This research paper presents a holistic approach to transform healthcare business strategies by predicting life expectancy for patients on ventilator support through a real-time, supervised, reinforcement machine learning model. The study addresses critical issues faced by the current healthcare system, including legal, social, and economic challenges related to end-of-life decision-making. Data collection is purely based on PRIMARY DATA collected from two different Private Hospitals. System Decision: Iterative Dichotomiser [ID3] algorithm is applied for construction of decision tree that obeys a top-down, greedy approach to split the dataset. Decision tree derived after first execution of ID3 algorithm, clearly depicts that during evaluation of Mortality Risk all Prognostic parameters (APACHE–II, SAPS, MMP, MODS, SOFA, LODS) do not affect the final outcome uniformly. The findings suggest that the proposed machine learning model can enhance group decision-making by enabling the legally appointed medical attorney/proxy and family members to jointly scrutinize treatment recommendations, ultimately leading to more informed and ethical healthcare practices. The paper underscores the importance of utilizing data-driven strategies for optimal resource allocation in healthcare and introduces the potential for future advancements through reinforcement learning techniques. The results indicate a need for larger datasets to improve predictive accuracy and support complex decision-making in critical care environments.