Aims: Asthma treatment has the potential to affect patients' voices. We undertook detailed characterisation of voice morbidity in patients attending a problem asthma clinic, and we determined how patients' perceptions related to objective assessment by an experienced observer.Methods: Forty-three patients took part in the study. Subjects completed the self-administered voice symptom score (VoiSS) questionnaire and underwent digital voice recording. These voice recordings were scored using the grade-roughness-breathiness-asthenicity-strain system (GRBAS). Laryngoscopy was also performed.Results: The median VoiSS was 26 (range three to 83). VoiSS were significantly lower in the 17 patients with normal laryngeal structure and function (range four to 46; median 22), compared with the 26 patients with functional or structural laryngeal abnormality (range three to 83; median 33) (95 per cent confidence intervals for difference 0.0-21.0; p ¼ 0.044). The overall grade score for the GRBAS scale did not differ between these two groups, and only 13 patients had a GRBAS score of one or more, recognised as indicating a voice problem. There were positive correlations between related GRBAS score and voice symptom score subscales. Although voice symptom scores were significantly more abnormal in patients with structural and functional abnormalities, this score performed only moderately well as a predictive tool (sensitivity 54 per cent; specificity 71 per cent). Nevertheless, the voice symptom score performed as well as the more labour-intensive GRBAS score (sensitivity 57 per cent; specificity 60 per cent). Patients' inhaled corticosteroid dose (median dose 1000 mg beclomethasone dipropionate or equivalent) had a statistically significant relationship with their overall grade score for the GRBAS scale (r ¼ 0.56; p , 0.001), but not with their VoiSS. Only one patient had evidence of laryngeal candidiasis, and only two had any evidence of abnormality suggesting steroid-induced myopathy.Conclusions: Vocal morbidity is common in patients with asthma, and should not be immediately attributed to steroid-related candidiasis. The VoiSS merits further, prospective validation as a screening tool for ENT and/or speech and language therapy referral in patients with asthma.