This study introduces a set of metrics for evaluating temporal preservation in synthetic longitudinal patient data, defined as artificially generated data that mimic real patients' repeated measurements over time. The proposed metrics assess how synthetic data reproduces key temporal characteristics, categorized into marginal, covariance, individual-level and measurement structures. We show that strong marginal-level resemblance may conceal distortions in covariance and disruptions in individual-level trajectories. Temporal preservation is influenced by factors such as original data quality, measurement frequency, and preprocessing strategies, including binning, variable encoding and precision. Variables with sparse or highly irregular measurement times provide limited information for learning temporal dependencies, resulting in reduced resemblance between the synthetic and original data. No single metric adequately captures temporal preservation; instead, a multidimensional evaluation across all characteristics provides a more comprehensive assessment of synthetic data quality. Overall, the proposed metrics clarify how and why temporal structures are preserved or degraded, enabling more reliable evaluation and improvement of generative models and supporting the creation of temporally realistic synthetic longitudinal patient data.
Longitudinal modeling of MS patient trajectories improves predictions of disability progression
KU Leuven（鲁汶大学）；Hasselt University（哈瑟尔特大学）；Charles University in Prague（布拉格查理大学）；General University Hospital（综合大学医院）；University of Bari（巴里大学）；Hospital Universitario Virgen Macarena（马卡雷纳圣母大学医院）；Dokuz Eylul University（多库兹埃卢尔大学）；University G. d’Annunzio（邓南遮大学）；University Hospital Basel（巴塞尔大学医院）；University of Basel（巴塞尔大学）
Deep Survival Analysis of Longitudinal EHR Data for Joint Prediction of Hospitalization and Death in COPD Patients
B2SLab, Institut de Recerca i Innovació en Salut (IRIS), Universitat Politècnica de Catalunya（B2SLab、创新与健康研究所（IRIS）、巴塞罗那理工大学）；Networking Biomedical Research Centre in the subject area of Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN)（生物医学研究网络（CIBER-BBN））；Institut de Recerca Sant Joan de Déu（圣约翰德德乌研究所）；Evidenze Health；Servei de Vigilància Epidemiològica i Resposta a Emergències de Salut Pública a Barcelona Sud, Subdirecció Regional a Barcelona, Secretaria de Salut Pública, Departament de Salut, Generalitat de Catalunya（巴塞罗那南部公共卫生流行病监测与应急响应服务，巴塞罗那地区副秘书处，公共卫生秘书处，加泰罗尼亚自治区卫生部门）；Fundació Docència i Recerca Mútua Terrassa, Catalonia, Spain（特拉维撒互惠教育与研究基金会，加泰罗尼亚，西班牙）
NOAH: Learning the Full Patient Journey. A Longitudinal Multimodal Time-Aware Model for Representation and Forecasting
TUM University Hospital and Technical University of Munich (TUM)（慕尼黑工业大学医院与慕尼黑工业大学）；Institute for Diagnostic and Interventional Radiology, TUM University Hospital（慕尼黑工业大学医院诊断与介入放射学研究所）；Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心）；Institute for Diagnostic and Interventional Neuroradiology, TUM University Hospital（慕尼黑工业大学医院诊断与介入神经放射学研究所）；University Medical Center Hamburg-Eppendorf（汉堡埃彭多夫大学医学中心）
Patient Similarity Analysis with Longitudinal Health Data
University of Zurich（苏黎世大学）；University Hospital of Zurich（苏黎世大学医院）；University of Applied Sciences and Arts Northwestern Switzerland（西北瑞士应用科学与艺术大学）；Yale University School of Medicine（耶鲁大学医学院）
Utilizing Longitudinal Chest X-Rays and Reports to Pre-Fill Radiology Reports
National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health（美国国立卫生研究院国家医学图书馆国家生物技术信息中心）；National Institutes of Health Clinical Center（美国国立卫生研究院临床中心）；Weill Cornell Medicine（威尔康奈尔医学院）
Validated Synthetic Patient Generation for Small Longitudinal Cohorts: Coagulation Dynamics Across Pregnancy
Robert F. Smith School of Chemical and Biomolecular Engineering, Cornell University（康奈尔大学罗伯特·F·史密斯化学与生物分子工程学院）；Department of Biochemistry, The Robert Larner, M.D. College of Medicine, University of Vermont Medical Center（佛蒙特大学医学中心罗伯特·拉纳医学院生物化学系）；Department of Obstetrics, Gynecology, and Reproductive Sciences, The Robert Larner, M.D. College of Medicine, University of Vermont Medical Center（佛蒙特大学医学中心罗伯特·拉纳医学院妇产科与生殖科学系）
EHR2Path: Comprehensive Pathway-Level Modeling of Longitudinal Patient Trajectories from Multimodal Electronic Health Records
Technical University of Munich（慕尼黑技术大学）；TUM School of Computation, Information and Technology（慕尼黑技术大学计算、信息与技术学院）；Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心）
'A net for everyone': fully personalized and unsupervised neural networks trained with longitudinal data from a single patient
German Cancer Research Center (DKFZ)（德国癌症研究中心）；Heidelberg University（海德堡大学）；University Hospital Essen (AöR)（埃森大学医院）；University Medicine Essen（埃森大学医学院）；German Cancer Consortium (DKTK)（德国癌症联盟）
High-Fidelity Longitudinal Patient Simulation Using Real-World Data
Department of Biomedical Informatics, Graduate School of Medicine, The University of Tokyo（东京大学医学系研究生院生物医学信息学部门）；Department of Healthcare Information Management, The University of Tokyo Hospital（东京大学医院医疗信息管理部）；Department of Cardiology and Nephrology, Mie University Graduate School of Medicine（名护大学医学研究生院心内科与肾内科部门）；Department of Emergency and Critical Care Medicine, Nippon Medical School（日本医疗学校急诊与重症医学部）；Artificial Intelligence and Digital Twin in Healthcare, Graduate School of Medicine, The University of Tokyo（东京大学医学研究生院人工智能与数字孪生在医疗中的应用部门）；Graduate School of Health Data Science, Juntendo University（立命馆大学健康数据科学研究生院）
Uncertainty-Aware Prediction of Lung Tumor Growth from Sparse Longitudinal CT Data via Bayesian Physics-Informed Neural Networks
Department of Mathematics, Vanderbilt University（范德比大学数学系）；John A. Paulson School of Engineering and Applied Science, Harvard University（哈佛大学约翰·A·保罗森工程与应用科学学院）
Machine learning using longitudinal prescription and medical claims for the detection of nonalcoholic steatohepatitis (NASH)
IQVIA（艾昆纬）
目的开发和评估机器学习模型，以检测疑似未确诊的非酒精性脂肪性肝炎（NASH）患者，用于诊断筛查和临床管理。方法在这项回顾性观察性非干预性研究中，使用来自1，463，089例患者的行政医疗索赔数据，训练梯度提升决策树，以从具有肥胖、2型糖尿病（T2 DM）、代谢紊乱或非酒精性脂肪肝（NAFL）病史的高危患者人群中检测可能的NASH患者。对模型进行训练，以检测所有风险患者或既往无NAFL诊断的亚组（非NAFL风险患者）中可能的NASH。使用回顾性医疗索赔数据对模型进行训练和验证，并使用精确召回面积和受试者工作特征曲线（AUPRC，AUROC）进行评估。在索赔数据中，NASH的6个月发病率为1/1，437例高危患者和1/2，127例非NAFL高危患者。训练用于检测所有风险患者中NASH的模型的AUPRC为0.0107（95% CI 0.0104 - 0.011），AUROC为0.84。在10%的召回率下，模型精度为4.3%，是NASH发病率的60倍。训练用于检测非NAFL患者中NASH的模型的AUPRC为0.003（95% CI 0.0029 - 0.0031），AUROC为0.78。在10%的召回率下，模型精度为1%，是NASH发病率的20倍。结论NASH在医疗索赔数据中的低发生率证实了NASH在临床实践中的漏诊模式。基于索赔的机器学习可以促进检测可能的NASH患者，以进行诊断测试和疾…
Mortality Prediction with Adaptive Feature Importance Recalibration for Peritoneal Dialysis Patients: a deep-learning-based study on a real-world longitudinal follow-up dataset
Peking University（北京大学）；Health Data Research（健康数据研究所）；University of Edinburgh（爱丁堡大学）；Peking University Third Hospital（北京大学第三医院）；Peking University People’s Hospital（北京大学人民医院）；University of Exeter（埃克塞特大学）
EVA: Generating Longitudinal Electronic Health Records Using Conditional Variational Autoencoders
Georgia Institute of Technology（佐治亚理工学院）；IBM Research（IBM研究院）；Vanderbilt University（范德堡大学）；Sutter Health（萨特健康）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Multi-classifier prediction of knee osteoarthritis progression from incomplete imbalanced longitudinal data
School of Computing Science, Newcastle University（纽卡斯尔大学计算科学学院）；University Medical Center Utrecht（乌得勒支大学医学中心）；Merck（默克公司）；Bioscience Institute, Newcastle University（纽卡斯尔大学生物科学研究所）；Institut de Recherches Internationales Servier（赛诺菲国际研究所）；GlaxoSmithKline（葛兰素史克公司）；Delft University of Technology（代尔夫特理工大学）；State Research Institute Centre for Innovative Medicine（创新医学国家研究中心）；University of Oulu（奥卢大学）；Versus Arthritis（对抗关节炎组织）
Graph Representation Learning of Longitudinal Medical Imaging Trajectories for Treatment Response Prediction
Technical University of Munich（慕尼黑工业大学）；TUM University Hospital Rechts der Isar（慕尼黑工业大学伊萨尔河右岸大学医院）；Helmholtz Munich（慕尼黑亥姆霍兹中心）；Universitat de Barcelona（巴塞罗那大学）；Medical University of Vienna（维也纳医科大学）；King’s College London（伦敦国王学院）；Munich Center for Machine Learning（慕尼黑机器学习中心）
