Research map / Probabilistic and scientific modeling / Treatment effects
PEACE: A Dataset of Pharmaceutical Care for Cancer Pain Analgesia Evaluation and Medication Decision
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Introduces PEACE, a multidisciplinary cancer-pain pharmaceutical care dataset (~38,766 patients, 103 Delphi-selected features) and benchmarks 13 ML/DL models on treatment effectiveness classification and medication recommendation regression.
Problem
Existing datasets lack multidisciplinary team (MDT) assessments, patient self-perceptions, and long-term follow-up needed for treatment effectiveness assessment (TEA) and medication recommendation (MR) in cancer pain management.
Approach
A cancer-pain medication dataset combining hospital records and an online follow-up platform, with features chosen through a two-round Delphi expert consensus, including pharmacist DRP evaluations and patient-reported adherence and impact on life; benchmarked with classical, boosting, time-series and EHR-specific models.
Key results
- LightGBM achieved the highest TEA accuracy (0.8023) and recall (0.7297) among the 13 models, with XGBoost close behind (0.7947 accuracy).
- On the MR regression task, tree-based models (decision tree, random forest, LightGBM) achieved the top results in most metrics, with e.g. random forest ERSO MSE 0.0189.
- The PEACE dataset comprises 38,766 patients, 103 features, and 2,601 patients with multiple (>=2) records.
Surprising findings
- Neural network models, including advanced time-series and tabular architectures, underperformed tree-based models on this tabular dataset and were prone to overfitting.
Limitations
- Expert panel drawn from only nine Chinese provinces, introducing subjectivity.
- Dataset from a single hospital population may limit generalizability; models not validated in clinical settings.
- Access requires training, a data use agreement, and institutional approval.
Experimental setup
Datasets: PEACE
Benchmarks: PEACE
Metrics: accuracy AUROC F1 score recall precision MSE MAE
Baselines: Decision Tree Logistic Regression Random Forest SVM MLP XGBoost LightGBM AdaBoost TransTab iTransformer Mamba StageNet
Builds on: Delphi method MIMIC-IV AdaCare StageNet
Closest papers by meaning
- KG-TREAT: Pre-training for Treatment Effect Estimation by Synergizing Patient Data with Knowledge Graphs (AAAI 2024 · similarity 0.83)
- Constructing Confidence Intervals for Average Treatment Effects from Multiple Datasets (ICLR 2025 · similarity 0.82)
- Unraveling Pain Levels: A Data-Uncertainty Guided Approach for Effective Pain Assessment (AAAI 2024 · similarity 0.82)
- Data-Driven Discovery of Dynamical Systems in Pharmacology using Large Language Models (NeurIPS 2024 · similarity 0.82)
- LLM-Driven Treatment Effect Estimation Under Inference Time Text Confounding (NeurIPS 2025 · similarity 0.81)
- Pharmacokinetics-Informed Neural Network for Predicting Opioid Administration Moments with Wearable Sensors (AAAI 2024 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.