Exploring Patient Focused and Machine Learning interventions for Rare Disease Diagnosis

在该项目中，患者持有的数据主题将与人工智能驱动的方法一起利用，以跟踪各种个人的旅程，以提示潜在的罕见疾病。通常，这一领域的工作集中在管理医疗保健提供者持有的数据（例如，Hu 2016），但在目前英国IT系统的结构中，由于数据流的整合较差，这将难以实现。相反，该项目将探索使患者能够在疾病诊断中发挥更大作用的机会。在罕见疾病组中，患者在诊断后往往成为“专家伙伴”（Aymé 2008年），然而，在诊断过程中，患者缺乏基本知识，无法对潜在疾病的讨论作出贡献。通过将干预措施交给患者，有可能就潜在的疾病进行早期对话。在以前的作品中，一种这样的方法已经看到了“医疗护照”的使用，它被用作一种方式来告知个人治疗中的众多利益相关者他们的健康状况和预定的预约[Leavey 2016]。（Bowen 2010），该项目将设计一个基于数字的健康旅程护照干预，以跟踪个人与医疗保健提供者的互动。这一过程的结果将告知：数字干预的关键要求，将用于捕获患者的健康状况-已经诊断为罕见疾病的患者的经验的知识库以及他们参与的途径。然后，该知识库将形成一个查询集，该查询集将提交给门户网站，如SAIL数据库项目，以产生一个数据集，可用于分析更多罕见疾病患者的途径。然后采用机器学习和数据分析的方法（例如，Perer 2013），将比较其他诊断患者之间的相似旅程模式，并与患者输入的症状一起向患者概述潜在疾病。然后，干预将提供罕见疾病的候选列表，并使患者能够突出可能向医疗保健提供者解释病情的潜在罕见疾病。罕见病的性质通常意味着许多医疗服务提供者不熟悉疾病本身或疾病的典型症状（Vandeborne 2016）。在创建患者数字工具和开发分析旅程的系统之后，该系统将在Amicus和其他项目合作伙伴确定的目标成员中进行野外测试。本评价的主要目的将围绕参与本实践的患者的经验形成，并更定量地评价数据驱动方法。

In this project, the theme of a patient held data will be leveraged alongside AI driven approaches to track the journeys of various individuals for the purpose of suggesting potential rare diseases. Typically, work in this area has focused on curating data held by healthcare providers (e.g., Hu 2016), but in the current structure of UK based IT systems this would be difficult to achieve due to poor integration of data streams.Instead, this project will explore opportunities to empower patients to have a greater role in the diagnosis of their diseases. Within rare disease groups, patients often become `expert partners' (Aymé 2008) after diagnosis, however while on the journey to diagnosis patients are lacking in fundamental knowledge to be able to contribute todiscussions about potential conditions. By placing an intervention in the hands of patients, there is potential for earlier conversations to be had about potential conditions. One such approach trailed in previous works has seen the use of a 'healthcare passport' utilised as a way to inform a multitude of stakeholders in an individuals treatment of the status of their health and scheduled appointments [Leavey 2016].By following a participatory design process (Bowen 2010), this project will design a digital based health journey passport intervention to track an individual's engagements with healthcare provides over time. The outcome of this process will inform both:the key requirements for a digital intervention that will be used to capture patient health engagements-a knowledge base of the experiences of patients already diagnosed with rare conditions and the pathways that they engaged withThis knowledge base will then form a query set that will be submitted to a portal such as the SAIL Databank project in order to produce a data set that can be used to analyse the pathways of a greater number of rare disease patients. By then adopting methods of machine learning and data analytics (e.g., Perer 2013), similar journey patterns amongst other diagnosed patients will be compared to and potential diseases will be outlined to a patient, in partnership with patient entered symptoms. The intervention will then provide a candidate list of rare disease and enable the patient to highlight potential rare diseases that might explain a condition to a healthcare provider. The nature of rare diseases typically means that many healthcare providers are unfamiliar with either the disease itself, or the symptoms that are typical of a condition (Vandeborne 2016).Following on from the creation of a patient digital tool, and a system to analyse journeys has been developed, the system will be tested in the wild with target members identified by Amicus and other project partners. A key aim of this evaluation will be formed around the experiences of patients engaging in this practise, and also to evaluate more quantitively the data driven approach.