Validation of a three-factor measurement model of dyspnea in hospitalized adults with heart failure

Abstract

Objective: The purpose of this study was to validate a 3-factor measurement model of dyspnea sensory quality (WORK-EFFORT, TIGHTNESS, SMOTHERING-AIR HUNGER) originally derived in patients with exacerbated chronic obstructive pulmonary disease.

Methods: In this validation study, adult patients with heart failure were enrolled after hospital admission (median hospital day 1) and asked to rate the intensity of dyspnea sensory quality descriptors on the day of enrollment (study day 1; N = 119) and in a recall version for the day of admission (study day 0; n = 97).

Results: Confirmatory factor analysis demonstrated good model fit for both days. Cronbach's α for each factor was greater than .87 for both study days.

Conclusion: This is the first study to validate a previously specified measurement model of dyspnea sensory quality in an independent sample. Results indicate that measurement of dyspnea sensory quality in exacerbated cardiopulmonary disease does not necessarily require disease-specific questionnaires.

Figure 1

Assessed for eligibility, exclusions, eligible to participate, and agreed versus declined to participate. M = male; F = female; HF = heart failure; STEMI = ST-elevation myocardial infarction; ACS = acute coronary syndrome; Ca = cancer.

Figure 2

Dyspnea descriptor percent endorsement and mean ratings for Study Day 1 and Study Day 0. (2A) Percentage of participants endorsing dyspnea descriptors (rating > 0); (2B) Mean intensity ratings for dyspnea descriptors (All Day 0 to Day 1 comparisons, |z| > 6.5, P < .0001, paired Wilcoxon test)

Figure 2

Dyspnea descriptor percent endorsement and mean ratings for Study Day 1 and Study Day 0. (2A) Percentage of participants endorsing dyspnea descriptors (rating > 0); (2B) Mean intensity ratings for dyspnea descriptors (All Day 0 to Day 1 comparisons, |z| > 6.5, P < .0001, paired Wilcoxon test)

Figure 3

Standardized estimates for 3-factor / 7-indicator model. (3A) Heart failure, Study Day 1 (N = 119). (3B) Heart failure, Study Day 0 (n = 97). (3C) Chronic obstructive pulmonary disease (data from ref. ; N = 100). Ovals = factors (latent variables); rectangles = indicators (measured variables, descriptors); circles = unique measurement error; capital letters = factor names; SMOTHERING = SMOTHERING-AIR HUNGER factor; lowercase letters = variable (descriptor) names (see Table 1 for complete items); double-headed curved arrows = correlations between factors; single-headed arrows from factors to indicators = factor loadings (standardized partial regression coefficients, ie, β-weights); single-headed arrows from errors to indicators = standardized residuals.

Figure 3

Standardized estimates for 3-factor / 7-indicator model. (3A) Heart failure, Study Day 1 (N = 119). (3B) Heart failure, Study Day 0 (n = 97). (3C) Chronic obstructive pulmonary disease (data from ref. ; N = 100). Ovals = factors (latent variables); rectangles = indicators (measured variables, descriptors); circles = unique measurement error; capital letters = factor names; SMOTHERING = SMOTHERING-AIR HUNGER factor; lowercase letters = variable (descriptor) names (see Table 1 for complete items); double-headed curved arrows = correlations between factors; single-headed arrows from factors to indicators = factor loadings (standardized partial regression coefficients, ie, β-weights); single-headed arrows from errors to indicators = standardized residuals.

Figure 3

Standardized estimates for 3-factor / 7-indicator model. (3A) Heart failure, Study Day 1 (N = 119). (3B) Heart failure, Study Day 0 (n = 97). (3C) Chronic obstructive pulmonary disease (data from ref. ; N = 100). Ovals = factors (latent variables); rectangles = indicators (measured variables, descriptors); circles = unique measurement error; capital letters = factor names; SMOTHERING = SMOTHERING-AIR HUNGER factor; lowercase letters = variable (descriptor) names (see Table 1 for complete items); double-headed curved arrows = correlations between factors; single-headed arrows from factors to indicators = factor loadings (standardized partial regression coefficients, ie, β-weights); single-headed arrows from errors to indicators = standardized residuals.