Translation in Arabic, adaptation and validation of the SF-36 Health Survey for use in Tunisia.

OBJECTIVE To translate into Arabic and validate the SF-36 quality of life index in a Tunisian Arabic population. BACKGROUND No validated Arabic quality of life index is published. DESIGN Arabic translation of the SF-36 scale was obtained by the "forward/backward translation" method. Adaptations were made after a pilot study involving 22 subjects from general population. Inter-rater reliability was assessed by use of intraclass correlation coefficient (ICC) and Bland and Altman method. Construct validity was assessed by Spearman rank correlation coefficient (convergent and divergent validity), and factor analysis with Varimax rotation. Internal consistency was assessed by Cronbach alpha coefficient. RESULTS We note that 130 Tunisian subjects were included in the validation study. No items were excluded. Inter-rater reliability was excellent (ICC=0.98). Cronbach alpha coefficient was 0.94 conferring to translated index a good internal consistency. Expected divergent and convergent validity results suggested good construct validity. Two main factors were extracted by factor analysis and explained 62.3% of the cumulative variance: the first factor represented mental component, the second physical component. The Cronbach alpha coefficient was 0.88 and 0.91 respectively for factor 1 and factor 2. CONCLUSION We translated into Arabic language and adapted the SF-36 scale for use in Tunisian population. The Arabic version is reliable and valid. Although the scale was validated in a Tunisian population, we expect that it is suitable for other Arab populations, especially North Africans. Further studies are needed to confirm such a hypothesis.