The rapid growth of IoT in healthcare results in large volume of healthcare data, analyzing such data helps in determining the patient health condition, identifying the disease etc. These data are distributed across various places and private in nature. InterPlanetary File System (IPFS) is often used by healthcare providers to securely store the raw data and Content Identifier (CID) is used to share it across different blockchain network. However, if CID is shared then the stored plaintext data will be available which is undesirable for highly sensitive healthcare data. Again, if we analyze data from a single healthcare provider it may result in a biased learning model. Thus, it necessitates robust privacy-preserving solutions for sensitive data sharing without depriving the analytics aspects of it. Traditional techniques based on encryption and decryption are not suitable as it is not scalable and does not support analytics on encrypted data. For this purpose, we proposed a system integrating Homomorphic encryption with blockchain technology to address these challenges. It allows us to securely share the sensitive health data in encrypted format across various healthcare provider network. This heterogeneous data is utilized to create a robust machine learning model while maintaining data privacy. The system is implemented using Hyperledger Fabric and the TenSEAL library. The devised architecture supports multiple Health Service Provider (HSP) networks which utilizes chaincodes and homomorphic encryption for data logging and storing it in IPFS enabling privacy preserving decentralized storage. The experimental results demonstrates that the proposed model exhibits high throughput and low latency and also maintains high accuracy while preserving the data privacy. This shows the effectiveness of the proposed approach in enhancing data security and privacy in digital healthcare environments.