Nasometric sensitivity and specificity: a cross-dialect and cross-culture study.

A series of 514 patients seen at three clinics in the United States and Spain were evaluated using clinical judgments of hypernasality, and nasometric assessment of oral-nasal resonance balance. Data from the nasometer were obtained while patients read a passage devoid of nasal consonants. Across all subjects, the Pearson correlation coefficient between the clinical and instrumental measures was 0.78. Prediction analyses revealed that maximum efficiency was obtained using a somewhat different threshold nasalance value for each of the three patient samples. When all 514 subjects were investigated as a single group, a threshold nasalance score of 28 was found to optimize identification of patients with and without clinically significant hypernasality. In that analysis, a sensitivity of 0.87, a specificity of 0.86 and an overall efficiency of 0.87 was obtained. The clinical relevance of these findings is discussed.