Jafri Malin, Siti Amirah Hanna
(2021)
The psychometric properties of World Health Organization quality of life brief assessment in Malay (WHOQOL-BREF).
Masters thesis, Universiti Sains Malaysia.
Abstract
Introduction: The World Health Organization Quality of Life Brief (WHOQOL-BREF) (WHOQOL Group, 1998) was translated into the Bahasa Malaysia language by Hasanah, Naing and Rahman (2003) and has been widely used in Malaysia. However, the lack of generalisability of the sample population and the type of validity used in the previous study required an update in psychometric properties of this instrument. Methodology: This study sample comprised of 301 Malaysian adults who completed the original WHOQOL-BREF, the Malay WHOQOL-BREF and DASS-21 through SurveyMonkey. To establish the model fit of the translated instrument, Confirmatory Factor Analysis (CFA) using Maximum Likelihood estimation were run in SPSS version 26 and AMOS version 24 were used. Results: Through CFA, the original four-factor model showed a good model fit in the current study sample population, Chisq/df = 2.08, comparative fit index (CFI) = .93 and root mean square of error approximation (RMSEA) = .06. Construct reliability was established through composite reliability and average variance extracted (AVE) values. All four domains had composite reliability more than .60, which ranged from .79 (social relationship factor) to .88 (psychological health factor). All AVE values were above .50, except for Environmental factor. Convergent validity was established with significant positive relationships between all four domains and two items from the instrument that evaluates overall QoL and general health facet (p <.001). Divergent validity was established with significant negative correlations between all four factors and Depression Anxiety and Stress-21 (DASS21) (p <.001). However, only Social Relationships domain of the instrument showed good discriminant ability between healthy and non-healthy participants (p = .04). Conclusion: The Malay WHOQOL-BREF showed appropriate model fit, reliable construct and convergent validity and good construct reliability. However, the instrument was unable to discriminate between healthy and non-healthy participants. Future research should discuss the findings in clinical samples with cautions.
Actions (login required)
 |
View Item |
