Achieving interoperability is a significant challenge for Internet of Things (IoT) developers due to the diverse communication protocols, data formats, and IoT devices, as well as the absence of global standards. We propose an ontology-driven semantic interoperability approach for big data analytics in IoT (ODSIAB-IoT) to address this issue, enhancing healthcare analytics by enabling semantic interoperability among diverse healthcare IoT systems. The model recommends appropriate medications and identifies potential adverse effects based on symptoms collected from various IoT devices. We utilize two datasets: one that links diseases with treatments, and another that details medications and their side effects. A lightweight model efficiently handles big data from diverse healthcare IoT devices. The Hadoop Distributed File System (HDFS) stores the data, which the Resource Description Framework (RDF) represents in triples for semantic clarity. RDF annotations, guided by ontologies, significantly enhance semantic interoperability. The model is validated using photoplethysmography (PPG) sensor data collected from 53 patients, achieving an accuracy 88.68% in disease prediction and medication recommendation. We employ SPARQL queries for RDF-based data analysis, which enables efficient retrieval of disease-specific parameters and demonstrates the efficacy of the ODSIAB-IoT model in enhancing semantic interoperability and healthcare decision-making.