Converting three general-cognitive function scales into persian and assessment of their validity and reliability

Glasgow Outcome Scale Extended (GOSE), Galveston Amnesia and orientation Test (GOAT) and Disability Rating Scale (DRS) are three popular outcome measure tools used principally in traumatic brain injury (TBI) patients. We conducted this study to provide a Farsi version of these outcome scales for use in Iran.

Following a comprehensive literature review, Farsi transcripts were prepared by "forward-backward" translation and reviewed by subject experts. After a pretest on a few patients, the final versions were obtained. 38 patients with closed head injury were interviewed simultaneously by two interviewers. Main statistics used to assess validity and reliability included "Factor analysis" for construct validity, Cronbach's alpha for internal consistency, and Pearson Correlation and Kappa Coefficient for inter-rater agreement.

Factor analysis for Farsi-GOAT (FGOAT) revealed 5 independent factors with a total distribution variance of 80.2%. For Farsi-DRS (FDRS), 3 independent factors were found with a 92.3% variance. The Cronbach's alpha (95% confidence interval) was 0.84 (0.763- 0.919) and 0.91 (0.901-0.919) for FGOAT and FDRS, respectively. Pearson Correlation between total scores of two raters was 0.98 and 0.97 for FGOAT and FDRS, in order. Kappa coefficient (95% CI) between outcome rankings of raters was 0.73 (0.618-0.852) and 0.68 (0.594-0.770) for FGOAT and FDRS, respectively. As for Farsi-GOSE scale, Kappa value was 0.4 (0.285-0.507) for 8-level outcome ranking and improved to 0.7 (0.585-0.817) for 5-level scale. We found a good correlation between FDRS and FGOSE predicted prognoses (Spearman's rho= 0.74, 95% CI: 0.676-0.802).

FDRS and FGOAT had appropriate validity and reliability. The 8-level outcome FGOSE scale disclosed a low inter-rater agreement, but a suitable observer agreement was achieved when the 5-level outcome was applied.