Maximum Phonation Time: Variability and Reliability

Summary

The objective of the study was to determine maximum phonation time reliability as a function of the number of trials, days, and raters in dysphonic and control subjects. Two groups of adult subjects participated in this reliability study: a group of outpatients with functional or organic dysphonia versus a group of healthy control subjects matched by age and gender. Over a period of maximally 6 weeks, three video recordings were made of five subjects' maximum phonation time trials. A panel of five experts were responsible for all measurements, including a repeated measurement of the subjects' first recordings. Patients showed significantly shorter maximum phonation times compared with healthy controls (on average, 6.6 seconds shorter). The averaged interclass correlation coefficient (ICC) over all raters per trial for the first day was 0.998. The averaged reliability coefficient per rater and per trial for repeated measurements of the first day's data was 0.997, indicating high intrarater reliability. The mean reliability coefficient per day for one trial was 0.939. When using five trials, the reliability increased to 0.987. The reliability over five trials for a single day was 0.836; for 2 days, 0.911; and for 3 days, 0.935. To conclude, the maximum phonation time has proven to be a highly reliable measure in voice assessment. A single rater is sufficient to provide highly reliable measurements.

Introduction

Voice is a multidimensional phenomenon.¹ The multidimensional voice assessment proposed by the European Laryngological Society includes aerodynamic measurements.² Maximum phonation time is usually used for practical reasons: determination of maximum phonation time is a noninvasive, fast, and low-budget measurement. The ability to maximally sustain a vowel after having taken a maximal inspiration is considered an objective measure of the efficiency of the respiratory mechanism during phonation.³ In the literature, maximum phonation time has been used to objectify the degree of severity of dysphonia and to determine the effects of voice therapy.⁴ However, some authors question the usefulness of this measure as an evaluation tool in therapy, as no significant change could be demonstrated after therapy in a group of patients with vocal nodules (eg, Treole and Trudeau⁵).

Many studies have provided normative data on maximum phonation time in diverse subject populations (see for an overview, Baken and Orlikoff⁶). The variability is high, depending on subject characteristics, such as gender or age⁷ and testing conditions.3, 8 In 1968, Hirano et al stated that although tests, such as maximum phonation time, indicate the degree of vocal function, no diagnosis of laryngeal disease could be established,⁹ an assertion that has since been confirmed by other authors.10, 11 In a study on procedural aspects of eliciting maximum phonation time, Neiman and Edeson³ concluded that subjects should receive a complete verbal and visual model of the experimental task before its elicitation; instructions should be standardized; and at least three trials of maximum phonation time were required before performance could be expected to approximate the criterion. However, in a study by Lewis et al,⁷ the authors found that using three trials to determine maximum phonation time in children yielded inadequate data, as most subjects had not yet reached their maximum achievement after three trials. Using more trials, higher achievement levels were found. On the other hand, Finnegan¹² demonstrated the presence of fatigue effects after practice effects. Furthermore, Shanks and Mast¹³ considered the differential operation of fatigue to be partially supported by the progressive increase in standard deviations found when raising the number of maximum phonation trials. Thus, the outcomes of studies on fatigue versus practice effects while performing maximum phonation tasks are not quite consistent. Although quite a few studies have described maximum phonation time in diverse subject populations and under various testing conditions, limited information is available on the reliability of the data over time.⁷

To our knowledge, no study has thus far determined how many days the subjects should be repeatedly measured, nor have any authors indicated how many trials and raters would be necessary to obtain reliable maximum phonation time measurements using dysphonic subjects and control subjects matched by gender and age. The purpose of the present study is to determine the reliability of maximum phonation time as a function of the number of trials, days, and raters in dysphonic and control subjects.