Background and aims: The aim of the present study was to examine the approximate value of maximum phonation time in community-dwelling Japanese people and to set the range of error after ascertaining relative and absolute reliability.
Methods: Participants in the present study were adults living in Kobe and Himeji City, Hyogo Prefecture, Japan. Participants were seated in the upright sitting position and were asked to produce an /a/ sound at their normal speaking volume for as long as they possibly could. The examiner measured the time until the sound became inaudible. Two trials were performed.
Results: The subjects in the final analysis comprised 380 adults (246 women, 134 men; age, 72.7 ± 5.9 years). The values of trial 1 and trial 2 were analyzed with intraclass correlation coefficient (ICC(1,1)), and Bland-Altman plot. Average values of the women, men, and both sexes combined were 16.1, 18.7, and 17.0 s, respectively. Average values of age groups 60-69, 70-79, and 80-89 years were 17.9, 16.8, and 15.9 s, respectively. The ICC(1,1) was 0.87 (P < 0.001). Both fixed bias and proportional bias existed. Limits of agreement (LOA) were -5.0 to 7.7 and % change ranged from -27.1 to 41.8 %.
Conclusions: These normative data indicated that sufficiently reliable values could be measured in two trials. LOA showed that the second value tended to be higher. In determining the effect of treatment, one must consider that the value will be slightly higher at re-test.
Keywords: Absolute reliability; Community-dwelling people; Maximum phonation time; Relative reliability.