Manca Davide, Caldiroli Dario, Storti Enrico
PSE-Lab, Process Systems Engineering Laboratory, Dipartimento di Chimica, Materiali e Ingegneria Chimica "Giulio Natta", Politecnico di Milano - Piazza Leonardo da Vinci 32, 20133 Milano, Italy.
Neuroanestesia e Rianimazione Fondazione IRCCS Istituto Neurologico Carlo Besta, Milano, Italy.
Comput Chem Eng. 2020 Sep 2;140:106945. doi: 10.1016/j.compchemeng.2020.106945. Epub 2020 Jun 4.
The different stages of Covid-19 pandemic can be described by two key-variables: ICU patients and deaths in hospitals. We propose simple models that can be used by medical doctors and decision makers to predict the trends on both short-term and long-term horizons. Daily updates of the models with real data allow forecasting some key indicators for decision-making (an Excel file in the Supplemental material allows computing them). These are beds allocation, residence time, doubling time, rate of renewal, maximum daily rate of change (positive/negative), halfway points, maximum plateaus, asymptotic conditions, and dates and time intervals when some key thresholds are overtaken. Doubling time of ICU beds for Covid-19 emergency can be as low as 2-3 days at the outbreak of the pandemic. The models allow identifying the possible departure of the phenomenon from the predicted trend and thus can play the role of early warning systems and describe further outbreaks.
重症监护病房（ICU）患者数量和医院死亡人数。我们提出了简单的模型，可供医生和决策者用于预测短期和长期的趋势。使用实际数据对模型进行每日更新，能够预测一些用于决策的关键指标（补充材料中的Excel文件可用于计算这些指标）。这些指标包括床位分配、住院时间、翻倍时间、更新率、每日最大变化率（正/负）、中点、最大平台期、渐近条件，以及一些关键阈值被突破的日期和时间间隔。在疫情爆发时，新冠疫情紧急情况下ICU床位的翻倍时间可能低至2至3天。这些模型能够识别该现象与预测趋势可能出现的偏差，因此可以起到早期预警系统的作用，并描述进一步的疫情爆发情况。
