Readmission Prediction on MIMIC-III (target)
74.42AUPRCOracle
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Oracletraining_mode=trained on target data2025.06 | 74.42 | 69.74 | — | — | — | |
| GraphCARE2026.06 | 73.4 | 69.7 | — | — | — | |
| RePrompTBackbone=Llama 3.1 1B2026.04 | 71.9 | 68.8 | — | — | — | |
| RePrompTsoft prompts (P)=102026.04 | 71.9 | 68.8 | — | — | — | |
| UDONCAREtraining_mode=trained on source data2025.06 | 71.17 | 67.68 | — | — | — | |
| Deepr2026.04 | 70.5 | 67.3 | — | — | — | |
| StageNet2026.04 | 70.2 | 67.6 | — | — | — | |
| DT-BEHRT2026.03 | 69.62 | 80.3 | 70.59 | — | — | |
| Adore2026.04 | 69.3 | 66.1 | — | — | — | |
| StageNet2026.06 | 69.3 | 66.7 | — | — | — | |
| GRASP2026.06 | 69.2 | 66.3 | — | — | — | |
| Deepr2026.06 | 68.8 | 66.5 | — | — | — | |
| ManyDGtraining_mode=trained on source data2025.06 | 68.63 | 67.03 | — | — | — | |
| AdaCare2026.06 | 68.6 | 65.7 | — | — | — | |
| GRU2026.06 | 68.2 | 65.4 | — | — | — | |
| AdaCare2026.04 | 68.1 | 64 | — | — | — | |
| CondAdvtraining_mode=trained on source data2025.06 | 67.72 | 65.47 | — | — | — | |
| RETAIN2026.04 | 67.6 | 66 | — | — | — | |
| HoT-SSMvariant=v12026.06 | 67.4 | 64.5 | — | — | — | |
| Transformer2026.06 | 67.3 | 63.9 | — | — | — | |
| GRAM2026.06 | 67.2 | 64.3 | — | — | — | |
| ARCI2026.04 | 67.1 | 65.2 | — | — | — | |
| IRMtraining_mode=trained on source data2025.06 | 67.07 | 65.93 | — | — | — | |
| COCONUTBackbone=Llama 3.1 1B2026.04 | 67 | 63.9 | — | — | — | |
| Prompt-TuningBackbone=Llama 3.1 1B2026.04 | 66.3 | 63.4 | — | — | — | |
| HoT-SSMvariant=v22026.06 | 66.1 | 63.2 | — | — | — | |
| MLDGtraining_mode=trained on source data2025.06 | 66 | 65.59 | — | — | — | |
| GRAM2026.04 | 66 | 62.7 | — | — | — | |
| SLDGtraining_mode=trained on source data2025.06 | 65.73 | 65.8 | — | — | — | |
| RETAIN2026.06 | 65.1 | 64.1 | — | — | — | |
| PCLtraining_mode=trained on source data2025.06 | 65.03 | 66.36 | — | — | — | |
| HEART2026.03 | 64.05 | 77.68 | 68.77 | — | — | |
| DANNtraining_mode=trained on source data2025.06 | 64.02 | 63.89 | — | — | — | |
| GRASP2026.04 | 63.9 | 61.7 | — | — | — | |
| Basetraining_mode=trained on source data2025.06 | 63.45 | 62.15 | — | — | — | |
| GRPO——advanceType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 63.18 | 57.2 | 69.12 | 53.97 | 60.87 | |
| Med-BERT2026.03 | 62.9 | 76.66 | 66.52 | — | — | |
| K2KType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 62.49 | 56.64 | 69.31 | 53.03 | 60.37 | |
| ExBEHRT2026.03 | 62.13 | 73.68 | 63.08 | — | — | |
| Prompt-basedType=LLM, Base Model=Meditron3-Qwen2.5-7B, Evaluation Protocol=Fine-tuned2026.04 | 60.25 | 55.38 | 60.15 | 43.01 | 54.7 | |
| K2KType=LLM, Base Model=Meditron3-Qwen2.5-7B, Evaluation Protocol=Fine-tuned2026.04 | 59.88 | 56.08 | 68.78 | 52.42 | 59.29 | |
| w/o retrieverType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 59.07 | 54.61 | 69.17 | 52.87 | 58.93 | |
| GRPOType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 58.9 | 54.7 | 66.8 | 51.2 | 57.9 | |
| Standard RAGType=LLM, Base Model=Meditron3-Qwen2.5-7B, Evaluation Protocol=Fine-tuned2026.04 | 58.14 | 53.9 | 63 | 45.98 | 55.26 | |
| w/o retrieverType=LLM, Base Model=Meditron3-Qwen2.5-7B, Evaluation Protocol=Fine-tuned2026.04 | 57.72 | 53.01 | 68.18 | 51.72 | 57.66 | |
| G-BERT2026.03 | 57.19 | 67.4 | 60.84 | — | — | |
| Standard RAGType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 57.09 | 52.99 | 69.73 | 53.52 | 58.33 | |
| KAREType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 56.67 | 50.97 | 61.64 | 44.55 | 53.46 | |
| HypEHR2026.03 | 56 | 68.28 | 48.09 | — | — | |
| KAREType=LLM, Base Model=Meditron3-Qwen2.5-7B, Evaluation Protocol=Fine-tuned2026.04 | 55.66 | 50.89 | 59.45 | 42.3 | 52.08 | |
| Prompt-basedType=LLM, Base Model=BioMistral-7B, Evaluation Protocol=Fine-tuned2026.04 | 54.19 | 49.71 | 66.51 | 49.82 | 55.06 | |
| RETAINType=ML2026.04 | 54.11 | 51.29 | 45.88 | 23.48 | 43.69 | |
| Zero-ShotBackbone=Llama 3.1 1B2026.04 | 53.7 | 50.5 | — | — | — | |
| GRUType=ML2026.04 | 52.94 | 50 | 68.28 | 51.84 | 55.77 | |
| AdaCareType=ML2026.04 | 52.83 | 52.27 | 63.49 | 46.51 | 53.77 | |
| DeeprType=ML2026.04 | 51.68 | 49.7 | 68.28 | 51.84 | 55.38 | |
| BEHRT2026.03 | 51.59 | 64.79 | 53.64 | — | — | |
| StageNetType=ML2026.04 | 50.38 | 48.27 | 51.56 | 34.74 | 46.24 | |
| TCNType=ML2026.04 | 49.84 | 47.65 | 65.46 | 48.66 | 52.9 |