{
  "abstract": "Background Multiple clinical risk scores have been developed for acute upper gastrointestinal bleeding (AUGIB) with limited sensitivity and specificity. We aim to develop a machine learning model to identify high-risk AUGIB patients associated with poor outcomes and to compare its performance with existing risk scores.Methods A population-wide database was developed with all AUGIB patients in 6 major hospitals in Hong Kong. The study endpoint is a composite endpoint of poor outcomes including mortality (all-cause mortality and bleeding-related mortality), rebleeding within 30 days, need for ICU, re-endoscopy, surgery and embolization. 15 factors were included ( IDDF2026-ABS-0249 Table 1). Machine learning-based predictive models were developed to estimate the probability of poor outcomes in patients with AUGIB. Several machine learning algorithms (decision trees, random forests, support vector machines, and gradient boosting machines (GBM)) were trained. Model performance was assessed using AUROC, sensitivity, and specificity. The results were validated with a 10-fold split and cross-validation.Results 6503 AUGIB patients were included, of whom 1500 patients had poor outcomes. The AUROC, sensitivity and specificity of the model with GBM were 0.9233, 0.8313 and 0.8563 respectively. Moreover, the GBM model outperforms the existing clinical risk scores in terms of the AUROC, F1 score and specificity of GBM ( IDDF2026-ABS-0249 Table 2). The AUROC, F1, sensitivity, specificity of GBS>=12 and Rockall Score >=4 were 0.751, 0.497, 0.588, 0.747 and 0.714, 0.483, 0.744, 0.565 respectively (IDDF2026-ABS-0249 Figure 1. The ROC curves of GBM model and other clinical risk scores).Conclusions The GBM model outperforms other clinical risk scores. This new model may replace existing clinical risk scores with higher accuracy.Abstract IDDF2026-ABS-0249 Figure 1Abstract IDDF2026-ABS-0249 Table 1The list of 15 factors included for the machine learning modelFactors1. Age2. Presence of cancer3. Systolic blood pressure (mmHg)4. Acidosis in Venous/ arterial blood gas5. Serum albumin6. Alkaline phosphatase (IU/L)7. INR8. Serum sodium (mmol/L)9. Serum bilirubin (umol/L)10. Serum urea (mmol/L)11. No. of packed cells transfused12. No. of platelets transfused13. Forrest classification14. Need for endoscopic therapy during Esophagogastroduodenoscopy (EGD)15. Need for ICU admissionAbstract IDDF2026-ABS-0249 Table 2Comparison of GBM model with other clinical risk scoresScore(Cut-off)GBMGBS(>=12)ROCKALL(>=4)AIMS65(>=2)ABC(>=4)TP266188238183218FP141248427200288FN5413282137102TN840733554781693AUC0.92330.7510.7140.7390.749F1 score0.73180.4970.4830.5210.528Sensitivity0.83130.5880.7440.5720.681Specificity0.85630.7470.5650.7960.706",
  "authors": [
    {
      "affiliations": [
        "The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Shannon Chan"
    },
    {
      "affiliations": [
        "The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Louis Lau"
    },
    {
      "affiliations": [
        "The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Azalea Tse"
    },
    {
      "affiliations": [
        "The University of Hong Kong, Hong Kong"
      ],
      "name": "Yui-Lun Ng"
    },
    {
      "affiliations": [
        "The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Ka-Wai Kwok"
    },
    {
      "affiliations": [
        "The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Philip Chiu"
    }
  ],
  "title": "IDDF2026-ABS-0249 Development and validation of a machine learning model to identify high-risk patients associated with poor outcomes in patients with acute upper gastrointestinal bleeding (AUGIB)",
  "uid": "e0410707-c100-51eb-a443-d949614b3a31"
}