Interobserver perceptual analysis of smokers voice

Clin Otolaryngol Allied Sci. 2004 Apr;29(2):124-7. doi: 10.1111/j.1365-2273.2004.00793.x.

Abstract

The purpose of this study was to investigate variations in the voice by three experienced speech-language therapists. Forty-eight men and nine women from the Associação dos Alcoólicos Anônimos, Santos were studied. Their ages were from 28 to 81 years, with median of 49 years and everyone was a smoker for 60 to 720 months. Most of them used more than 20 cigarettes a day and all of them had stopped alcohol use for 1 to 25 months. The perceptual analysis of the voices was performed by means of the GRBAS scale. The voice sample consisted of a sustained vowel /a/ at a comfortable pitch and loudness level. The three judges were blinded to the opinion of their colleagues and a kappa test was applied. For roughness, the concordance rates were 59.6% for observers 1 and 2 (kappa = 0.234); 47.4% for 1 and 3 (kappa = 0.047) and 52.6% (kappa = 0.016) for 2 and 3. For breathiness, the concordance rates were 80.7% for observers 1 and 2 (kappa = 0.191); 57.9% for 1 and 3 (kappa = 0.147) and 57.9% (kappa = 0.156) for 2 and 3. With regard to asthenic quality, there was concordance of 100% for observers 1 and 2, so kappa could not be applied; the concordance rate was 96.5% for 1 and 3 and for 2 and 3. The evaluation for strained voice revealed concordance rates of 71.9% for observers 1 and 2 (kappa = -0.017); 59.6% for 1 and 3 (kappa = 0.095) and 70.2% (kappa = 0.039) for 2 and 3. The disagreement among the observers was worst for pathological rather than normal voices; when disagreement was present among experienced judges, it was of only one point in the scale used.

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Alcoholism*
  • Auditory Perception*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Observer Variation
  • Smoking*
  • Speech Production Measurement* / statistics & numerical data
  • Voice Quality*