A comparison of otoscopy and tympanometry in the diagnosis of middle ear effusion

Ears were examined otoscopically by one or more of five experienced observers who were each asked to decide whether a middle ear effusion was present. Their decisions were then ordered according to the probability, estimated from tympanometry, that an effusion was present at the time of the observation. The decisions of three of the observers were distributed randomly among the ranked observations. Since it is known that tympanometry provides a reasonably accurate indication of the presence of an effusion it is concluded that otoscopic evidence is often unreliable.