インドの農業トラクター運転手におけるデータマイニング技術を用いた聴力測定データの分析

Noise from agricultural tractors is a critical occupational health hazard, often leading to Noise Induced Hearing Loss (NIHL). This study aims to assess the risk of NIHL among Indian Agricultural Tractor Drivers (ATD) and enrich the understanding of such risks using unsupervised data mining techniques.

A cross-sectional study was conducted on audiometric profile of Indian ATD with driving experience ranging from 5 to 43 years, and audiometry testing was done with 0.125 kHz to 8 kHz of frequencies. Participants were selected from Bhopal, Madhya Pradesh, India. The k-means clustering techniques, an unsupervised learning method, was applied to classify the audiometric data. Z-scores were used to evaluate cluster homogeneity and separation, while ANOVA was performed to determine the significance of various factors, including age, experience, and weight, on hearing impairment.

The mean hearing threshold levels were the lowest for drivers with less than 10 years of experience and highest for those with over 10 years of experience. The mean age of the drivers at the time of testing was 39.5 (±10.2) years. The audiometric data did not follow a normal distribution, necessitating the use of k-means clustering for analysis of both the ears. All audiometric frequencies showed statistically significant between-cluster differences, though with notably lower F-values compared to the right ear, ranging from 4.946 at 8 kHz (p = .011) to 19.461 at 1 kHz (p < .001). While age, experience, and weight were not significant for some parameters, other factors showed significant impacts on hearing impairment. Notably, the effect of different frequencies on both right and left ears was significant, highlighting the potential risks associated with prolonged tractor operation.

The study demonstrates the feasibility of using k-means clustering to analyze audiometric data effectively. This method could play a vital role in hearing conservation programs for individuals exposed to occupational noise at agricultural workplaces. Raising legislative awareness and implementing customized safety programs to promote tractors with reduced noise levels are recommended.