Integration of Artificial Intelligence (AI) and Distributed Ledger Technologies to Improve Interpretability and Reportability for Point-of-Care Medica

登革病毒（DENV）是一种蚊媒病毒性疾病，其全球发病率在过去十年中急剧增加，目前估计全球每年的发病率高达3.9亿例感染。纸基微流体装置由于其相对小的尺寸（长度和宽度约为几mm至2 cm）、便携性和相对便宜的成本，已经显示出作为现场使用的护理点设置的各种成功。然而，大多数uPAD设备仍然无法被非专业用户解读。由于病毒感染倾向于被低估，以及这些地方性病毒发生的低资源环境导致就医率低下，这一问题更加复杂。使用人工智能（AI）和分布式账本技术的集成系统可以与微流体一起解决这些痛点。库珀教授的研究小组最近开发了一个带有区块链通信的人工智能平台，用于检测乌干达的疟疾。该研究使用监督学习算法通过照片识别积极和消极的结果，并链接到基于could的学习平台。然而，区块链技术的使用是能源密集型的，并且与云的数据分析链接需要持续的网络访问。在这个项目中，我们将建立新的分布式账本架构，实现边缘计算（在无需网络访问的情况下执行分析的设备上）。新平台将与DHIS2等国家数字卫生平台整合，以自动创建一个（非）集中式监测系统，改善先前发现的官方报告不力问题，这有可能导致新的公共卫生措施（例如减少病媒种群）。该集成系统的改进的可用性将潜在地改进甚至对于其专业知识点位于病毒感染本身之外的医学专家的可解释性（例如，一般卫生工作者，如护士或全科医生）。总的来说，这将提高医疗保健系统管理登革热病毒的稳健性，以应对未来的地方性疫情。

Dengue virus (DENV) is a mosquito-borne viral disease whose global incidence has increased dramatically in the past decade with current estimates indicating an incidence level of up to 390 million infections per year globally. Paper-based microfluidic devices have shown various successes as a point-of-care settings for field use due to their relatively small size (approximately a few mm to 2 cm in length and width), portability and relatively inexpensive cost. However, most uPAD devices are still unable to be interpreted by non-professional users. This problem is compounded by the virus' infections propensity to be under-reported and the low-resource settings these endemic viruses occur in leading to poor rates in seeking medical treatment.An integrated system using artificial intelligence (AI) and distributed ledger technologies could be utilized to tackle these pain points together with microfluidics. Professor Cooper's research group has recently developed an AI platform with blockchain communication for the detection of malaria in Uganda. The research used supervised learning algorithms to identify positive and negative results via photograph with links to could-based learning platform. However, the use of blockchain technologies is energy-intensive and the link to the cloud for data analysis requires constant network access. In this project, we will establish new distributed ledger architectures enabling edge computing (on the device performing the assay without the need for network access). The new platform will be integrating with national digital health platforms, such as DHIS2, to create a (de-)centralized surveillance system automatically, improving the previous identified problem of poor official reporting, which has the potential to lead to new public health measures (e.g. to reduce vector populations). The improved usability of this integrated system would potentially improve interpretability even for medical experts whose point of expertise lie outside of the viral infection itself (e.g., general health workers such as nurses or general practitioners). Overall, this would increase the robustness of healthcare systems management for dengue virus for future endemic outbreaks.