The Utility of Disease-Specific Health Status Instruments in Ambulatory Care

Abstract

Purpose: Currently, a major obstacle in evidence-based otolaryngology is incorporating validated instruments into daily practice. To address this need, we proposed a series of interrelated projects which evaluate the utility of disease-specific health status instruments for patients in the ambulatory care setting. Aim 1 examined the ability of a disease-specific quality of life instrument to predict vocal diagnoses made via laryngoscopy/stroboscopy, as compared to common clinical inquiries about vocal characteristics. Aim 2 evaluated disease-specific and general health status to determine whether general health status data alone can accurately represent vocal health in this population. Aim 3 determined the extent to which voice-specific heath status contributes to general quality of life, when simultaneously adjusting for what have traditionally been considered more concerning comorbidities. Methods: Patients presenting for their first visit to a tertiary care academic medical center laryngology subspecialty clinic completed disease-specific health status instruments (Voice Handicap Index-10 [VHI-10] and Singing Voice Handicap Index-10 [SVHI-10]), common clinical queries, and a standardized global health instrument (Patient-Reported Outcome Measurement Information System [PROMIS] global short form), as part of their routine clinical intake. Results were reviewed against objective voice diagnoses and patient comorbidity data obtained through chart review. For Aim 1, the area under the receiver operating characteristic curves (AUC) and sensitivity/specificity of the VHI-10 and mainstream clinical queries were determined. For Aim 2, Spearman rho analyses were applied to test the null hypothesis that there were no correlations between the VHI-10 or and PROMIS scores. For Aim 3, the influence of voice and comorbid conditions on general health status scores was assessed through multivariate ordinal regression. Results: For vocal fold paralysis, the VHI-10 score had an observed AUC of 0.78 (95% CI 0.68-0.88) and had better discrimination than 13 common clinical queries. For laryngeal stenosis, the VHI-10 score demonstrated moderate discrimination, with an associated AUC of 0.79 (95% CI 0.56-1.00) and higher discrimination than 5 common clinical queries. VHI-10 scores were moderately correlated with PROMIS social item scores (Spearman rho 0.37, p<0.0001), however there was no significant correlation between the SVHI-10 and PROMIS scores. In all analyses, voice-related quality of life was a significant predictor of general health status, even when adjusting for comorbid conditions (physical health β= -1.097, p<0.001; mental health β= -0.735, p<0.001; global item β= -1.334, p<0.001; social item β= -0.818, p<0.001). Conclusions: Our results indicate that (1) the VHI-10 is an effective predictor of some voice diagnoses, when compared to mainstream queries about vocal characteristics; (2) voice-specific measures correlate with the PROMIS assessment, however correlations are not so robust that voice-specific instruments and general quality of life instruments can be directly substituted; and (3) voice health has a significant, multi-dimensional impact on general health status, which is not subsumed by the presence of classically more concerning comorbid conditions.