported in validation studies among medical educators in Hong Kong [ 43 ], outpatients with epilepsy in Denmark [ 44 ], adults living with epilepsy and HIV in Kenya [ 45 ] and among Chinese university students [ 46 ].…”

Assessing Burnout and Well-Being in Higher Education Health Science Faculty

“…In a 2015 systematic review, Topp et al found that the tool had high clinimetric validity and could be used effectively as a non-invasive, subjective measure of well-being [17]. Previous research has also reported structural and construct criterion validity using the WHO-5 Index in numerous countries and populations [18,19]. Cronbach alphas ranged from 0.81-0.90 in a study of medical educators concluding the instrument to be a psychometrically sound tool for measuring well-being [18].…”

“…Previous research has also reported structural and construct criterion validity using the WHO-5 Index in numerous countries and populations [18,19]. Cronbach alphas ranged from 0.81-0.90 in a study of medical educators concluding the instrument to be a psychometrically sound tool for measuring well-being [18]. However, in a study of adolescent mental well-being spanning 15 countries, the elimination of the first item on the questionnaire ("I have felt cheerful and in good spirits") resulted in better psychometric properties in that population [20].…”

Applying Factor Analysis for Assessing Knowledge Structure of Students in Grade 10: The Subject of Redox Reaction

“…Researchers generally use CFA to test whether the data match the existing models (Chan et al, 2022). If no model has been proposed, EFA can be used to explore the relationships between variables (model), and then CFA is employed to test and modify the model (Alotaibi & Alotaibi, 2021;Pereira et al, 2021).…”