Hearing aid evaluation: predicting speech gain from insertion gain

In this study, hearing aid gain for speech was defined as the difference in level between the aided and unaided performance-intensity functions measured at any specific value of percentage of items correct. The articulation index method was used to predict speech gain based on the subject's unaided sound field thresholds, ambient room noise, hearing aid internal noise, hearing aid insertion gain, and the subject's unaided performance-intensity function. Predicted speech gain agreed with measured speech gain with rms errors of only 3 dB for 11 subjects with mild or moderate hearing loss tested with monosyllabic words and continuous discourse. The speech gain provided by a hearing aid can thus be predicted from electroacoustic measures, which generally can be obtained in a shorter time. Importance functions believed to be applicable to nonsense syllables, words, and continuous discourse were used to make the predictions, but prediction accuracy was not affected by the importance function chosen. Speech gain measured with the monosyllabic word test was highly correlated with speech gain measured with the continuous discourse test, provided that similar presentation levels were used.