1990 年 11 巻 2 号 p. 45-48
Accuracy of six different procedures, Jerger formula (Jerger), Niemeyer-Sesterhenn formula (NS), Baker-lilly formula (B-L), broad-band noise regession by Hyde (Hyde I), multiple regression on broad-band noise by Hyde (Hyde II) and bivariate plotting method by Silman et al. (B-P), used in predicting hearing level from the stapedial reflex threshold data were compared from 63 normal hearing and 66 hearing-impaired ears.
As results, none of both predictors suggested by Hyde was accrurate for clinical use. With each formula of Jerger, N-S and B-L, hearing level was correctly predicted in more than one half of all ears. However, the three formulae showed a low rate (23~35%) in accuracy predicting a mild or moderate loss. With the B. P, no method accurately estimated magnitude of hearing loss while maintaining a high proportion (ca.85%) of separation from normal hearing ears with the hearing-impaired on the basis of position on the bivariate graph.
Among these methods, the Jerger suggested to be adequate to predict normal hearing for younger children, and the B. P was particularly useful in identifine simulation in adults as well as functional deafness in elder children.