Translation, adaptation and validation of Parkinson’s disease quality of life − 39 questionnaire to Yoruba language

Abstract Background The Parkinson’s disease quality of life questionnaire (PDQ-39) is a widely accepted measure available in many African languages, none of which is a Nigerian language. This study was aimed at translation and adaptation of the PDQ-39 to Yoruba language of the southwest Nigeria, and assessing the validity and reliability of Yoruba version of PDQ (PDQ-39 (Y)). Methods This adaptation followed the American Association of Orthopedic Surgeons guidelines. The English version of PDQ-39 was cross-culturally adapted into Yoruba using forward- and back-translation, expert panel review, pre-testing of pre-final version on 10 patients with PD for cognitive debriefing interview to produce the final PDQ-39(Y). The PDQ-39 and PDQ-39(Y) were administered concurrently on 40 respondents with PD for known-group validity. The PDQ-39(Y) was re-administered on 33 respondents at 14-day interval for test-retest reliability. Respondents’ socio-demographic data were also recorded. Socio-demographic variables, PDQ-39 and PDQ-39(Y) scores were summarised using descriptive statistics. Known-group validity was assessed using independent t-test by gender and one-way ANOVA by age and educational level. Pearson’s correlation method examined agreement between PDQ-39 and PDQ-39(Y). Internal consistency and test-retest reliability of PDQ-39(Y) were determined by Cronbach’s alpha and intra-class correlation coefficient (ICC) respectively. Construct validity was determined through first-order confirmatory factor analysis (CFA). Level of significance was at p < 0.05. Results The adaptation process revealed experiential and semantic equivalence issues in certain words, which were resolved during the expert panel meeting. Respondents’ mean (SD) age was 61.6 (6.8) years, overall scores on PDQ-39 and PDQ-39(Y) were 54.4 (11.6) and 53.8 (10.4) respectively. All domain scores correlated ( r = 0.780 to 0.990; p = 0.001), and did not differ significantly ( p = 0.180 to 0.880) for both versions. Cronbach’s Alpha values were 0.930 to 1.000 for internal consistency while ICC was 0.830 to 0.990 for test-retest reliability for domains and overall scores. Respondents scores did not differ significantly by educational levels ( p = 0.103 to 0.961) in most, and across age groups ( p = 0.103 to 0.961) and gender ( p = 0.180 to 0.880) in all PDQ-39(Y) domains or overall scores. First-order CFA models for all eight domains of PDQ-39(Y) converged successfully, with factor loadings ranging from 0.501 to 0.867. Conclusion PDQ-39(Y) is valid and reliable with acceptable to excellent structural validity and uni-dimensionality across standard fit indices.