Resident Harvard Medical School / Mass Eye and Ear Otolaryngology–Head and Neck Surgery Boston, Massachusetts
Objective: (1) To examine the 10-item Voice Handicap Index-10 (VHI-10) instrument with item response theory (IRT), (2) to assess whether the dominant latent trait could be accurately measured using a subset of items, to improve efficiency of use in high-volume clinical scenarios, and (3) to determine whether the VHI-10 items correlate with diagnoses made via laryngoscopy/stroboscopy.
Methods: Adults with a primary voice complaint completed the VHI-10 upon presenting for evaluation at a tertiary care academic center. Factor analysis and IRT evaluation of prospective cohort data was performed to assess unidimensionality and item properties. Regression methods were used to evaluate correlation with objective diagnoses made by laryngoscopy/stroboscopy.
Results: The VHI-10 is unidimensional. Information parameter estimates ranged from 1.54 to 4.49, with higher values indicating more useful items. Item information functions demonstrated which items provided less value compared with others, and were removed to create a 5-item scale with preserved ability to reflect the underlying construct. There was no difference in the ability of the 10- and 5- item scales to predict voice diagnoses.
Conclusions: The VHI-10 demonstrates strong psychometric properties, as well as correlational validity with laryngoscopy/stroboscopy findings. IRT and other modern psychometric methods can be used to improve efficiency by maximizing the amount of relevant information per item administered, while reducing survey fatigue.