clinical significance or statistical significance: which one should be considered for analyzing the gain- frequency responses of hearing aids?

Background and aim: when analyzing the gain-frequency response of a hearing aid, data can be analyzed both clinically and statistically. this study aimed to investigate and compare the gain-frequency responses using statistical and clinical analyses under active digital noise reduction (dnr-on) and inactive digital noise reduction (dnr-off) conditions. methods: the gain-frequency responses of a hearing aid for one of the most well-known commercial digital hearing aid manufacturers were measured using the fp35 hearing aid analyzer (frye electronics inc., usa) by presenting two types of signals (digital speech and composite noise) at input levels of 65 and 80 db spl under the dnr-on and dnr-off conditions. data analysis was performed both statistically (using wilcoxon signed rank test) and clinically (using 3 db difference criterion). results: a statistically significant difference was found in the gain-frequency responses for all speech and noise input levels between the two conditions; while a clinically significant difference was observed only at noise input levels of 65 and 80 db spl. conclusion: for analyzing the hearing aid performance, both clinical and statistical analyses should be considered.