| dc.description.abstract | In healthcare systems worldwide, the manual
assignment of ICD-10 (International Classification of
Diseases, 10th Revision) codes presents significant
challenges, including resource constraints, lengthy
processing times, and potential inaccuracies. This
comprehensive review of the literature summarizes and
analyzes existing research on automated ICD-10 coding
systems, focusing on machine learning methodologies such
as decision trees, natural language processing (NLP), and
deep learning models. The review comprehensively
evaluates the performance, accuracy, and implementation
challenges of these techniques across diverse healthcare
settings. By examining studies from multiple healthcare
settings, this paper highlights the potential of automated
systems to improve diagnostic precision, reduce manual
workloads, and enhance overall healthcare efficiency. The
evaluation highlights major obstacles, including data
availability, integration with existing systems, and the need
for ongoing training of healthcare professionals, with brief
implications for developing countries like Sri Lanka.
Finally, this comprehensive analysis recommends future
research areas to help automated ICD-10 coding systems
become more widely used, which would ultimately lead to
better healthcare outcomes worldwide. | en_US |
