Comparison between self-reported and observed locomotion ability scores in patients with stroke: a cross-sectional...

This study aimed to investigate (i) the validity of self-reported scores of locomotion ability in stroke survivors using the ABILOCO-Benin questionnaire, and (ii) the congruence between online Rasch analysis of the ABILOCO-Benin raw scores and a conversion table set along with the questionnaire development. Thirty-five adult poststroke patients [mean age: 52 (SD: 10) years, 60% men] recruited from the rehabilitation department of National University Hospital of Cotonou (Benin), self-reported their locomotion ability, filling out the ABILOCO-Benin questionnaire [self-reported scores (SRS)]. Afterwards, a physical therapist observed and rated the patients' locomotion ability, while they performed each of the 15 items of the questionnaire [performance-based scores (PBS)]. Both raw SRS and PBS were then converted to linear measures using (i) an online Rasch analysis method and (ii) an ordinal-to-interval transformation table. Analyses showed high correlation (intraclass correlation coefficient=0.74, P<0.001) and nonsignificant difference (P=0.778) between SRS and PBS. Linear measures from online Rasch and a conversion table also exhibited high correlation (intraclass correlation coefficient=0.92; P<0.001). However, the difference between online analysis and the conversion table was significant (P=0.022). In conclusion, self-report is a valid method to administer ABILOCO-Benin in stroke patients. The conversion table offers a valid opportunity for quick transfer of raw scores to linear measures. However, the accuracy of linear measures from the conversion table may be slightly affected by missing responses.