Pediatric Gastrointestinal Risk Dataset (PedGI-FRAD)

The PedGI-FRAD 2026 dataset is a large-scale longitudinal pediatric gastrointestinal healthcare dataset developed for federated gastrointestinal risk prediction in distributed healthcare environments. The dataset contains 485,605 temporally ordered healthcare records collected at 15-minute intervals until March 2026 from five distributed pediatric hospital clients operating under heterogeneous healthcare conditions. The dataset was constructed using clinically grounded pediatric gastrointestinal monitoring variables derived from authentic healthcare observation protocols, wearable physiological sensing streams, hospital electronic health record systems, pediatric laboratory investigations, symptom progression monitoring frameworks, and digital health interaction environments. The dataset includes heterogeneous multimodal pediatric healthcare information covering demographic indicators, gastrointestinal symptom progression patterns, inflammatory laboratory biomarkers, nutritional and lifestyle measurements, medication history, relapse progression indicators, wearable physiological observations, and federated healthcare context variables. Gastrointestinal symptom attributes include abdominal pain severity, bloating patterns, nausea progression, vomiting episodes, diarrhea frequency, constipation duration, stool abnormalities, appetite variation, and nocturnal gastrointestinal discomfort. Laboratory biomarker variables include C-reactive protein, erythrocyte sedimentation rate, fecal calprotectin, white blood cell count, platelet count, hemoglobin, albumin, vitamin D, and iron-related indicators. In addition, the dataset incorporates temporal progression descriptors such as symptom progression trends, relapse frequency, stool abnormality dynamics, and longitudinal weight-loss progression to support sequential pediatric gastrointestinal intelligence modeling. Wearable and IoT-driven physiological variables including resting heart rate, heart-rate variability, sleep efficiency, stress score, and body temperature measurements are also integrated within the dataset. Federated healthcare context attributes such as hospital-client distribution, device source, missing-value ratio, and regional healthcare information are included to simulate realistic heterogeneous non-IID pediatric healthcare environments across distributed medical institutions. The primary prediction target of the dataset is GI_Risk_Level, representing binary pediatric gastrointestinal risk prediction, while GI_Severity_Level provides an additional clinical stratification label for gastrointestinal severity assessment.