Country differences in assessng UPDRS part III

Objective: To assess the potential impact of regional differences in UPDRS part III assessments. Background: Subjective outcome measures contribute significantly to measurement error. In Psychiatric indications significant rating differences in the use of scales have been found between countries. We therefore conducted a retrospective analysis of a large multinational clinical study in Parkinson's disease with the aim of assessing potential cultural and regional differences in the use of the UPDRS. Methods: All raters intending to participate in this clinical study needed to meet a pre-defined level of experience and were required to score one UPDRS part III video exam for certification. Potential country differences in UPDRS part III total score, number of item deviations and distance from the consensus score were assessed by the analysis of variance. Results: Two hundred and twenty three raters from 12 countries submitted scores for the video. The mean group total score was 35.0 (±7.0), the mean number of item deviations 2.5 (±3.2) and the mean number of absolute item deviations 3.1 (±3.5). Single factor ANOVA yielded significant results in the UPDRS part III total score (p=0.03), number of item deviations (p=0.008) and absolute item deviations (p=0.004). Conclusions: The analysis demonstrates that significant differences between individual countries exist in assessing UPDRS part III and that these differences can potentially be a source of signal noise. Well designed and comprehensive rater training methods need to be implemented at the beginning of any multinational clinical trial to decrease the potential regional/country effect on the study outcomes.