The primary goal of the translation effort is a translated version of the KDQOL-SF™ that is linguistically and conceptually equivalent to the U. S. English version.  Equivalence can only be obtained if the original and translated versions of the KDQOL-SF have the same meaning.  Translators should be instructed to produce colloquial translations that will be understood by the general public (Acquadro, Jambon, Ellis, & Marquis, 1996).

As in the Renal Outcomes Study, the KDQOL-SF™ instructions, items, and response choices should be translated independently by at least two trained bilingual translators.  These translations should be carried out by local teams and both translators should be native speakers of the language into which the KDQOL-SF™ is being translated.  Translators should rate the difficulty of translating each item and response scale using a 0 (not at all difficulty) to 100 (most difficult) scale (Ware, Gandek, Keller, & the IQOLA Project Group, 1996).  To help select equivalent response options, the Thurstone and Chave (1929) method of equal-appearing intervals can be employed.  In this method, a sample of raters (n = 25 or so) is asked to rate the position of intermediate response choices using a 10-cm line anchored by the extreme (lowest and highest) response choices (Ware et al., 1996).  The translators should compare their translations and reconcile discrepancies.

Two different translators should then rate the quality of the reconciled forward translation.  Each item and response scale is rated for its conceptual equivalence to the U. S. English version using a 0 (not at all equivalent) to 100 (exactly equivalent) scale.  Items and response scales that are rated less than 75 on the 0 to 100 scale are re-translated by the original translators until an acceptable independent rating of equivalence is obtained.

The resulting translation should then be cognitively tested in a small sample (about 10) of dialysis patients.  This testing should take the form of concurrent or think-aloud interviews as well as self-administration followed by retrospective interviews (Jobe & Mingay, 1990). Following cognitive testing, the KDQOL-SF item and response options should be rewritten as necessary and a new version of the translation produced. Ideally, the next step is to convene a panel that includes the forward translators, experts in kidney disease (nephrologist or nurse), a survey design expert, and an individual with kidney disease. The panel's job is to evaluate the conceptual equivalence of the translation and the original survey instrument. The forward translation is then finalized based on the panel's feedback.

The final forward translation should be back-translated into U.S. English by two other translators. Both of these translators should be native English speakers. These two translators should compare their backward translations and come to agreement about discrepancies. The reconciled back translation will then be compared against the U.S. English original and each item and response scale rated for equivalence to the original U. S. English version by the KDQOL Working Group using a 0  (not at all equivalent) to 100 (exactly equivalent) scale. Items and response scales must be rated 75 or higher on the 0 to 100 scale for the translation to be approved. A memo summarizing the back-translation should be written.

The next step is field testing the translated KDQOL-SF. At a minimum, the KDQOL-SF should be administered to a sample of 75 people with kidney disease who are native speakers of the target language. Scale equivalence should be assessed by performing standard reliability and validity testing and comparing these results to those obtained for the U.S. English sample (Hays, Anderson, & Revicki, 1995; Hays, Kallich, et al.,1994). Ideally, the translated and the U.S. English versions of the KDQOL-SF should be administered to a bilingual sample in counter-balanced order to allow for direct comparisons of responses for the same respondent (c.f., Coons, Alabdulmohsin, Draugalis,& Hays, 1998).  A final report should be written including the information in the forward and back-translation memos as well as results of the field test.

References