Introduction and Validation of a Novel Acute Pancreatitis Digital Tool: Interrogating Large Pooled Data From 2 Prospectively Ascertained Cohorts

Abstract

Objectives: Acute pancreatitis (AP) is a sudden onset, rapidly evolving inflammatory response with systemic inflammation and multiorgan failure (MOF) in a subset of patients. New highly accurate clinical decision support tools are needed to allow local doctors to provide expert care.

Methods: Ariel Dynamic Acute Pancreatitis Tracker (ADAPT) is a digital tool to guide physicians in ordering standard tests, evaluate test results and model progression using available data, propose emergent therapies. The accuracy of the severity score calculators was tested using 2 prospectively ascertained Acute Pancreatitis Patient Registry to Examine Novel Therapies in Clinical Experience cohorts (pilot University of Pittsburgh Medical Center, n = 163; international, n = 1544).

Results: The ADAPT and post hoc expert-calculated AP severity scores were 100% concordant in both pilot and international cohorts. High-risk criteria of all 4 severity scores at admission were associated with moderately-severe or severe AP and MOF (both P < 0.0001) and prediction of no MOF was 97.8% to 98.9%. The positive predictive value for MOF was 7.5% to 14.9%.

Conclusions: The ADAPT tool showed 100% accuracy with AP predictive metrics. Prospective evaluation of ADAPT features is needed to determine if additional data can accurately predict and mitigate severe AP and MOF.