[French validation of the revised Piper Fatigue Scale].

The need for a valid and reliable multi-dimensional instrument in French to enable nurses and clinicians to assess intensity and changes in cancer-related fatigue (CRF) and evaluate efficacy of intervention stratégies on patient outcome, prompted us to translate the Piper Fatigue Scale (PFS) and assess its psychometric properties in a French speaking population. The revised PFS is a 22 item self-report instrument with four sub-scales: behavioural/severity, affective, sensory and cognitive/mood. The PFS was translated using a procédure of translation and back translation and tested on 20 bilingual subjects. Correlation coefficients were measured between the item scores of the English and French version of the PFS. The French version of the PFS was adjusted and then completed by 229 adult cancer patients with solid tumours under-going treatment by chemo-, radio-, hormono- or immunotherapy. A factorial analysis and a multi-traits analysis were used to assess the reliability and the dimensional structure of the translated version of the PFS. Cronbach's alpha correlations ranged between 0.85 and 0.92 for the various sub-scales, thus indicating good sub-scale reliability. Item internal consistency was good with 100% of item-scale correlations greater than or equal to 0.4. Item discriminant validity checking showed 3 items to be insufficiently discriminating. Item-scale correlations greater than the correlation of the item with other sub-scales were 98%. The Cognitive/Mood dimension appeared to group 2 separate dimensions. Convergent validity, tested by correlating the PFS with a Visual Analogue Scale measuring fatigue and with the OMS Physical Performance Status Scale, revealed significant relationships. A five sub-scale model of the PFS is proposed containing 22 items to increase performance of the French version of the PFS comparatively to the 4 sub-scale model of the original PFS. Translation of 2 items has been slightly modified to reduce ambiguity. The overall results of this preliminary study are satisfactory. We expect results to be further improved by the modifications proposed and testing is now underway in homogenous populations to evaluate the sensitivity of the instrument to changes over time. This will enable us to ascertain the definitive French version of the Piper Fatigue Scale.