Application of artificial intelligence model in managing patients with voice disorders
Main Article Content
Abstract
The study was conducted at Hanoi Medical University Hospital from 2024 to 2025 to evaluate the effectiveness of the artificial intelligence (AI) model AI-VoiceCare in managing patients with voice disorders. A total of 312 patients diagnosed with voice disorders were enrolled. The AI model utilized a convolutional neural network (CNN) to analyze Mel-spectrogram features and integrated the AI-VoiceCare system for outpatient electronic medical record management. The model achieved an accuracy of 92.1%, sensitivity of 0.90, and specificity of 0.93 in classifying pathological voices. Record-processing time was reduced by 61.1%, and documentation errors decreased by 60.9%. After six months of implementation, the proportion of patients returning for scheduled follow-up visits increased from 48.7% to 76.3%, and 68.5% of patients reported significant improvement in voice quality after AI-assisted therapy. In total, 84.3% of physicians rated the system as “useful” or “very useful.” The application of AI-VoiceCare enhances diagnostic accuracy, patient management, and voice rehabilitation, contributing to the development of a personalized care model for individuals with voice disorders.
Article Details
Keywords
Voice disorder, artificial intelligence, AI-VoiceCare, voice analysis
References
2. Verdolini K, Ramig LO. Review: occupational risks for voice problems. Logoped Phoniatr Vocol. 2001; 26(1): 37-46. doi:10.1080/140154301300109150.
3. Behlau M, Zambon F, Guerrieri AC, Roy N. Epidemiology of voice disorders. J Voice. 2017; 31(1): 127.e1-127.e7. doi:10.1016/j.jvoice.2015.08.017.
4. Zhang Y, Jiang J. Acoustic analysis in voice pathology detection. IEEE Trans Biomed Eng. 2020; 67(8): 2197-2205. doi:10.1109/TBME.2020.2969782.
5. Phạm TTB, Võ TH, Nguyễn DM. Ứng dụng trí tuệ nhân tạo trong chẩn đoán bệnh lý giọng nói. Tạp chí Tai Mũi Họng Việt Nam. 2023; 18(3): 45-52.
6. Alhindi T, Elshafeey A, Alotaibi M, et al. CNN-based detection of voice disorders. J Voice. 2022; 36(3): 412-420. doi:10.1016/j.jvoice.2021.03.010.
7. Paja W, Stylianou Y. Automatic detection of pathological voices. IEEE Access. 2021; 9: 10456-10467. doi:10.1109/ACCESS.2021.3051234.
8. Chen L, Zhang Y, Yang X, et al. Deep learning in voice pathology. Speech Commun. 2021; 128: 14-22. doi:10.1016/j.specom.2021.02.004.
9. Umapathy K, Krishnan S, Rao A, et al. Feature extraction for pathological voice detection. Biomed Signal Process Control. 2020; 59:101904. doi:10.1016/j.bspc.2020.101904.
10. Chien CY, Lee YC, Lin MH, et al. AI in laryngeal image diagnosis. Otolaryngol Head Neck Surg. 2022; 166(4): 703-710. doi:10.1177/01945998211066723.
11. Arora R, Sharma D, et al. Voice pathology detection using transfer learning. Comput Speech Lang. 2023; 80: 101470. doi:10.1016/j.csl.2022.101470.
12. Hu T, Wang X, Li Q, et al. Multimodal AI for laryngeal diseases. IEEE Trans Neural Netw Learn Syst. 2024; 35(2): 543-556. doi:10.1109/TNNLS.2023.3298715.
13. Wang S, Zhao Y, Lin H, et al. AI-powered screening of voice disorders. Expert Syst Appl. 2023; 224:119990. doi:10.1016/j.eswa.2023.119990.
14. International Organization for Standardization. ISO/IEC 19794-13:2021. Information Technology-Biometric Data Interchange Formats-Part 13: Voice Data. Geneva: ISO; 2021.
15. Patel RR, Huber JE, Kreiman J, et al. Clinical implications of AI in speech pathology. Int J Speech Lang Pathol. 2023; 25(3): 251-260. doi:10.1080/17549507.2022.2142658.