Intelligent Decision Support Systems for Prioritization in Health Care

等候名单是指名单上的一群人在等待通常供不应求的服务。这种清单存在于许多提供服务的环境中，如银行、零售商店、医院以及保健和社会服务诊所。挑战在于定义个人接收服务的优先顺序。根据预先定义的标准（例如重要性和重要性）对等候名单上的个人进行排序的过程称为优先排序。在现实世界中，在复杂和动态系统（如医疗保健和社会服务）的不确定性下做出优先级决策，需要考虑我们拥有的大量数据、相关风险、多个（有时相互冲突的）目标和供应限制，因此极具挑战性。该研究计划的总体目标是更好地理解和分析现实世界的优先问题（例如，在初级医疗保健环境中），并通过将运筹学（OR）与人工智能（AI）方法相结合，开发针对优先问题挑战的创新解决方案。本研究项目有三个研究方向。在研究方向1中，我们将为不确定和动态复杂环境下的优先级决策开发健壮的、数据驱动的优化解决方案。在研究方向2中，我们将使用OR和AI方法开发可解释的优先级解决方案。为优先级设计可解释的解决方案非常重要，因为这样的解决方案使用户能够理解所做决策背后的逻辑。这可以增加用户对开发的解决方案的信任，这在医疗保健等关键环境中非常重要。最后，人们对大数据及其潜在价值的商业和研究越来越感兴趣。统计和优化方法可以帮助分析这些大数据。在研究方向3中，我们将利用这些方法开发混合解决方案，以优先处理大数据。这项研究计划的目的是通过整合人工智能方法来开发针对优先问题的创新解决方案，并将其应用于初级卫生保健环境，从而在手术室领域取得突破性进展。每个研究方向都有明确的预期里程碑和产出，我们将使用这些里程碑和产出来评估进展，并确保我们按计划实现目标。该计划的影响将通过以下方式衡量：(a)对OR和人工智能领域的方法和知识的贡献；(b)在顶级科学期刊上发表的论文数量；(c)受过训练的高素质人员人数及其进展情况；(d)在实践中将已开发的知识和见解有效地转移给其他研究人员和决策者；最终(e)短期和长期的可交付成果将在加拿大产生学术、经济、社会、培训和健康方面的影响。

A waiting list is a queue of individuals who are on a list awaiting a service that is often in short supply. Such lists exist in many service-providing settings such as banks, retail stores, hospitals, and health and social services clinics.The challenge is to define the individual's order of priority to receive the service. The process of ranking individuals on the waiting list according to predefined criteria (e.g. importance and significance) is referred to as prioritization. In the context of real-world, making prioritization decisions happens under uncertainty in complex and dynamic systems (such as healthcare and social services) and requires considering big amount of data that we have, associated risks, multiple (sometime conflicting) objectives and supply constraints, is thus, extremely challenging. The overarching objective of this research program is to better understand and analyze real-world prioritization problems (e.g., in primary healthcare settings), and to develop innovative solutions for prioritization challenges by integrating operational research (OR) with artificial intelligence (AI) methods. This research program has three research directions. In the Research Direction 1, we will develop robust, data-driven optimization solutions for prioritization decisions under uncertainty and in dynamic and complex environments. In the Research Direction 2, we will develop interpretable solutions for prioritization, using both OR and AI methods. Designing interpretable solutions for prioritization is important because such solutions make it possible for the user to understand the logic behind the decisions made. This can increase users' trust in the developed solutions and this is important in critical settings such as healthcare. Lastly, there has been an increasing interest in business and research regarding big data and the potential value of its use. Statistical and optimization methods can help in analyzing these big data. In the Research Direction 3, using these methods we will develop hybrid solutions to deal with big data in prioritization. The aim of this research program is to lead ground-breaking advances in the field of OR by integrating AI methods to develop innovative solutions for prioritization problems, and applying them in the primary healthcare setting. Each research direction has clear intended milestones and outputs that we will use to assess progress and ensure we meet our objectives as planned. This program's impact will be measured by: (a) contributions to the methodology and knowledge in the OR and AI fields; (b) number of publications in top tier scientific journals; (c) number of highly qualified personal trained and their progress; (d) the effective transfer of the developed knowledge and insights to other researchers and also decision makers in practice; and ultimately (e) the deliverables in the short and long terms will lead to academic, economic, social, training and health-related impacts in Canada.