Back to Annual Meeting Program

PURPOSE:
Preventable readmissions represent a major burden on the health care system and by risk stratifying patients resources can be directed to prevent these costly complications. This study examines patient characteristics, surgical factors, and postoperative complications associated with 30-day readmissions in gastrointestinal (GI) resections.

METHODS:
Inpatients undergoing major GI surgery were selected from the 2011 ACS National Surgical Quality Improvement Program prospectively collected database. Procedures were classified into esophageal, gastric, small bowel, large bowel, liver, and pancreatic resections using Common Procedural Terminology codes. Postoperative complications were divided into pre- and post-discharge groups by comparing time to complication and discharge. Operative times were grouped by 75th percentile(>/=4 hours). Univariate analysis using Chi-square, Mann Whitney-U, and Student's T-test were used to compare patient comorbidities, surgical characteristics, and postoperative complications with 30-day unplanned readmission rates. Factors with a p<0.1 were included in multivariate logistic regression. Odds ratios(OR) and 95% confidence intervals(CI) are reported and p-value <0.05 was considered statistically significant.

RESULTS:
For 43,894 patients undergoing GI resection, the overall 30-day unplanned readmission rate was 12.0% ranging from 11.4% for colorectal resections to 15.7% for pancreatic resections. Median postoperative length of stay was longer in the readmission group (7 vs. 6 days p<0.0001). Major predictors of 30-day readmissions included pre-discharge major complications (OR=1.28; CI: 1.14,1.44, p<0.0001), preoperative steroid use (OR=1.62; CI: 1.39,1.89, p<0.0001), operative time >/=4 hours (OR=1.61; CI: 1.45,1.78, p<0.0001) and discharge to a facility other than home (OR=1.48; CI: 1.28,1.70, p<0.0001). Other factors associated with increased readmission included dependent functional status, open surgery, pulmonary comorbidity, neurologic comorbidity, higher ASA score, diabetes, and preoperative anemia(table 1). Post-discharge major and minor complications were highly correlated with 30-day readmission rates (OR=59.3; CI: 52.2,67.3, p<0.0001 and OR=6.3; 95%CI: 5.8,6.9, p<0.0001) and not included in the final model.

CONCLUSIONS:
Unplanned 30-day readmissions represent a major medical and financial concern, but some may be foreseeable and thus preventable. Although previous studies have identified major complications as a strong risk factor for readmissions, this might represent an overestimate of the risk due to confounding by including post-discharge complications that may in fact cause, not predict, readmissions. This model provides insight into factors that could inform resource utilization and post-operative care to help prevent readmissions in select high-risk GI surgical patients.
Table 1: Factors associated with unplanned 30-day readmissions following GI resection.

Risk Factor	% Readmitted	Adjusted OR	95% CI	p-value
Preoperative Steroid use	18.6 v 11.5	1.62	1.39, 1.89	<0.0001
Operative time (>/=4hrs vs. <4hrs)	15.4 v 12.8	1.61	1.45, 1.78	<0.0001
Discharge Destination (Facility vs. Home)	15.1 v 11.6	1.48	1.28, 1.70	<0.0001
Open Surgery	13.1 v 9.1	1.46	1.30, 1.63	<0.0001
Pre-discharge Major Complication	14.9 v 11.1	1.28	1.14, 1.44	<0.0001
Neurologic Comorbidity	15.2 v 11.9	1.26	1.06, 1.51	0.01
Pulmonary Comorbidity	14.2 v 11.5	1.22	1.03, 1.45	0.02
Dependent Functional Status	15.8 v 11.8	1.24	1.01, 1.51	0.037
ASA (3/4 vs 1/2)	13.5 v 9.9	1.17	1.06, 1.30	0.002
Pre-discharge Minor Complication	12.8 v 11.9	1.16	0.98, 1.38	0.077
Diabetes	13.8 v 11.6	1.13	1.01, 1.27	0.032
Preoperative Anemia (HCT<36)	13.5 v 11.3	1.11	1.01, 1.22	0.024

Logistic regression model also controlled for age, hepatic insufficiency, cardiac comorbidity, renal insufficiency, wound class, smoking, bleeding disorder, chemotherapy/radiation, weight loss, preoperative sepsis, and LOS.