Automatic COVID-19 pregnant patient prognosis algorithm

我们的社会能否从COVID-19大流行中恢复过来、保护患者生命并恢复正常生活，不仅取决于我们能否早期诊断，还取决于我们能否预测哪些患者面临严重危及生命的并发症的高风险，从而从早期干预中受益。在怀孕患者的情况下，我们预测这一点的能力变得非常具有挑战性，他们的免疫力被改变，导致各种不同的表现和疾病的进展。为了更好地了解COVID-19在怀孕人群中的进展情况，我们需要一个关于既往怀孕COVID-19病例的大型数据库，但目前无法获得。 为了解决这个问题，我们的短期目标是从国家和国际来源创建一个大型COVID-19妊娠数据库，同时确保没有数据重复。 一旦创建了存储库，由于患者症状、临床结果和疾病结果之间的关系非常重要，手动分析数据集仍然非常具有挑战性。因此，我们的目标是利用先进的人工智能方法和我们的大型数据库，为孕妇创建一个快速可靠的COVID-19疾病诊断系统，同时考虑病毒行为随时间的可能变化。 作为该项目的最后阶段，我们将在加拿大和北美最大的产科医院之一西奈山医院以及所有安大略合作医院将该算法投入临床使用。 该数据库和算法有望从根本上改变加拿大以及潜在的国际医疗保健提供者识别、管理和治疗孕妇COVID-19病例的能力。它们将使我们更深入地了解各种妊娠患者参数之间的相互作用，以决定疾病的结局。

Our society's ability to recover from the COVID-19 pandemic, preserve patient lives, and resume normal routines depends on our ability to not only diagnose early, but also to predict which patients are at high risk of severe life-threatening complications and would, therefore, benefit from early intervention. Our ability to predict this becomes extremely more challenging in the case of pregnant patients, whose immunity is altered, leading to various and different manifestations and progression of their disease. To better understand the progression of COVID-19 in the pregnant population, we require a large data repository of prior pregnant COVID-19 cases, which is currently unavailable. To address this, our short term aim is to create a large COVID-19 in pregnancy repository from National and International sources, while ensuring there is no data duplication. Once a repository is created, it will still be very challenging to analyze the dataset manually due to non-trivial relationships between patient symptoms, clinical results, and disease outcomes. Thus, our aim is to use advanced artificial intelligence approaches with our large data repository to create a fast and reliable COVID-19 disease prognostication system for pregnant patients while accounting for possible changes in the virus' behaviour over time. As a final stage in this project, we will bring the algorithm into clinical use at Mount Sinai Hospital, one of the largest obstetric hospitals in Canada and North America, and to all Ontario partner hospitals. The repository together with the algorithm are expected to radically transform Canadian and, potentially, international healthcare providers' ability to identify, manage and treat cases of COVID-19 in pregnant patients. They will provide us with a deeper understanding of the interactions between the various pregnant patient parameters in dictating disease outcome.