CT-HEG: A Bidirectional, Timestamp-Attributed Event Graph for ICU In-Hospital Mortality Prediction - An Architectural Ablation Study
CT-HEG: A Bidirectional, Timestamp-Attributed Event Graph for ICU In-Hospital Mortality Prediction - An Architectural Ablation Study
CT-HEG：一种用于ICU住院死亡率预测的双向时间戳属性事件图——架构消融研究
Abstract: Accurate ICU mortality prediction requires modeling irregular clinical observations across heterogeneous entity types. Existing sequence models handle irregular sampling but ignore typed relational structure; existing graph models assume fixed-interval inputs. We introduce the Continuous-Time Heterogeneous EHR Graph (CT-HEG) schema and evaluate which architectural choices drive predictive performance.
摘要： 准确的ICU死亡率预测需要对跨异构实体类型的非规则临床观察进行建模。现有的序列模型可以处理非规则采样，但忽略了类型化的关系结构；而现有的图模型则假设输入为固定时间间隔。我们引入了连续时间异构电子健康记录图（CT-HEG）架构，并评估了哪些架构选择能够提升预测性能。
CT-HEG encodes each ICU stay as a typed, timestamped graph with three node types (visit, vital, lab_event) and 2D edge attributes (t_hours/48, value_norm) encoding timing and value without imputation. We instantiate CT-HEG as CHIRP-Net, a four-layer heterogeneous GATv2Conv network, evaluated on MIMIC-IV v3.1 (31,142 ICU stays, LOS>=48h, 13.4% mortality) with five seeds and bootstrapped confidence intervals, against logistic regression, mTAND, a Transformer, and GRU-D, plus an ablation study.
CT-HEG将每次ICU住院记录编码为一个带有时间戳的类型化图，包含三种节点类型（就诊、生命体征、实验室事件）和二维边属性（t_hours/48，value_norm），在无需插补的情况下对时间和数值进行编码。我们将CT-HEG实例化为CHIRP-Net，这是一个四层异构GATv2Conv网络。我们在MIMIC-IV v3.1数据集（31,142次ICU住院，住院时长>=48小时，死亡率13.4%）上进行了评估，采用了五个随机种子和自助法置信区间，并与逻辑回归、mTAND、Transformer和GRU-D进行了对比，同时进行了消融研究。
CHIRP-Net achieved 5-seed mean AUROC 0.8449+/-0.0071 (AUPRC 0.4958+/-0.0209); the ensemble achieved AUROC 0.8618 (95% CI: 0.8485-0.8745). Removing reverse edges disconnected observation nodes from the visit readout, cutting AUROC by 0.1968+/-0.0073. Time-attentive edge features contributed 0.0247+/-0.0093 AUROC. Collapsing heterogeneous edge types into one relation (7x fewer parameters) outperformed the full model on all seeds. Post-calibration ECE was 0.0307.
CHIRP-Net在五个种子下的平均AUROC为0.8449+/-0.0071（AUPRC为0.4958+/-0.0209）；集成模型达到了0.8618的AUROC（95%置信区间：0.8485-0.8745）。移除反向边会导致观察节点与就诊读出节点断开，使AUROC下降了0.1968+/-0.0073。时间注意力边特征贡献了0.0247+/-0.0093的AUROC。将异构边类型合并为单一关系（参数减少了7倍）在所有种子测试中均优于完整模型。校准后的预期校准误差（ECE）为0.0307。
Temporal and demographic subgroup analyses were explored but not reported here, pending follow-up work. Bidirectional connectivity was necessary for the model to use its inputs at all, and CT-HEG was reasonably well calibrated after validation-fitted temperature scaling. These results support CT-HEG for irregular EHR data, while external validation, a pre-specified temporal evaluation, and a demographic fairness audit remain necessary before any claim of robustness.
我们探索了时间维度和人口统计学子组分析，但由于后续工作尚未完成，此处未予报告。双向连接对于模型利用输入信息至关重要，且在经过验证集拟合的温度缩放后，CT-HEG表现出良好的校准性。这些结果支持将CT-HEG应用于非规则电子健康记录数据，但在宣称其稳健性之前，仍需进行外部验证、预先指定的时间评估以及人口统计学公平性审计。