Measurement of dyspnea: word labeled visual analog scale vs. verbal ordinal scale

Abstract

We previously used a verbal ordinal rating scale to measure dyspnea. That scale was easy for subjects to use and the words provided consistency in ratings. We have recently developed a word labeled visual analog scale (LVAS) with labels placed by the subjects, retaining the advantages of a verbal scale while offering a continuous scale that generates parametric data. In a retrospective meta-analysis of data from 43 subjects, individuals differed little in their placement of words on the 100 mm LVAS (mean±S.D. for slight=20±2.5 mm, moderate=50±5 mm and severe=80±6 mm) and ratings were distributed uniformly along the scale. A significant stimulus–response correlation was obtained for both the LVAS (r²=0.98) and for the verbal ordinal scale (Spearman r=0.94). The resolution of the two scales differed only slightly. With meaningful verbal anchors, well-defined end-points, and clear instructions about the specific sensation to be rated, both scales provide valid measures of dyspnea.

Introduction

The measurement of the perception of dyspnea (uncomfortable breathing) presents challenges which have been addressed with different kinds of rating scales (Harver and Mahler, 1990, Adams and Guz, 1991, Lansing and Banzett, 1996). The most widely used devices are visual analog scales, verbal category scales (comprising words having an inherent order, i.e. ordinal scales), and a hybrid of these, the Borg numerical category scales. Each scale type has specific advantages and limitations; for instance verbal scales provide semantic anchors that may aid day-to-day consistency and are more easily communicated in some circumstances, whereas analog scales have good resolution and lend themselves to parametric statistical analysis. The ‘Borg scales’ were developed to combine the advantages of both, but have a more limited resolution than analog scales. Borg positioned words on the scale to achieve a particular relationship between ratings of exertion and physiological variables during exercise (Harver and Mahler, 1990). Borg scales thus lack the inherent semantic validity of scales whose word placement is based on measurement of semantic meaning (Heft and Parker, 1984).

Verbal descriptor rating scales have been widely used to measure the intensity of body sensations such as the discomfort of pain because they are readily understood by the average person and are easy to use even under distressing conditions (Melzack, 1975, Gracely and Dubner, 1987, Duncan et al., 1989). We introduced such a verbal rating scale to measure the intensity of air hunger, the strong sensation of discomfort that can be produced by increasing the drive to breathe while limiting ventilation (Wright and Branscomb, 1954, Banzett et al., 1989, Banzett and Lansing, 1996). The words none, slight, moderate, and extreme, with intermediate half-steps, provided a 7 point rank-order scale which we used to measure changes in air hunger under a variety of experimental conditions (Banzett et al., 1990, Manning et al., 1992, Banzett et al., 1996, Bloch-Salisbury et al., 1996, Harty et al., 1996, Lansing et al., 2000, Moosavi et al., 2000). But this scale shares the limitations of other ordinal category scales: the sensitivity is limited by the number of categories, and the derived data must be analyzed with non-parametric statistics (Aitken, 1969).

Analog scales have been extensively used in the study of pain and respiratory discomfort (Adams and Guz, 1991, Lansing and Banzett, 1996, Price, 1999) but rarely to study the specific sensation of air hunger (an exception is Adams et al., 1985). Analog scales permit subjects to make finely graded judgements of discomfort on a continuous scale, and yield data that are appropriate to analyze with parametric statistics. Visual analog scales with word labels placed at intervals along the scale retain some of the advantages of verbal category scales (Freyd, 1923). Word labeled scales have been developed for the study of pain; subjectively meaningful ‘guide words’ help subjects make consistent use of the scale (Scott and Huskisson, 1976, Heft and Parker, 1984, Rainville et al., 1992). We devised a ‘word labeled visual analog scale’ (LVAS) to measure dyspnea. Before each experiment subjects placed descriptive labels along a vertical scale between the end-points none and extreme at positions that were semantically appropriate to them. We have now used this method to measure air hunger and other discomforts of breathing in response to a variety of stimulus conditions: hypercapnia, hypercapnia with chest vibration, hypercapnia with inhaled furosemide, hypoxia, and bronchoconstriction.

Here we report a retrospective analysis of the data derived from the 43 subjects in those studies to answer several questions. (1) Do people have a common understanding of the meaning of the descriptive labels? This question was addressed by determining whether different subjects placed the words in similar locations along the intensity scale. (2) Do subjects actually use the labeled scale as a continuous scale? This question was addressed by determining whether the labels attracted a disproportionate number of ratings compared to unlabelled regions of the scale. (3) What scale resolution is required for the measurement of air hunger? This question was addressed by determining if the scale could resolve differences in air hunger at the limits of our ability to reliably produce and measure changes in the hypercapnic stimulus. In addition to answering these questions, we compared the performance of the new LVAS with our prior verbal ordinal scale under similar stimulus conditions.