Evidence-driven spatiotemporal COVID-19 hospitalization prediction with Ising dynamics.
Evidence-driven spatiotemporal COVID-19 hospitalization prediction with Ising dynamics.
复制标题
DOI：
10.1038/s41467-023-38756-3
复制
发表时间：
2023-05-29
影响因子：
16.6
通讯作者：
Sun, Jimeng
中科院分区：
文献类型：
--
作者：
Gao, Junyi;Heintz, Joerg;Mack, Christina;Glass, Lucas;Cross, Adam;Sun, Jimeng
在这项工作中，我们旨在通过建立时空预测模型来准确预测COVID-19大流行期间的住院人数。我们提出了一种基于Ising动态的深度学习模型葫芦，用于COVID-19住院时间的时空预测。通过在统计力学中绘制位置和点阵位置之间的类比，我们使用伊辛动力学来指导模型提取和利用位置之间的空间关系，并模拟来自现实世界临床证据的颗粒信息的复杂影响。利用美国各地丰富的关联数据库，包括保险理赔、人口普查信息和医院资源使用数据，我们在美国2299个县的大尺度时空COVID-19住院预测任务上对HOIST模型进行了评估。在4周住院预测任务中，HOIST平均绝对误差达到368.7，一致性相关系数平均得分为0.6和0.89。我们详细的治疗所需人数（NNT）和成本分析表明，未来的COVID-19疫苗接种工作可能在农村地区最具影响力。该模型可作为未来县和州一级疫苗接种工作的资源。在2019冠状病毒病大流行期间，准确的住院预测至关重要。在这里，作者表明，基于统计力学的深度学习模型能够预测住院情况，支持有针对性的疫苗接种工作。
In this work, we aim to accurately predict the number of hospitalizations during the COVID-19 pandemic by developing a spatiotemporal prediction model. We propose HOIST, an Ising dynamics-based deep learning model for spatiotemporal COVID-19 hospitalization prediction. By drawing the analogy between locations and lattice sites in statistical mechanics, we use the Ising dynamics to guide the model to extract and utilize spatial relationships across locations and model the complex influence of granular information from real-world clinical evidence. By leveraging rich linked databases, including insurance claims, census information, and hospital resource usage data across the U.S., we evaluate the HOIST model on the large-scale spatiotemporal COVID-19 hospitalization prediction task for 2299 counties in the U.S. In the 4-week hospitalization prediction task, HOIST achieves 368.7 mean absolute error, 0.6 and 0.89 concordance correlation coefficient score on average. Our detailed number needed to treat (NNT) and cost analysis suggest that future COVID-19 vaccination efforts may be most impactful in rural areas. This model may serve as a resource for future county and state-level vaccination efforts. Amid the COVID-19 pandemic, accurate hospitalization predictions are vital. Here, the authors show that a deep learning model based on statistical mechanics is able to forecast hospitalizations, supporting targeted vaccination efforts.
登录
查看更多内容
影响因子：
5.8
作者：
Gao, Junyi;Yang, Chaoqi;Heintz, Joerg;Barrows, Scott;Albers, Elise;Stapel, Mary;Warfield, Sara;Cross, Adam;Sun, Jimeng
通讯作者：
Sun, Jimeng
影响因子：
13.6
作者：
Godoy-Lorite A;Jones NS
通讯作者：
Jones NS
DOI：
10.1093/jamia/ocaa322
发表时间：
2021-03-18
期刊：
Journal of the American Medical Informatics Association : JAMIA
影响因子：
--
作者：
Gao J;Sharma R;Qian C;Glass LM;Spaeder J;Romberg J;Sun J;Xiao C
通讯作者：
Xiao C
影响因子：
3.7
作者：
Coletti P;Libin P;Petrof O;Willem L;Abrams S;Herzog SA;Faes C;Kuylen E;Wambua J;Beutels P;Hens N
通讯作者：
Hens N
影响因子：
13.8
作者：
Bennett TD;Moffitt RA;Hajagos JG;Amor B;Anand A;Bissell MM;Bradwell KR;Bremer C;Byrd JB;Denham A;DeWitt PE;Gabriel D;Garibaldi BT;Girvin AT;Guinney J;Hill EL;Hong SS;Jimenez H;Kavuluru R;Kostka K;Lehmann HP;Levitt E;Mallipattu SK;Manna A;McMurry JA;Morris M;Muschelli J;Neumann AJ;Palchuk MB;Pfaff ER;Qian Z;Qureshi N;Russell S;Spratt H;Walden A;Williams AE;Wooldridge JT;Yoo YJ;Zhang XT;Zhu RL;Austin CP;Saltz JH;Gersing KR;Haendel MA;Chute CG;National COVID Cohort Collaborative (N3C) Consortium
通讯作者：
National COVID Cohort Collaborative (N3C) Consortium