Abstract
The Indonesian version of the SF-36 Item Health Survey instrument measured quality of life in patients with chronic diseases.However, the validity and reliability of this instrument have not been tested in the post-heart attack patient population.This study aimed to assess the construct validity and reliability of the Indonesian version of SF-36 in postheart attack patients in Mataram, Indonesia.This study was a cross-sectional study.Thirty participants were recruited using consecutive sampling.Confirmatory Factor Analysis using Partial Least Squares Structural Equation Modeling (PLS-SEM) with WarpPLS version 6.0 was carried out to assess convergent and divergent validity, indicator reliability and internal consistency reliability, and evaluating the outer model of the instrument that consisted of three orders.In the first order, 35 items (indicators) constructed eight domains quality of life (PF, RLP, P, FH, RLE, E/F, EWB and SF), and in the second order, the previous eight domains constructed the second-order domains.In the third order, domains in the second order constructed the main construct, Health Status.Two alternative models were tested.The difference of the two models lied on the domains of the second order.In the first order of both models, 12 out of 35 items showed low indicator reliability.All the eight first-order domains had good internal consistencies.Six and seven domains had good convergent validity and good discriminant validity, respectively.Alternative model 2, which had three second-order domains, PCS, MCS, and WB, demonstrated good indicator reliability and convergent validity on more domains in the second order than alternative model 1.Therefore, it was concluded that the Indonesian version of the SF-36 Item Health Survey showed some evidence of construct validity and reliability, although 12 items showed low indicator reliability.In addition, the alternative model that had three second-order domains had better validity and reliability.