Creation and Institutional Validation of a Readmission Risk Calculator for Elective Colorectal Surgery

BACKGROUND: Readmissions reflect adverse patient outcomes, and clinicians currently lack accurate models to predict readmission risk. OBJECTIVE: We sought to create a readmission risk calculator for use in the postoperative setting after elective colon and rectal surgery. DESIGN: Patients were identified from 2012-2014 American College of Surgery-National Surgical Quality Improvement Program data. A model was created with 60% of the National Surgical Quality Improvement Program sample using multivariable logistic regression to stratify patients into low/medium- and high-risk categories. The model was validated with the remaining 40% of the National Surgical Quality Improvement Program sample and 2016-2018 institutional data. SETTINGS: The study included both national and institutional data. PATIENTS: Patients who underwent elective abdominal colon or rectal resection were included. MAIN OUTCOME MEASURES: The primary outcome was readmission within 30 days of surgery. Secondary outcomes included reasons for and time interval to readmission. RESULTS: The model discrimination (c-statistic) was 0.76 ((95% CI, 0.75-0.76);p< 0.0001) in the National Surgical Quality Improvement Program model creation cohort (n = 50,508), 0.70 ((95% CI, 0.69-0.70);p< 0.0001) in the National Surgical Quality Improvement Program validation cohort (n = 33,714), and 0.62 ((95% CI, 0.54-0.70);p= 0.04) in the institutional cohort (n = 400). High risk was designated as >= 8.7% readmission risk. Readmission rates in National Surgical Quality Improvement Program and institutional data were 10.7% and 8.8% overall; of patients predicted to be high risk, observed readmission rate was 22.1% in the National Surgical Quality Improvement Program and 12.4% in the institutional cohorts. Overall median interval from surgery to readmission was 14 days in the National Surgical Quality Improvement Program and 11 days institutionally. The most common reasons for readmission were organ space infection, bowel obstruction/paralytic ileus, and dehydration in both the National Surgical Quality Improvement Program and institutional data. LIMITATIONS: This was a retrospective observational review. CONCLUSIONS: For patients who undergo elective colon and rectal surgery, use of a readmission risk calculator developed for postoperative use can identify high-risk patients for potential amelioration of modifiable risk factors, more intensive outpatient follow-up, or planned readmission.