Particle Swarm Optimization (PSO) is a well-built ensemble-based optimization technique was driven by the social behaviour of fish and avian creatures. It surpasses in solving tricky optimization problems by recurrently improving candidate solutions based on the community intelligence of swarm of particles. Its prominence lies in simplicity, proficiency, adaptability, and sturdy capability to vast search domain, making it extensively pertinent over diverse domains. Over the last few years, PSO and its variations have attained significant consideration in healthcare applications, primarily for feature selection (FS) in high-dimensional medicinal datasets. In healthcare domain, the rapid growth and difficulty of medical data impose significant obstacles in accurate diagnostic outcomes and decision-making. FS is a decisive phase in scrutiny of healthcare data, as it helps diminish dimensionality, mend model accuracy, and identify major biomarkers. By electing applicable medical features, health practitioners can make rapider, more accurate treatment plans, ultimately refining patient outcomes. This paper provides an all-inclusive review of PSO including categorization into three classes standard, improved and hybrid versions are addressed, aiming on their application in FS for Clinical datasets such as disease diagnosis, prognosis, medical imaging, genomics, wearable remote health monitoring and medical data privacy are reviewed. The review methodology adheres to PRISMA guidelines, ensuring structured literature selection, screening, and evaluation.