Health Care Management Science最新文献

This work proposes a two-phase approach for optimizing the layout of a referral cancer center in Latin America, which results from a collaboration between a university, in Brazil, and a consultancy company. The objective of the problem is to minimize the total transportation cost of patients, medical/non-medical staff, materials, and equipment. In the first phase, an integer programming model is used to assign departments to floors in such a way that the vertical transportation cost between departments is reduced. In the second phase, a heuristic method is employed to determine the layout of the blocks across a given floor, when applicable, while minimizing the transportation cost within the floor. Process mining is employed to gather data associated with the movement flow between rooms and departments within the unit. The developed approach was used by the managers of the cancer center to help design a new hospital building, and the layout produced by the optimization procedure was compared with the initial layout originally built by the architects of the hospital. The results obtained demonstrate that our method was capable of significantly reducing both vertical ( $-19.7\%$ ) and horizontal ( $-22.7\%$ ) transportation costs within the hospital.

Life expectancy is typically shorter in areas with higher deprivation, highlighting the need for policymakers and health care managers to focus on reducing health inequalities through efficient and effective care. This study aims to assess the impact of deprivation on primary health care performance using data from the National Health Service (NHS) in England. Two methods are applied: Data Envelopment Analysis (DEA) to evaluate the performance of 188 Clinical Commissioning Groups (CCGs), whose duties were recently taken on by the new Integrated Care Systems (ICSs), and the Malmquist Index (MI) to assess deprivation's effect on performance. The DEA results reveal significant variation among CCGs in equity, efficiency, and effectiveness, indicating substantial room for improvement. The MI results show that while CCGs in more deprived areas had more resources per capita and higher efficiency, they were generally less effective than those in less deprived areas. This emphasizes the need to enhance health and social policies to address persistent health inequalities due to deprivation, a critical challenge for the new ICSs. This study illustrates how DEA and the MI can support policymakers and managers in this effort.

Ambulatory care facilities play a critical role in many healthcare systems worldwide. To ensure efficient care provision, we must match care demand with care supply. To support provider decision-making, this article reviews Operations Research planning problems, the corresponding planning and control decisions that must be made when opening up or running an ambulatory care facility, and their data requirements. We give an overview of demand and supply-related data that an ambulatory care facility can collect and comment on the consequences for decision-making if some of that data is missing. We briefly discuss three healthcare systems and their influence on data collection and decision-making. We also take a closer look at several real-world appointment data sets and their usefulness for planning decisions. In addition, we discuss model implementation barriers and give recommendations for modelers and practitioners to bridge the gap between theory and practice. Finally, we present future research directions for Operations Research in ambulatory care.

Cardiovascular diseases (CVDs) are one of the primary reasons for death worldwide. These diseases often occur due to the occlusion of coronary arteries, thereby leading to insufficient blood and oxygen supply that damages cardiac muscle cells. Electrocardiogram (ECG) signals which reflect heart electrical activity are being used for diagnosing various cardiac diseases. Typically, a standard ECG consists of 12 channels referred to as leads which enable practitioners to monitor heartbeats through different channels where each heartbeat lasts approximately 600 ms. The majority of studies focus on the classification and early diagnosis of arrhythmias. Although the current studies on change-point methods have acquired massive accuracy in detecting potential shifts during a multi-channel process, they lack flexibility in manually assigning more weights to the channels, which are of more importance for experts. This could be addressed by implementing the weighted multivariate functional principal component analysis (WMFPCA). The objective of this study is to develop a novel change-point detection method to monitor long-term cardiovascular treatment. A third-order tensor structure was employed to represent the 12-lead ECG data in three dimensions (beats × samples × leads). Exploiting intra-beat, inter-beat, and inter-lead correlations along with channel significance in the third-order tensor, the WMFPCA is incorporated into Hotelling's $T^2$ statistic to construct monitoring schemes. Simulation results show that the proposed approach outperforms the existing methods in monitoring multi-channel processes. Finally, applying the suggested model on a real-world dataset containing Myocardial Infarction (MI) subjects verifies the model.

Multi-billion-dollar opioid settlement agreements have been reached with pharmaceutical manufacturers and distributors to address their liability in contributing to the opioid epidemic in the United States. These agreements stipulate that within the state, the settlement funds must be directly allocated to local government (e.g., counties) and used for abatement activities to remediate the harm of the opioid epidemic in communities. This naturally leads to an important question of how the funds should be distributed to meet the diverse needs of the counties consistently across all counties to be deemed fair. Although there exist various definitions of fairness in the literature, it remains unclear how to empirically quantify the fairness of settlement allocation based on data, which is crucial for developing evidence-based allocation policies. To fill this gap, we define two allocation fairness measures, deviation and maximum regret, and formulate the fair settlement allocation as convex optimization problems. To further enhance the interpretability of the allocation policies, we restrict the allocation to a weighted sum of the given empirical metrics. We apply our analytical framework in a case study of the settlement allocation in Pennsylvania using real-world empirical metrics. We identify the frontiers of the non-dominated allocation policies between min-deviation and minimax-regret allocations, which dominate all alpha fairness-based and formula-based allocation policies. All allocation policies show lower fairness (with higher deviation or maximum regret) in counties that are rural, low-income, and with lower-ranking health factors. The price of interpretability is more significant in terms of maximum regret compared with deviation.

With the advancement in computing power and data science techniques, reinforcement learning (RL) has emerged as a powerful tool for decision-making problems in complex systems. In recent years, the research on RL for healthcare operations has grown rapidly. Especially during the COVID-19 pandemic, RL has played a critical role in optimizing decisions with greater degrees of uncertainty. RL for healthcare applications has been an exciting topic across multiple disciplines, including operations research, operations management, healthcare systems engineering, and data science. This review paper first provides a tutorial on the overall framework of RL, including its key components, training models, and approximators. Then, we present the recent advances of RL in the domain of healthcare operations management (HOM) and analyze the current trends. Our paper concludes by presenting existing challenges and future directions for RL in HOM.

This study examines the inpatient service efficiency of safety-net and non-safety-net hospitals using a two-stage approach at both the hospital and physician levels. For the hospital-level analysis, we conducted 430 Data Envelopment Analysis (DEA) models at the first stage to measure efficiency at the Diagnosis-Related Groups (DRG) level. In the second stage, Tobit and logistic regression models were applied to compare safety-net hospitals to non-safety-net hospitals. For the physician-level analysis, we conducted 386 DEA models to measure individual physician efficiency within specific DRGs. In the second stage, we compared the performance of the same physicians working in safety-net versus non-safety-net hospitals. The findings reveal that non-safety-net hospitals demonstrate significantly higher efficiency than safety-net hospitals. However, comparisons of the same physicians across settings show no significant differences in individual efficiency. This suggests that the efficiency gap arises not from the support or motivation provided by hospitals but from differences in the quality of physicians employed. These results underscore the need for policies that help safety-net hospitals attract and retain high-quality physicians to bridge the efficiency gap and better serve vulnerable populations.

Standing orders allow triage nurses in emergency departments (EDs) to order tests for target patients prior to a physician evaluation. Standing orders specify the medical conditions for which a triage nurse is permitted to order tests but typically do not specify the operational conditions under which ordering tests is desirable, from either a system or a patient point of view. We examine the operational impacts of standing orders on the ED as a whole, and propose a threshold policy for activating standing orders as a function of ED congestion. To parameterize the threshold policy we develop three simplified models: 1) an infinite-server model to derive an easily-computed feature for predicting whether activating standing orders would be beneficial, 2) a Jackson network model, to demonstrate that standing orders can lead to diverse outcomes for different patient populations, and 3) a Markov decision process model, to quantify the optimality gap for our threshold policy. We confirm the tentative findings from the simplified models in a more realistic setting using a simulation model that is calibrated with real data. We find that the threshold policy, with a threshold that is a simple function of the aforementioned feature, performs well across a wide range of parameter values. We demonstrate potential unintended consequences of the use of standing orders, including overtesting and spillover effects on non-target patients. Medical studies demonstrate that the use of standing orders decreases average ED length of stay (LOS) for target patients. Our research shows the importance of investigating the impact of standing orders on the ED as a whole.

Ambulance location in Emergency Medical Services (EMS) is a widely studied problem requiring efficient resource allocation within budgetary constraints. The literature has focused on enhancing EMS performance with limited attention given to their economic performance. This study addresses EMS performance with an emphasis on budget constraints by revising three coverage maximization models: the time-dependent Maximum Expected Coverage Location Problem (time-dependent MEXCLP), the multi-period Double Standard Model (mDSM), and the multi-period Queueing Maximal Availability Location Problem (Q-MALP-M2). These models are adapted to incorporate ambulance types, multi-period relocation, and budget constraints related to costs associated with ambulance station openings, ambulance acquisition, transport, and multi-period relocation. The revised models, along with two hybrid models (model 1 and model 2), were evaluated and compared using a discrete-event simulation model based on three key performance indicators: 1) coverage, 2) waiting time, and 3) time to arrive at the hospital. Additionally, the study investigates ambulance sharing as a policy to enhance EMS performance, wherein a single ambulance serves two patients whenever feasible. The study uses data from the Fez-Meknes region in Morocco, collected in 2021. Results indicate that hybrid model 1 outperformed the other models in most scenarios, as it allows for the decentralization of ambulances by investing the allocated budget in constructing new ambulance stations and acquiring new ambulances, contrasting with the other models that allocate almost the entire budget to purchasing new ambulances. Furthermore, the findings reveal that ambulance sharing significantly improves EMS performance, particularly under tightening budgetary restrictions and increasing demand; however, the benefits of ambulance sharing diminish as the allocated budget increases.

Benchmark efficiency analysis in public health typically focuses on hospitals rather than primary care providers. Data Envelopment Analysis (DEA) is widely used to assess resource efficiency among decision-making units (DMUs). However, traditional DEA struggles to differentiate between efficient units and is sensitive to the selection of inputs and outputs. Methods like super-efficiency and cross-efficiency address some of these limitations but often exclude outliers and may overlook efficiency related to specialisation. DEA Visualisation integrates DEA with multivariate statistical methods allowing for the identification of inefficiency sources and specialisation patterns without losing discriminatory power or removing extreme cases from the sample. This study analyses 82 public primary health centres in Madrid serving senior citizens in 2018. The findings reveal inefficiencies such as a preference for prescribing specific rather than generic drugs, increasing public health costs. Additionally, two extreme cases (outliers or mavericks) were identified as having high infrastructure costs and disproportionate staffing. Redistributing patients from overcrowded centres could enhance efficiency, while centres focused on preventive care showed greater cost-effectiveness, particularly in reducing prescription costs.