Validation of the Total Disability Index (TDI) in French Version

Abstract

IntroductionThe Total Disability Index (TDI) questionnaire has been developed to provide a more complete assessment of low back and neck pain, as they frequently co-occur. This study aimed at validating the TDI questionnaire in French, to determine if it could be used in France. HypothesisThe TDI French version is valid, reproducible and comparable to the English version. MethodsThis multicentric study prospectively included French-speaking volunteers, both patients admitted for spine surgery in two specialized spine centers and healthy individuals. Healthy subjects were recruited among students of an engineering school and medical staff. A booklet was given to the participants containing a Lumbar and Cervical Visual Analog Scale (respectively LVAS and CVAS), and the French versions of Oswestry Disability Index (ODI), Neck Disability Index (NDI) and TDI questionnaires. Statistical analysis included Cronbach's α calculation for internal consistency assessment, correlation analysis with ODI and NDI items for convergent validity, principal component analyses and factor analysis. Discriminant validity was assessed by comparing healthy subjects and patients using Student's t tests, and floor and ceiling effects search. Results71 participants were included, with 34 (48%) healthy volunteers and 37 (52%) patients. Mean age was 45.2±19.6 years and 57% of the cohort were males. Internal consistency was good: Cronbach's α was calculated at 0.96 (95%CI: [0.95–0.98]). For each TDI item, a high correlation was found with ODI corresponding items, between 0.81 and 0.97 (p<0.001), and good correlation with NDI items, ranging from 0.65 to 0.96 (p<0.001). TDI correlated also with LVAS and CVAS (respectively 0.70 and 0.65, p<0.001). Principal component analyses indicated good correlation between the TDI items and between each item and TDI total score. Factor analysis indicated two main factors explaining 77% of TDI variance, constituted by all TDI items. Regarding discriminative validity, healthy subjects and patients presented significantly different TDI scores (p-values ≤0.01 for each item). Barplot representations of each TDI item revealed no major floor nor ceiling effects. ConclusionThis study confirms the reliability, feasibility and validity of the Total Disability Index questionnaire in its French version. Its validation allows its use in France. Level of evidenceII.