In this paper, a framework for optimizing data transfer efficiency in cloud-integrated healthcare Internet of Things (IoT) systems is proposed. The framework aims to enhance real-time health monitoring and predictive analytics while addressing the challenges of bandwidth consumption, latency, and data overload in healthcare environments. Data from IoT-enabled healthcare devices, such as wearable sensors and patient monitoring systems, are first collected and pre-processed. Outlier detection is performed using the Isolation Forest algorithm, which identifies anomalous data points, ensuring accurate data for further analysis. Feature extraction is performed using ResNet, a deep learning model, to effectively capture hierarchical features from sensor and environmental data. To optimize data transfer, the MQTT protocol is employed for its lightweight, low-bandwidth communication capabilities, reducing overhead during data transmission. Selective data transmission is then applied, prioritizing critical and time-sensitive information, such as ECG or oxygen levels. Cloud processing involves the integration of IoT data with medical records for predictive health analytics and optimized resource allocation. The proposed framework improves system efficiency, ensuring seamless and secure data flow in healthcare IoT systems, ultimately enhancing patient care and health monitoring.