Accuracy, reproducibility and variability of quantitative assessments of bulbar and respiratory function in motor neurone disease.

The validity of quantitative tests to assess bulbar and respiratory function used in therapeutic trials in motor neurone disease (MND) is studied in 26 MND and 21 age- and sex-matched control subjects. Five raters timed repeated visual and auditory stimuli generated by a metronome simulating bulbar tests. For all raters, mean error rate was 5.07%, correlation coefficient was 0.89 (and %difference of 1.53), and coefficient of variation (CV) was 3.18%. The experienced rater obtained significantly better %differences and CV. For all subjects and tests, mean correlation coefficient was 0.98, and mean CV was 8.8%. There were no significant differences in reproducibility or variability in readings obtained in MND and control subjects assessed by the experienced rater. The use of composite bulbar, but not respiratory, scores resulted in significant improvements in reproducibility and variability.