Improving Acute GI Bleeding Management Through Artificial Intelligence: Unnatural Selection?

Excerpt

Gastrointestinal bleeding (GIB) is the leading digestive disorder responsible for hospitalizations in the USA [1], a statistic not expected to change in the near future given the current shift of care toward more medically complex and anticoagulated patients, who are at increased risk for GIB. Accurate risk stratification of patients with GIB at initial presentation can facilitate improved triage efficiency and superior allocation of hospital-based resources. The ideal risk stratification tool should have both high positive and negative predictive values, which would result in low-risk patients’ discharge for outpatient follow-up and early endoscopy in those with high-risk predictors. A health system embracing this model would likely reduce costs while potentially improving meaningful clinical outcomes such as overall mortality and hospital length of stay. …