The cost of repair and maintenance of medical devices can be fairly burdensome to the healthcare industry. Healthcare providers often consider plans such as increasing in-house repairs, using multiple repair service providers, or timely replacement of devices to overcome this issue. This study aims to develop a data-driven Markov Decision Process (MDP) framework based on Discrete-Time Markov Chain (DTMC) model to optimize medical equipment repair and replacement decisions. Effective decision-making on whether to repair or replace is crucial to managing the product lifecycle cost and the costs to healthcare facilities. The study determines the optimal repair or replacement decision based on the product lifecycle data and current product failure status. It utilizes a net present value model to maximize expected value over an infinite time horizon. The study uses a dataset of 24,516 repair and maintenance records of 5,171 individual medical devices of a particular type to extract parameters needed for the optimization model. The dataset provides a rich baseline for analyzing different failures and event modes during product lifespan such as battery-related issues, random failure, preventive maintenance, and physical damage. It further quantifies the chance of moving from one product status to another. The model outcomes are discussed for a particular case. The findings reveal the most frequent reasons for failures and the most economically viable repair and replacement decision for each end-of-use device based on their current condition. Several sensitivity analyses are conducted to clarify the impact of operating revenue and warranty time on the repair or replacement.

医疗设备的修理和维护成本可能对医疗保健行业造成沉重负担。医疗保健提供商通常考虑诸如增加内部维修，使用多个维修服务提供商或及时更换设备等计划来克服此问题。这项研究旨在开发基于离散时间马尔可夫链（DTMC）模型的数据驱动的马尔可夫决策过程（MDP）框架，以优化医疗设备的维修和更换决策。关于是否要维修或更换的有效决策对于管理产品生命周期成本和医疗机构成本至关重要。该研究根据产品生命周期数据和当前产品故障状态确定最佳的维修或更换决策。它利用净现值模型在无限的时间范围内最大化期望值。该研究使用了特定类型的5,171个单独医疗设备的24,516个维修和保养记录的数据集，以提取优化模型所需的参数。数据集提供了丰富的基线，可用于分析产品寿命期间的不同故障和事件模式，例如与电池有关的问题，随机故障，预防性维护和物理损坏。它进一步量化了从一种产品状态转变为另一种产品状态的机会。针对特定情况讨论了模型结果。这些发现揭示了每个最终用途设备根据其当前状况而出现故障的最常见原因以及最经济可行的维修和更换决定。进行了几项敏感性分析，以阐明营业收入和保修时间对维修或更换的影响。