Predicting Acute Pancreatitis Length of Stay: ‘Hotspotting High-cost Admissions 90

• Published in: The American journal of gastroenterology Vol. 113; no. Supplement; pp. S49 - S50
• Main Authors: Bowns, Emily, Hinton, Alice, Lara, Luis F., Hart, Phil, Conwell, Darwin
• Format: Journal Article
• Language: English
• Published: New York Wolters Kluwer Health Medical Research, Lippincott Williams & Wilkins 01.10.2018
• Subjects: Ablation; Cysts; Gastroenterology; Length of stay; Pancreatitis
• ISSN: 0002-9270; 1572-0241
• DOI: 10.14309/00000434-201810001-00090

Introduction: AP LOS directly impacts costs and resource allocation. We chose to develop a risk score to identify subjects with AP at risk for prolonged hospitalization. 1. Determine risk factors that lead to an increased LOS for patients with AP. 2. Derive a risk score to stratify AP LOS. Methods: The 2013 Nationwide Readmissions Database (NRD), was used to develop the AP risk score. The NRD contains information on 36 million discharges, containing ICD-9/ICD-10 diagnosis codes. Included in the analysis were adults (>18 years of age) with a primary diagnosis of AP. Excluded were patients who were pregnant, transferred during their stay, or died during admission. SAS 9.4 was used to develop a logistic regression model identifying significant independent predictors of LOS > 7 days. Based on the coefficients and standard errors from this model, a risk score was developed for LOS > 7 days. Results: Overall there were 217,535 patients included in the analysis (52% male). LOS was strongly associated with costs of admission. (Figure 1; Correlation coefficient 0.71; p-value <0.001). Predictors of LOS > 7 days included age, comorbidities, admission day, hospital type and size, morbid obesity, gallstones, and alcohol use. An AP LOS risk score was developed from these predictors. The AP LOS risk scores were normally distributed (Figure 2) and ranged from 0-105 (Table 1). The model (OR; 95% CI) was stratified into low (reference), moderate (2.09; 1.97-2.21) or high risk (4.15; 3.93-4.39) with an AUC of 0.644 for increased LOS > 7 days. Conclusion: 1. A risk score for LOS was derived from administrative claims data. 2. Further external validation and testing of this AP risk score is planned. 3. Clinical Implication: "Hotspotting" AP can facilitate reallocation of resources to a subset of high-needs, high-cost AP patients.