RESUMO
The aim of the study is to evaluate the possibility to implement machine learning to create a digital auditory profile for elderly patients and to analyze the hearing aid fitting efficacy depending on involvement of the peripheral and central auditory pathways in a pathological process. Data analysis of 375 people aged 60-93 years is presented. 355 patients with chronic bilateral hearing loss (230 of them used hearing aids) were included in the main group, and 20 normal hearing elderly people were included in the control group. Audiological examination consisted of standard tests (pure tone audiometry, impedancemetry, speech audiometry in quiet) and tests to evaluate the central auditory processing (binaural fusion, dichotic digits, speech audiometry in noise, random gap detection). The Montreal Cognitive Assessment was used to detect cognitive impairment. The hearing aid fitting efficiency was evaluated with COSI questionnaire and speech audiometry in free field. Processing of the results was carried out using Pearson's correlation analysis aimed at creating a polynomial model of a patient's hearing on the basis of the limited test battery. There were close correlations between the state of cognitive functions and age, results of tests to evaluate the central auditory processing, as well as patients' satisfaction of hearing aid. The results of the work indicate the possibility of using computer technologies of data analysis to develop rehabilitation programs for elderly hearing impaired patients.