A Kiswahili version of the SF-36 Health Survey for use in Tanzania: translation and tests of scaling assumptions

The objective of the study was to translate and adapt the SF-36 Health Surv ey for use in Tanzania and to test the psychometric properties of the Kiswa hili SF-36. A cross-sectional study was conducted as part of a household su rvey of a representative sample of the adult population of Dar es Salaam, T anzania. The IQOLA method of forward and backward translation was used to t ranslate the SF-36 into Kiswahili. The translated questionnaire was adminis tered by trained interviewers to 3,802 adults (50% women, mean (SD) age 31 (13) years, 50% married and 60% with primary education). Data quality and p sychometric assumptions underlying the scoring of the eight SF-36 scales we re evaluated for the entire sample and separately for the least educated su bgroup (n=402), using multitrait scaling analysis. Forward and backward tra nslation procedures resulted in a Kiswahili SF-36 that was considered conce ptually equivalent to the US English SF-36. Data quality was excellent: onl y 1.2% of respondents were excluded because they answered less than half of the items for one or more scales; ninety percent of respondents answered m utually exclusive items consistently. Median item-scale correlations across the eight scales ranged from 0.47 to 0.81 for the entire sample. Median sc aling success rates were 100% (range 87.5-100.0). The median internal consi stency reliability of the eight scales for the entire sample was 0.81 (rang e 0.70-0.92). Floor effects were low and ceiling effects were high on five of the eight scales. Results for n=402 people without formal education did not differ substantially from those of the entire sample. The results of da ta quality and psychometric tests support the scoring of the eight scales u sing standard scoring algorithms. The Kiswahili translation of the SF-36 ma y be useful in estimating the health of people in Dar es Salaam. Evidence f or the validity of the SF-36 for use in Tanzania needs to be accumulated. Q ual. Life Res. 8:101-110 (C) 1999 Kluwer Academic Publishers.