David Roberson has chosen two distinct paths of research. His recent basic studies focus on hair cell regeneration; his studies of practice patterns focus on medical errors.

In the first, he has currently developed and refined an avian model for hair cell damage. In this work, a single large dose of gentamicin is used to achieve essentially synchronous total hair cell death in the basal one third of the cochlea. As the hair cells in the basal cochlea regenerate, he maintains a continuous high concentration of the cell division marker bromodeoxyuridine via a pump-cannula system implanted in the inner ear. Since all original hair cells are killed by the gentamicin treatment, all hair cells seen at the end of the experimental period are new, regenerated cells. Because bromodeoxyuridine is present continuously at high levels in the cochlea, all dividing cells are labeled and all unlabeled cells are known to have arisen via direct transdifferentiation--the conversion of other cell types to hair cells. This model provides a unique environment in which to study direct transdifferentiation, because each cell's mode of regeneration can be identified.

In the second, he recently conducted a survey of practice errors among 2,500 members of the American Academy of Otolaryngology-Head and Neck Surgery. Two hundred ten physicians--45 percent of respondents--reported that a medical error had occurred in their practice in the past six months. The design of the study suggested a system for classifying and tracking errors.