Abstract
BACKGROUND: The Motricity Index (MI) is a commonly used method of measuring muscle strength in post-stroke hemiparesis. This study aimed to produce the MI Italian version (MI-IT) and assess its reliability in subjects with stroke. METHODS: Phase-1: stepwise approach to MI-IT production and pilot-testing with 10 health professionals to ensure clarity of each item and instructions for administration and scoring. Phase-2: evaluation of MI-IT reliability on stroke subjects, each independently assessed by 2 raters randomly selected from a group of 10 physiotherapists; the first rater re-administered the MI-IT 1-3 days later. Intraclass correlation coefficients, Spearman's rho and, limited to the more affected side, non-parametric limits of agreement (LOA) were computed for total MI-IT scores, squared weighted kappa and percentage of observed agreement for individual item scores. RESULTS: The back-translated versions showed no discrepancies with original MI, but 3 items were revised after pilot-testing. Complete data on 50 (test-retest) and 51 (inter-rater) participants demonstrated excellent reliability of all MI-IT total scores on the more affected side (Spearman's rho range: test-retest 0.953-0.975; inter-rater: 0.965-0.970), with LOA ranging from 9-25%), but poor inter-rater reliability for some scores on the less affected side (Spearman's rho range: test-retest, 0.816-0.976; inter-rater: 0.508-0.721). Moderate to almost perfect agreement was found for all individual item scores, except for 2 items on the less affected side. CONCLUSIONS: The MI-IT is sufficiently reliable to evaluate motor impairment of the more affected side after stroke, with acceptable measurement error for all scores.