Hospitals have collected extensive patient health data over the last few years, including reports on clinical tests, treatment updates, and disease diagnoses. Clinical decisions and treatment recommendations are based on the information extracted from this data. Collaborative filtering techniques have achieved significant success in health recommender systems, but traditional algorithms face challenges related to data sparsity and scalability, resulting in lower accuracy and efficiency. This paper proposes a model for predicting patient needs and making real-time storage decisions based on health data. Machine learning classifiers are trained to map health data attributes to storage features using training sets derived from correlations identified in small samples of expert data. Data sparsity and scalability issues have been overcome with the proposed machine learning approach, making it a suitable alternative to the collaborative filtering technique. The integration of machine learning classifiers into health data systems enhances the prediction of patient needs and optimizes real-time storage decisions, providing a more accurate and efficient solution compared to traditional collaborative filtering techniques.