Machine learning and IoT in healthcare: Recent advancements, challenges & future direction

摘要

The integration of Machine Learning and Deep Learning with IoT-enabled devices for real-time health monitoring has significantly revolutionized healthcare. These technologies facilitate the analysis of intricate medical datasets, yielding actionable insights that promote evidence-based clinical decision-making. Although significant advancements have been made, there is still an absence of a thorough synthesis regarding current applications, primary challenges, and prospective research directions. This review aims to synthesize recent applications, identify significant gaps, and propose clear direction for future research. A comprehensive narrative review was performed where a systematic literature search was conducted in PubMed and Scopus for studies published between 2020 and 2025. A total of 300 pertinent papers on ML and IoT's applications in healthcare were selected and analyzed to synthesize technological advancements, trade-offs, practical implications, challenges, and potential directions for future research. Neural network models, such as CNNs and ANNs, along with ensemble methods like Random Forest and XGBoost, often attain predictive accuracies ranging from 85 % to 95 %. Advanced technique, like generative imaging models, reinforcement learning, and transformer-based architectures, improve diagnostics, chronic disease management, robotic-assisted surgery, and predictive analytics, while explainable AI promotes clinical trust. Cloud-edge integration utilizing lightweight machine learning models enables real-time, energy-efficient applications, enhancing diagnosis, decision support, personalization, and cost-effectiveness, notwithstanding current challenges. To conclude, the integration of ML and IoT is transforming healthcare through enhanced monitoring, improved predictive capabilities, and tailored treatment approaches. Addressing persistent limitations is crucial for fully realizing its potential and directing future research in this evolving field.