The Malay version of SF-36 health survey instrument: testing data quality, scaling assumptions, reliability and validity in post-coronary artery bypass grafting (CABG) surgery patients at the National Heart Institute (Institut Jantung Negara—IJN), Kuala Lumpur

Abstract

Background The Short Form 36 (SF-36) is a scoring system comprising of 36 items categorized into eight constructs corresponding to patients’ health-related quality of life. It has been used extensively in various countries on different sub-populations and used to indicate the health status and help to ascertain the effect of clinical interventions on the particular population. Objective To examine the psychometric properties of the Malay version of SF-36 (Malay SF-36) summated rating scales and validate the scale among post-coronary artery bypass grafting surgery (CABG) patients at the National Heart Institute (IJN), Kuala Lumpur. Methods Five hundred and nine post-CABG patients at the IJN, Malaysia completed the questionnaires between 1 July and 31 December 2017. Psychometric tests endorsed by the “International Quality of Life Assessment Project” were utilised. Results The data quality was excellent with a high questionnaire completion rate (100%). As hypothesized, the ordering of item means within scales was clustered. In unison, scaling assumptions were satisfied. Good discriminant validity was shown between subsets of patients with various levels of health status. Notwithstanding, there were probably translation issues of the Physical Functioning scale which showed small ceiling effects. We clearly observed high ceiling and floor effects in both Role Physical and Role Emotional scale most probably attributed to the dichotomous style of their choice of responses. Cronbach alpha values of the eight scales ranged from 0.73 to 0.90, showing good internal consistency reliability. Confirmatory Factor Analysis (CFA) confirmed the 8-factor solution and Composite Reliability revealed internal consistency reliability except for Vitality and Social Functioning. Based on the Average Variance Extracted (AVE), convergent validity was adequate except for two domains. Discriminant Validity is good for the eight constructs as the √AVE are generally higher than the correlation coefficients between the latent constructs. Conclusion The scoring for the Malay SF-36 based on the summated ratings method was proven to be valid to be applied in our local clinical population. The CFA, fitness estimates, reliability and validity assessments suggest that the Malay version of SF36 is a valid and reliable instrument. However, further work is warranted to further refine the convergent validity and reliability of some scales.